CONTROLS and DRIVETRAIN
STICKS
CONTROL RODS
CLUTCH DRUM
& FAN
TRANSMISSION
& MAST
MAIN ROTOR HEAD
MAIN BLADES
TAIL ROTOR
MISCELLANEOUS
General
One of the things to watch out for when working with and assembling the stick controls is both freedom of movement and a lack of slop ... it can be a delicate balance. During initial checks, one should first look at all the areas where components pivot and verify that there is no binding. I found that in several rotating areas where welding had occured, there was some binding due to heat penetration that needed to be cleaned up in order to get smooth pivoting. The most difficult one was the center of the pitch pivot tube (the one with three linkage arms) since the center welded area is not easily accessible and heat distortion appeared to have slightly warped the tube. This was attacked with a flapper wheel and a brake hone on an extension ... since the center weld area is not the main bearing area, I could be relatively aggresive but I still did repeated checks so as not to remove too much material and weaken the tube. Another builder's was heat warped enough that he had to actually straighten the tube in a press.
I chose to paint all the sliding / rotating surfaces with Perma-Slik G which does cause a bit of a build-up and dimensional change. Although it takes extra work, this actually helped to get a smooth low-slop fit. I found that most of these coated surfaces needed to be burnished to remove any excess Perma-Slik G and this process allowed me to get the exact fit that I wanted. Areas that had a very tight fit were started with maroon Scotch-Brite but most of this burnishing was done with grey (the finest that I had) Scotch-Brite. Final polishing was done with just shop towel.
Cyclic Sticks
There are several changes that I have seen made to the sticks.
Some builders have chosen to go with the "bent" type,
which allows more clearance for the seat bottom cushion and easier entry. Until I've designed,
fabricated and installed the seats I'm not sure if this going to
be a problem. I can also see where the "bent" stick
allows more room for sideways movement before hitting the pilot's
legs; again I'll have to try remember to check this out and think
about slope landings when I'm doing that. It should be noted that
the height of the cyclic stick will probably need to adjusted for
the individual pilot in order that their arm can comfortably rest
on their leg but still allow full movement. Again, this is
something that should be deferred until the seats are installed.
Update: There is no question that the stock stick with the
rigging manual throws will hit my legs, especially when aft of
neutral. I will eventually look at making new sticks but for now
I'm just going to try finish the craft. I need to do a lot more
measurements, but it seems to me that instead of just a curved
stick, the ideal solution would actually have a bit of a straight
horizontal part at the top. This would allow maximum clearance
for the pilot's legs.
Another change I've seen is to add reinforcements and then make the right stick removable via the use of a pin or bolt. This does make sense for safety reasons when flying with a non-pilot in the passenger seat. It would be possible to remove the right cyclic as designed and delivered but this also requires the removal of the tie rod at the bottom of the stick which would be inconvenient. This is a change I will definately consider after completion, as it would be an easy retrofit. In anticipation of this change, I've added a D-sub connector at the bottom of the stick so the electricals can be easily disconnected.
The factory seems to be very concerned about any slop in the
cyclic system and have introduced both a frame brace and a
recommendation to put tack welds on the stop collar to eliminate
play. While these may have benefit, my bigger concern for slop in
the cyclic system is with the bellcranks at the rear end of the
torque tubes. Because of the various tubing sizes used, there is
a slight gap between the torque tube and the bellcrank which the
factory has compensated for by putting slits in the bellcranks
that will be squeezed together as the bolt is tightened. I intend
to add some shim stock in here since I'm concerned that one would
be tempted to over-torque the bolt in order to get a proper fit.
A bigger concern to me are the bolt holes that were pre-drilled
at the factory and allow for a rotation of the bellcranks that is
only eliminated by the compression force of the bolt. With the
bolt installed and lightly snugged, I can see several degrees of
play in both bellcranks as the result of over-sized holes. I'm
not sure which way I'll go, but I'm either going to add some weld
material to the holes and then ream them to 1/4" or more
likely I'll over-size them to reamed 5/16" holes. Before
doing either of these, I'll check to see whether close-tolerance
bolts could eliminate the issue, but I don't think so.
Update: After painting the torque tube and the bores of the
bellcranks, it would appear that the paint has taken up all of
the play that existed between them; in fact, I had to lightly
sand a couple of high spots. For now, I am going with just the
stock setup but I will be monitoring this area to see if any slop
works its way into here.
Another place where I noticed a bit
of play is the holes in the ears on the cyclic torque tube that
hold the pivot bolt for the cyclic stick. When I installed a
regular AN4 bolt, I could feel some play in it and obviously the
laser cutting is at least .250" rather than being slightly
undersize and reamed. Since the cyclic grip moment arm is much
longer than pivot to rod end moment, this means that any slop in
this area will be magnified by the ratio of these two arms. I've
tried an AN174 close tolerance bolt and it's somewhat better but
its still a slide in fit without any resistance ... for now I'll
go with the AN174 but I'll monitor this area to see whether it
really needs to be drilled and reamed for an AN5 bolt.
Update: With both of the cyclic fore-aft control tubes in place,
it's easy to verify if there is any play in the system by firmly
holding one cyclic stick and then trying to rock the other one.
The above noted play was more obvious and is also compounded by
the slightly oversize pivot bolt hole in the cyclic sticks
themselves. I sincerely wish the factory had chosen to ream these
holes to size per the notes on the construction prints. There is
slightly more play on one of the sticks and that stick was moved
to the co-pilot side. One of my "someday" projects will
be to install precision bearings in the sticks to try remove this
slop.
My kit came with grips from Automatic Flagman and while they appear to be okay, they only have two push button switches on top plus a rocker trigger. The trigger switch is a DPDT ON-NONE-MOMENTARY type that does not appear to support a dual PTT intercom and radio function (i.e. 1/2 way is intercom, all the way is radio). The black push button switch is of the SPDT momentary style while the red switch is a SPDT alternating latch style (such as would be used for landing light etc.). Even if one of the push button switches was changed to a hat style, I'm having trouble figuring out how to add all the obvious functions plus possibly leave a spare. At this time I can see possibly having the following functions on the pilot's grip:
Update: I've decided to wire my sticks for the first set of the above functions. Unlike fixed wings, we normally don't call for the transponder code until we're actually ready to lift off. Once the code has been entered, we normally lift off nearly immediately and thus the transponder can be set active right away. My starter button will be on the collective which frees the right hand during startup. The other thing I considered was an air restart. What with the altitudes that we normally fly at, the primary concern in an engine-out would be to establish and maintain a good auto with very little time to consider a restart. The starter switch on the collective is still accessible but doesn't distract in any way from the cyclic which is being used for manoeuvering and speed control.
Although I haven't finalized my requirements, I'm thinking I will probably change the pilot side grip to something like an Infinity grip in order to add flexibility and extra switches. While the provided grip may be able to be modified to include all the required functions, I think ergonomically these switches would be grouped too close together and could lead to confusion. For some reason, I also find the supplied grip doesn't feel that comfortable for my hand; perhaps it's because I consider my hand size to be average (or even a little smaller) and these grips are STC'd for Ag planes, or just the fact that I'm trying them out before they're mounted on a real stick. It should also be noted that the Infinity grips are slightly smaller than the true military grips; according to their website, they're 1/2" shorter and a little smaller in the girth. I tried the Infinity grips while wearing my Nomex gloves and found that I definitely prefer the feel of the military grips.
Update: I managed to acquire a pair of new surplus grips from a CH-47 that are like the B8 grips I've seen advertised. These are definately very solid and comfortable grips with very high quality switches. For now, I'm thinking that I'll probably use the hat switch for radio flip-flop and channel select with the other switches used for transponder ident, landing light and cargo release. Although the trigger switch has two positions (first detent is intercom, all the way is radio PTT), I'm thinking that I'll wire it so that either postion will actually key the radio. Since the intercom position stays closed when the switch is further pressed to the PTT position, if the radio PTT is wired to the intercom switch then either position will key the radio. The switch for the landing light will need to be changed from a momentary style to an ON-OFF alternating style. Looking at the Otto Engineering data sheets, I believe that the Aircraft Spruce 11-00796[7/8] press-fit switches will work for this.
Like the CHR-supplied grips, these grips also required a sleeve to match them to the cyclic sticks. I used some 1.25" x .125" wall 6061-T6 tubing to make these sleeves, but they did require a bit of machining to get the ideal fit.
Now that I've had a little more time to think about this, I realize that these are GREAT grips but the co-pilot's grip is seldom used (especially if the stick is made removeable). It almost seems a shame to install both of these rather than just one on the pilot's side and a similar knock-off one on the passenger side. At least this is the kind of dilemma I don't mind since its like having too much high quality equipment.
I have heard second-hand about one ship that was using simple grips and had a cyclic stick shake (i.e. vibration) problem. This was cured by adding some lead to the top of the stick and apparently this extra weight dampened the shake to the point that it wasn't noticeable. Since my grips are heavier than the stock ones, I don't anticipate this being a problem.
Once the seats have been fabricated and everything is in place, the length of the cyclic sticks can be checked. I found that mine were slightly long in the as-delivered state ... the objective was to have my forearm gently resting on my leg. I ended up cutting off 1/2" which made them 17" from the pivot hole to the end. Note that this is dependant upon the grips, the seats and the pilot which means they should be tested and adjusted for each unique combination. Another way of looking at this is that I ended up with a measurement of ~16-1/2" from the pivot point to the grip's "wing" at the bottom of the grip. I tried sitting in another Safari which had it's stick cut about 1-1/2" shorter than this and I found it to be quite restrictive as my forearm was pressing relatively hard against my leg.
When drilling the hole to attach the grip, I found that it was most comfortable to have a slight twist in the grip for a comfortable neutral. However, I found that the rotation I selected was a lot less radical than some other ships that I've seen. My objective was just to keep my wrist relatively straight. While a more rotated postion may seem more comfortable or natural at first, I found that it also created a bit of an illusion when moving the stick. I also found that I had to notch both the stick and the spacer to proved extra clearance for the wires from the lowest button.
Bottom stick ready for installation, top one ready to have
the grip attached
One issue I found with my grips is that they were prewired with wire that was much heavier and larger diameter than required (M5086/2-18-9). Since I wanted to minimize the size of the wire hole in the stick, I ended up splicing these leads with normal tefzel wire and then routing them inside tygon tubing (1/4" ID x 3/8" OD). The other end of these wires is terminated in a D-sub connector which allows the sticks to be removed. In the above picture one can also see how I used Perma-Slick G on the the areas where the sticks pivot on their mount. I didn't want to build this area up with paint but I still wanted corrosion protection and minimal friction. A little bit of MoS2 grease on the pivot really makes for free movement.
Cyclic Trim
Most helicopters that I'm aware of have cyclic stick forces in flight that must be neutralized for minimum pilot fatigue. While many of these are pretty fancy and use hydraulics or electric servos, even the R22 has a fixed longitudinal spring and a lateral trim spring that can be engaged to cancel the left stick forces required during cruise.
The Safari kit and documentation do not have any references to a cyclic trim system, but I am aware of several builders who have opted to install some mechanism to accomplish this task. Unfortunately, as far as I know they have not chosen to provide a lot of detail or pictures. While I can see how these trims can be very worthwhile, especially to counter-act the right stick force, I don't have a solid design at this time that I'm willing to share. At least in the R22 one is pushing against the cyclic with their palm, while the force required in the Safari is towards the open part of the right hand. I do have a concept for an electric trim system and it will be noted that I've even pre-installed a hat switch to control it. It would have been nice to weld any required mounting tabs before painting, but that would have held up other assembly tasks.
I was talking to another Safari owner that temporarily installed a piece of surgical tubing to ease the cyclic stick forces on his injured shoulder. He was extremely impressed with how this worked and likened it to having hydraulic controls. Needless to say, I still plan on investigating this change after my basic construction is complete.
I am also aware of at least one builder who has experimented with changing the swashplate angle when the cyclic stick is in the neutral position in order to reduce the in-flight forces. Basically, he tilted the swashplate forward and to the pilot's side, but I'm not sure how much.
Collectives
I am not happy with the new design of the collective sticks where they're pre-greased at the factory with no zerc fittings and have the end caps fuse welded in place. I have several issues with this:
I discussed this with the factory and their response was to just build it as designed and fly it for a hundred hours or so before changing anything; after this interval one could remove the collectives and then make appropriate changes. I looked at this very carefully and realized this would be a painful undertaking since the only way I can see to physically remove the collectives after assembly is through the "firewall" area (it does just barely fit once the left and center interior panels are removed). However, after final assembly this would require the bulkhead ("firewall") to be removed which also implies that the engine, transmission, blades etc. would also all have to be pulled off first ... ouch! Also, one would have to be very careful about the fabrication of the pilot's side panel that encloses the collective as it may not be possible to remove this panel with the collective installed due to interference with the down stop. Alternatively, if one was just going to replace the grease with light oil then it MIGHT be possible to just grind off the fuse weld while the collectives are installed and then perform the cleanup.
Contrary to the factory recommendation, I am definitely changing these BEFORE final assembly. The primary reason for this is to try make the throttles as friction free as possible so the governor can move in very small increments rather than "twitch" the throttle. I've flown a ship with a twitchy governor and it was not a pleasant experience, especially during power changes when landing. I've ground off the weld on the end of the collective tubes which allows the collar to be removed and then allows the tubes to separate. There was a lot of grease between the tubes and it took a lot of passes with solvent to try remove it all. I'm going to investigate getting both surfaces coated with DryFilm, the same stuff that is used on internal engine parts and impregnates a teflon-like material into the metal. The other option might be Policoat, but I need to find out more information about the differences. When I'm talking to the coater, I'll also verify whether to add a couple of small holes in the underside of the outer tubes so that a few drops of light oil can be added if required.
Update: I've decided to paint the collectives and control system myself and to use Perma-Slik® G on all the bearing surfaces. This is a MoS2 based solid film lubricant in an epoxy binder that is available in bulk or in aerosol cans and as of this writing, one of the distributors for Everlube products is Aviall. It's my opinion that Perma-Slik® G is not nearly as tough or thin as the high temperature DryFilm coatings that I've seen from professional coaters. However, the advantage is that it can be applied at room temperature by a non-professional without the need for special equipment. As builders, after working on these parts for a long time we know exactly which areas have exterior surfaces, interior surfaces, hidden areas, bearing areas etc. and I've realized how hard it is to communicate with a painter exactly how I want things done.
White and black are paint ... the greyish areas are Perma-Slik® G
While the collectives were apart, I welded a filler plate in
the largest lightning hole in the throttle arm on the pilot's
side as this area will actually be used as the connect point for
the governor actuator linkage. In order to re-assemble the
collectives, I turned some aluminum plugs that fit inside the
collective tube and are tapped to receive three 4-40 countersunk
screws that go through both the original collars and the
collective tube. On the passenger side there is just a small lip
on the plug while on the pilot's side there's some extra meat
since I'm also adding a small
aluminum box on the end to allow for more switches; specifically
the starter, governor, trim and landing light switches. When
putting the collectives back together, I applied a thin
layer of Royal (Royco) 11MS grease (manufactured by Anderol) between the inner and
out tubes. There's definitely a lot less friction with this setup
than as received from the factory.
Update: If I were re-building the box on the end of the pilot's
collective, I would probably offset it to the left as far as
practical. There is room between it and the door and the offset
would allow more leg room.
Hindsight says that the landing light switch on the right is very close to the pilot's left leg
When installing the collective grips, I first used a belt sander to grind off most of the "bump" at the forward end and then used an X-Acto knife to cut out the remaining center part of the end. I wanted to locate the grips below the collective's end caps so they would not impart friction aginst the stationary part yet I still wanted them fairly far forward. I read how motorcycle folks used air, soapy water and even hairspray to install this kind of grip. In the end, I found that just the basic push and twist method worked fine for me. Once I've finalized a collective friction / locking device, I'll probably come back and put just a very little bit of rubber glue under the forward edge of the grips.
For shorter pilots, they may want to try create a seat mockup if they're taking the collectives apart. What they should be checking for is how well they can reach the collective when it's at both the lower and full down positions ... I think they'll find it a bit of a difficult reach, especially if they don't have retractors on their shoulder belts. I haven't really explored the options, but perhaps shortening the collective about an inch or so might help ... there should still be just enough room for the grip. It might also be viable to raise the downstop on the collectives ... again, I haven't fully explored how this might change the linkage, especially the pitch rod.
There has been recent discussions on other kit helicopter forums about the ability to remove the passenger's collective when a non-rated passenger is aboard. This is both as a safety item to prevent the passenger from bumping the center collective and also to allow more room for the pilot's knee/cyclic movement. I looked at this quickly and realize that since this is not allowed for in the basic design one will need to be very careful in the design modification to allow for both strength, reliability and ease of removal. While I will keep this in the back of my mind, I don't plan on doing this initially unless I can find a design that I'm very confident with.
Sheet #16 of the construction prints show a 1/4" hole centered at the back of the pilot's collective to allow passage of the starter button wire. In my kit, the stubs on the collective pivot tube supports were actually 2-1/2" long instead of 2" as shown on sheet #20. What this means is that if one is to center the hole for the wire based on the collective, the pivot tube support stub actually encroaches on the hole. If this were to have been assembled this way, there would definitely have been a pinching or rubbing on the starter wire. If one hasn't already drilled the hole, then it might be possible to offset the hole such that there isn't any interference. In my case, I drilled the hole before discovering the interference and then had to shorten the stub by about 1/8" plus.
I was debating how to paint the collective pivot tube since the the center support has a "T" tube at the top that can't be removed from the pivot tube. As I was looking closer at this, I realized that this support was actually rubbing against the pivot tube and was not in perfect alignment with the end pivot tube supports. I finally just split the center support "T" tube and removed it. I then cut the top of the "T" down and welded a flate plate onto the top. A split Delrin® support, as used on many fixed wing pedal supports, was then fabricated to capture the pivot tube.
Collective Trim
The supplied collective trim consists of a basic extension spring that applies down force on the collectives to prevent them from climbing. Both myself and several other builders have expressed concern over this very basic spring that is held in place with a cable and S-hooks since there is no backup or containment system in the event that the spring breaks, which has been known to happen. If my cross-reference is correct, this spring is loaded way past its design rating (perhaps 3+ times) and also has the following warning from the distributor: "Not for use in cycling applications". I'm now personally aware of THREE examples of this spring breaking in flight plus indirectly another THREE for a total of *SIX* breakages. The latest one was after only 7 hours of service. The immediate result is that the pilot must exert a *LOT* of downforce on the collective while trying to analyze what went wrong, maintain rotor RPM, work the throttle, etc. This can create a VERY DANGEROUS situation, especially for low time pilots ... I hate to think what would happen if this occurred while one was using their left hand for reading a map, setting a radio / transponder or was just close to the ground. I don't know about others, but I REFUSE to install this supplied spring system on my craft as I believe it is too dangerous for my objectives.
I started looking at alternatives and as a minimum I considered switching this system to a safety drawbar spring with a similar rating which acts on compression rather than extension. This would be a very simple switchover, but I'd also make sure that both ends of the system are fully captured. Alternatively, one could add a lighter safety drawbar spring on each end of the pitch tube and gain a level of redundancy. Perhaps others have already found a better solution which I'll gladly listen to.
Springs that were checked but were too light
I now think that I've come up with a viable solution to the trim issue that should improve both the safety and reliability. Instead of pulling the pitch tube up, this spring assembly pushes it up and the spring is being compressed as the collective is lifted. The force of this new assembly is easily adjustable and in the highly unlikely event of a breakage, the spring is contained. The primary negative is that it does require welding a piece onto the frame in a place that may be somewhat difficult to access on a completed ship. There is more information in the Frame section.
I first installed the CHR supplied spring and measured its
force both with the collective full down and full up. This was
repeated with the eyebolt length at its longest, shortest and mid-point
in order to get a range of forces that were available with the
stock setup. I then installed my modified spring and measured its
force both at the weakest setting and at about its mid-point of
adjustment. The adjustment range of available forces was quite
comparable to the stock setup. As a result, I fully intend to go
with my new setup.
Update: I'm quite happy with my setup. I did find that I had to
tighten my spring to about the extreme of what the stock system
would have been capable of. I understand that this is due in part
to the fact that I have the round tip blades instead of the newer
square tip blades. The result is that the collective will hold
it's position without any hand pressure on it ... the ideal
scenario.
It appears that there may be another alternative that totally eliminates the spring. In the Jan. 2006 issue of Experimental Helo magazine, there is an article by Stu Fields showing how he's experimenting with a centrifugal force method to fully balance the collective forces. This looks very encouraging. It must be noted that this is still in the experimental stage and doesn't have a lot of testing on it yet. However, this sheet will provide a good idea of how these changes were fabricated:
While I have never seen factory documentation about their solution to the spring issue, I have heard them discuss the use of dual springs to add a measure of redundancy. I've now seen a craft with factory installed dual springs and while it certainly appears better than the orignal single spring method, I'm not sure if it's an ideal solution.
Also note the piece of rubber glued to the firewall as
seal for the control rod slot
Correlator
The throttle and collective linkages in the kit are arranged such that there is a correlator that increases the throttle as collective is increased. By altering the moment arm on one of the linkages, one can alter the effect of this interaction with the objective that the correct amount of throttle should be automatically applied as the collective is raised or lowered. Once I've actually flight tested the ship and adjusted this linkage, I'll try to update this document with the moment arm that I end up with. Note that in the Frame section there is a discussion about a change to the seat supports to allow a greater range of adjustments.
I've received information from one builder that his correlator spacing was at about 3/4" and that only small adjustments were required from there. Another ship's correlator spacing was at 1-5/16" while a third ship used 1-1/4". What I don't know about these ships is the exact linkage configuration such as the carb arm hole used and whether they used the normal linkage arm lengths. It does appear that a significant spacer is required and that one should make sure that there is sufficient clearance on the right seat support to allow for this.
Unlike the R22, there is no throttle over-ride detent which bypasses the correlator. I have heard that this causes some difficulty in practicing certain things, like engine failure in a hover, since the correlator increases the throttle as the collective is being raised. Once I get my linkage re-assembled, I need to check this very carefully as I would think that one can continue to manually roll off the throttle to counter-act the correlator.
Control Friction / Lock
There is a regulation in Canada that control locks such as on the cyclic and collective cannot become engaged when the craft is being operated. Obviously this is also a common sense issue and while the construction notes say to just use friction on the pivot bolts, I don't believe this is a long-term reliable solution. While I intend to build control locks similar to those shown on page #46 of the Construction Prints, I also intend to install positive restraints for when they're dis-engaged. The cyclic should be simple since it just needs a bracket on the seat front panel to engage the locking pin. I haven't figured out the collective restraint yet, but it could probably be something as simple as a short bungee cord that prevents the locking tab moving from the open position.
I fabricated the cyclic control lock basically per the plans but it should be noted that one needs to measure for the individual craft rather than use the dimensions on the plans. The pivoting part of my yoke was only 5" wide (vesus 6" on the plans) and the distance to the stick was shorter than indicated on the prints. I added a positive down lock so this assembly could never interfere with the cyclic during flight. I've also added thin (.010") Delrin® washers between this part and the angle stock on seat front to save the paint and make it smoother to operate. Before drilling various holes, the stick should be carefully centered, preferably to the mast but to the frame could also work. If the cyclic lock is rigged such that the stick is held in this exact position then it can be engaged during the initial setup of the control rods and it makes their alignment much easier. Also note that the fore-aft position of the stick can be slightly altered by adding washers under the joining bolt head conditional that the retaining pin holes haven't been drilled yet ... I purposely made the pivoting part of the yoke about 1/8" shorter than expected so that washers would be required under the bolt head (easier than making a new bracket if its too long). The plans show a bent pin for a retainer but I chose to make the holes a little larger (3/16") so that they will accomodate a quick release pin (McMaster part # 92385A016) similar to what is used on the ballast weight. Note that it's better to defer drilling this hole until the pin is available in case one moves up to a 1/4" pin such as West Marine's C0375-0740-W.
The red part of this assembly is only visible from above when it
is engaged
Update: I think I'm going to regret fabricating and painting
this piece before completing the rigging process. Using the
supplied material and construction prints, full aft movement of
the stick hits the lock's arm when adjusting the cyclic per the
Rigging Manual. This becomes even worse with full aft and lateral
movement of the stick. There is also the issue of thick seat
cushions that can also cause an interference with stick movement.
More importantly, I find that the vertical neutral position of
the stick feels too far back for me and the full aft postion
really feels like the "stick in the gut" syndrome. I'm
not sure if this is due to my seat design or the fact that I'm 6'.
What this does mean is that I will seriously be looking at either
changing to a curved cyclic stick or at least cutting and welding
the sticks to move the top further forward at neutral. With the
existing lock, I can only modify the stick in the area above the
lock or else fabricate a whole new lock.
Update 2: While the stick may hit the sides of the lock in the
rearmost position, I don't think this is going to be a problem
with the stock sticks. The bigger problem is that the stick hits
my legs before it hits the edges of the lock bracket.
The more I've thought about the collective lock, the more I
like the R22 system which incorporates both a collective friction
adjustment and a control lock. I've added a frame tab to allow
for connecting this device and have obtained the R22 pieces.
Since the R22 collectives use an internal rotating tube whereas
the Safari uses an external rotating tube, it will be necessary
to add a floating collar to the Safari collectives in order to
provide an attachment point.
N.B. Although this is a combined friction and
lock device, the primary purpose is as a locking device. While it
will allow the pilot to add varying amounts of friction, I do not
believe that it is a wise thing to do for most flight operations.
If the collective will not stay at exactly the same setting by
itself (i.e. a rising or falling collective) then it is
indicative that the collective trim system needs adjustment. The
friction / locking device is not meant
as an alternative to this.
Update: I've fabricated the floating collar and tested the friction / locking mechanism. It works great, BUT ... In order to keep the frame tab as short as possible, I fabricated it so the sliding arm is on the outside and the lever would be on the inside. The problem is that there is only about 1/4" between the collective and the sheeting around the pilot's seat. I knew it would be tight, but it's too tight to get everything in place without interference. I've now come up with a plan to keep the collar that I built but move the lever to the outboard side.
Update 2: After a lot of fiddling, Plan B didn't work as well as I'd hoped and I'm now working on Plan C. Basically both the slider and the lever will have to be towards the outside of the craft in order to get sufficient clearance. The basic concept works extremely well (after all, it works on all Robbies), but the execution is lacking. Since I'm in a time bind for a collective lock, I'll fabricate a quick and dirty temporary one similar to that shown in the Construction Prints. This is on the "to do" list as something to come back to, but I needed it off the "critical" list.
In order to get the craft airworthy I had to install some kind of a collective lock and opted to make one similar to that shown on the Construction Prints. Basically a piece of .063" aluminum and some bending and cursing. Seems someone left the lid off our tool dip ... a bit of naptha helped to thin it for brushing but it's not a great result. As expected, it also rubbed the paint off the collective stick.
Although I haven't finalized the installation of the pylon support rods which will lock the transmission and swashplate into position, it appears that the pitch link rod actually rubs on the frame. This only occurs in some positions, but it appears that I will need to correct it. There are two possible solutions: bend the rod or notch the frame. Since this area of the frame has the 1/8" transmission support front plate behind it, I think it is already reinforced sufficiently if the frame is notched. With my transmission squaring brace in place, I heated the tube in this area and compressed the front edge with a large tube to get the clearance; probably about 1/4" required. In talking with factory personnel, this is a known problem. While their solution is to just let the rod rub against the frame, I don't consider this good practice and changed this area in order to gain clearance.
One of the things I'd noticed is that the aluminum tubing used for the control rods does have some flex to it. In the back of my mind I'd planned to further investigate this when all the rods to the swashplate were installed and was even thinking that perhaps I'd make up a set of steel rods to eliminate, or at least reduce, any flexing. I happened to be talking to CHR and they said this has been tried and there was no perceptible difference in the 2/rev vibrations. Since I was going with the straight control rods made with steel, I changed the pilot's fore-aft cyclic rod and the side-to-side interconnect rod to steel. Thus all the rods from the pilot's cyclic to the swashplate are made from steel but at the penalty of a couple of pounds (minus the weight of the removed walking beams and rod-ends so it's probably a wash).
Straight Control Rods
One interesting idea that was discussed is the use of straight
rods to the swashplate and the elimination of the walking beams
which have been shown to introduce some flexing into the linkage.
In an effort to reduce the 2/rev vibrations, at least one ship
has been outfitted with this change and I'm aware of at least one
other that is undergoing this change. Limited testing has shown a
significant (0.7+ ips) reduction in the 2/rev measurements on one
ship and not as drastic a change on others. Instead of the rod
going from the rear of the cyclics to the walking beams, these
straight rods are routed to go through the firewall and then
directly to the swashplate. I am now aware of one frame which is
being delivered from the factory with this change so it is
possible that this change is being formally implemented by CHR.
Update: I haven't been able to find anywhere on CHR's website
where they discuss this change, especially any form of updated
plans. However, an NTSB
report indicates that they have redesigned and incorporated
this change. I'm also aware that several kits have been shipped
with this change included. Of more relevance is the fact that the
transmission side mounting arrangemant has also been changed on
these kits. Once again, this clearly
points out the lack of communication and documentation related to
this kit.
This modification requires at least four changes to ships that use walking beams:
It also appears that the rod ends on the bellcranks at the end of the cyclic tube may be SLIGHTLY rubbing on the rear upper edge of the arms since they are at a relatively steep angle. I need to look at this a more closely and decide whether the arms just need a bit of rounding or whether to slightly bend the arm and place a washer between the rod end and the bellcrank arm.
These changes are much easier to implement during construction and before paint. While I may initially install the walking beams, I'm making the changes for the installation of straight rods. This way I have a choice that only requires the physical installation of the appropriate rods without having to do any fitting, cutting, painting etc. Anyone considering this change should contact the factory and discuss the details.
Note that any dimensions in the following description are just what worked for me and may be different on each craft ... they need to be double checked!
As I have received some questions about how I performed these changes, I thought I'd add a little more detail. Specifically, the biggest concern is the cutting of the transmission support frame member:
With the transmission mounted, I ran a piece of fishing wire from the swashplate directly to the cyclic bellcrank. Although it touches the frame, by moving the swashplate through its range I was able to identify the location and width of the interference. Both of the tubes to be cut are ~7 1/2" wide at the center looking forward from the rear side. The actual edges of my cuts are angled slightly to follow the angle of the control rods and in my case they're:
The actual depth of these notches seems to be dependant upon the ship, but I found that the passenger's side needed to be just about full tube depth and the pilot's side is just slightly more than full tube depth. Before cutting the tubes, I chose to add an angle iron brace to try prevent any movement during welding. It doesn't show in the picture, but there is a 1/8" spacer at the intersection since the angle iron is above/below mounted.
I first made a circular cut in the tube with a cutoff wheel but left about 1/4"+ at the front that was still connected. Approximately 2/3 of the rear part of the tube was then cut and removed. The remaining part of the tube was then heated and splayed such that it created a vertical plate that became the inside of the notch. The reinforcement piece was shaped from a piece of 3/4" x 1-1/2" x .049 4130 which was cut to make a "U" shape (~3/4" x 1") and also to fit over the vertical plate that remained from the original tube. This was welded in place along with four end caps.
The notches in the shroud were first identified using the fishing wire technique and opened up by trial and error. Final size was approximately:
The biggest pain with the shroud modification was making the filler pieces that seal the notch in the top and the slots in the front panel. I first tried making these out of .032" 2024-T3 aluminum that I had, but couldn't easily get the required reversing bends. I then tried .025" 6061-T6 and these worked much better as the bends can easily be "tweaked" by hand. I found it was hard to get the angles for a full seal at the top and ended up using a thin layer of high-temperature RTV to seal all the gaps once the various panels had been painted. This was a lot easier than trying to reform new fillers multiple times.
The slots in the firewall were first identified by measuring from the wire to various mounting points with the firewall removed. On my ship, the rough center of the slot was located from the edges of the firewall sheet as follows:
A relatively small slot was first cut and then progressively opened up by trial and error as the swashplate was moved through its full range. In the photo below, the slots have full rod clearance but are still very rough as they were being opened up with a nibbler while the firewall was installed. After full clearance was obtained and the firewall was removed, they were opened up a bit more to provide extra clearance and to smooth out the slots. Also note that there is a slight perspective error since the photo was taken at an angle from slightly above the console support rods.
For the actual control rods, I purchased a couple of four foot lengths of 5/8" x .065" 4130 steel tubing along with a foot of 1/2" x .083 tubing. The smaller tube is to be used as welded end inserts and the control rods will be built the same way as the CHR supplied pitch change rod. Note that the 1/2" tubing will have to be SLIGHTLY turned down (to .492" or so) in a lathe before being able to be used as inserts.
The final step before installation of this modification is to remove the walking beams. Since I wanted to leave the option of either walking beams or the straight rods, I left the frame attachment brackets for the walking beam rod ends in place. The holes in the top "firewall" rib should also be capped off. I chose to first enlarge the slots in the ribs to allow for passage of the walking beam ends after installation and then added filler plates that were mounted with screws and nutplates. In the event that one decides to only go with the straight rods with no walking beam option, the walking beam mounts could be cut off which will save a bit of weight. If the forward walking beam mount is removed, this makes it much easier to install a centered inertia reel which can be used with a "Y"-belt shoulder harness.
Speaking of walking beams ... The walking beams as shown on Construction Print #22 do not have equal length arms (3" & 3-1/2"). Although it would be relatively easy to accommodate this ratio in the bellcranks shown on Construction Print #16, I have never seen any reference to this change. While either configuration (walking beams vs. straight rods) appears to work, I would assume that there is a slight difference in the how the craft responds to collective inputs. This is due to the different amounts of throw on the side control rods to the swashplate versus the main pitch rod when the collective is raised or lowered.
Tip: The pitch change rod can be used to perform much of the initial alignment for the various cuts. By screwing the rod ends most of the way out, its length is very close to what is required for the new rods.
Tail Rotor Control Cable
Like all the newer kits, I have a push-pull cable to control
the tail rotor. One of the difficulties was deciding where to
actually exit the cable through the firewall ... the plans didn't
really help. I wanted it to exit within the fixed panels so that
any future removal of the firewall center section would not
require the removal of the cable. The most obvious places were
either above or below the lowest frame member on the right side.
I looked at all the pictures I have and it appears that most
builders opted for above the lowest frame member and that's what
I chose. My concern with going to the lower position was that it
might require a slightly longer cable run and the cable might be
a bit too short ... such is the joy of making decisions without
the tail boom attached.
Update: With hindsight, I would have changed this exit hole to be
just inside the vertical frame tube for the firewall (about even
with the screw in the following picture) ... in fact, I'll
probably make this change. The cable is not overly long and this
will allow it to follow the tubes of the tail boom a bit more
closely. Additionally, it provides a minor benefit by moving the
cable a bit more inboard and provides a bit more protection for
it. The downside is that I mounted my strobe power supply in this
general area and it looks like the optimum position for me will
be right on the seam where the center firewall section meets the
ribs.
Update 2: I decided to go ahead and move the cable's exit position. Although the new hole is right on the firewall seam, I used the supplied grommet rather than the cable exit in the above picture. The grommet actually stretched quite well to cover the dual/single panels and it should still allow for firewall removal without having to first remove the cable. I'm using adel clamps both on the inside (about 1/2 way between the seat front panel and the firewall) and just after the cable exits the firewall ... these should allow for minimal movement of the cable within the grommet. The real big advantage is this small movement of the exit hole (probably just under 2") allows the fixed length cable to much more closely follow the boom contour.
Testing the new cable exit position
As noted in both the Controls and Rigging sections, I had issues with both the rear and forward mounts for the cable. The rear mount needed to be bent and just allowed the cable to be attached with minimum threads on the rod end. The front mount was welded to the frame too far back from the pedals and required an extension for the cable.
When I went to mount the adel clamps for the cable I ran into another issue. The kit supplies two #6 clamps for this and I had made sure that I had plenty of extras as I planned to use a lot more of them than what the plans indicate. However, I found that both the supplied DG6 clamaps and WDG6 clamps that I also had were prone to sliding on the cable. Upon closer checking, it would appear that the high point on the cable's sheath is 11/32" which puts it right between a #5 and #6 clamp. I chose to go with #5 clamps as the purpose is to lock the cable in position rather than just act as a guide. Ironically, the supplied grommets for the cable were only 1/4" ID.
When clamping the cable in place, I found it was easiest to start at both ends and work towards the front part of the boom where the cable tends to bow. The front of the cable was progressively clamped up to the end of the main frame (i.e. rear landing gear socket) and the back part was clamped from the tail rotor gearbox up to where the cable can no longer follow the straight path. It is then pretty obvious how much of an offset one has for the cable and where is the best place to put the curve. Interestingly, I only needed one spacer for the clamps and this had the holes for the adel clamps just under 1" apart ... perhaps this is due to the spacer at the front mount which effectively makes the cable a bit longer. The rest of the clamping was done with two adel clamps back to back ... the rearmost one used a bracket to allow the clamps to mount closer together. One problem area was around the vertical stabilizer where it needed to be notched in order to clear the clamp.
There seems to be a bit of a conflict on the information about the hold downs for the cable. I have heard several sources state that it is imperative to get the cable as well supported and rigid as possible in order to avoid any play or spongy feeling to the linkage. I've also seen a factory assembled craft that just had a few clamps along the curved area and just a few tie wraps to the boom. I chose to try make my cable quite rigid and ended up with 17 clamps plus two grommets ... possibly a bit of overkill.
One of the things to be careful of with the cable routing is how close it gets to the exhaust. I am aware of one case where the outer sheath of the cable seemed fine but it appears that there was some heat damage to the inner part of the cable and it caused some sticking. Since my engine is running very rich, I could see the exhaust residue and it's very obvious that it's going both further back than I expected plus a bit upwards. I've been told that a heat deflector is a wise addition if there is any question about heat from the exhaust on the cable.
When inspecting the push-pull cable prior to installation, I did a bit of playing around with it while it was still coiled. One of the things that I found most disconcerting is the apparent slop in the cable. When pulling on an end, there is an almost immediate transfer of the motion to the other end. However, when pushing on the cable there was about 1/2" of movement on the end being pushed before I could detect any motion at the other end!!! When looking at the specifications for the cable it is interesting to note that the cable's normal load is listed as 120 lb in tension and only 25 lb in compression. I'm assuming that the only reason the cable works in this application is because the normal anti-torque pedal pressure results in tension on the cable. However, it makes me seriously wonder about the effect in gusty conditions where one might need to "dance" on the pedals. Besides the alternative of going to dual cables for a pull-pull setup, I intend to keep an eye out for alternative cables such as as those from Teleflex. The CompAir piston planes use what appears to be Teleflex cables for flap actuation and while they do exhibit the same symptoms, I noticed considerably less slop due to compression (the cable was shorter than the one supplied in the Safari kit).
One of the things that the above points out is the need to carefully and fully secure the push-pull cable. I believe that any movement of it will cause either a slight control position change or at least introduce some slop on the next movement. I intend to use a lot of adel clamps to firmly hold it in position. While zip ties are quick to install, it is hard to get a real tight movement free connection with them. Furthermore, if the cable is just attached directly to the boom with zip ties it creates a lot of areas where any dust or dirt between the two surfaces will result in abrasion.
Note that there are rod ends on both the push-pull cable and also within the tail rotor swash assembly. These should be checked for freedom of movement just like all the other rod ends in the craft. I am aware of one tail rotor assembly that was received with stiff rod ends ... spinning the rod ends resulted in smoother operation. All of mine (2 for the cable and 8 on the tail rotor assembly) were extremely stiff and required a LOT of pressure to move them before they were loosened up.
When installing the cable, I found it necessary to remove the black dust boots. After a bit of head scratching, I used a piece of safety wire slipped between the boot and the rod to tease the boot onto/over the "bump". Putting the boot back on should not be a problem.
Rod Ends
The rod ends that I received in my kit were the kind that have a PTFE liner bonded to the race. The one thing I noticed is that there is a significant difference in friction among the various rod ends and most were quite stiff. Per the instructions, this can be cured by mounting a bolt in the rod end bearing, then chucking it in a drill and spinning it for a bit followed by water quenching. Essentially the PTFE begins to melt/shred and slightly opens the race. However, there are a couple of tricks that I spent way too long learning the hard way. The most important one is to use a drill that has both a relatively fast speed and LOTS of torque. I first tried my air drill and it sort of worked on some of the rod ends, but only about a quarter of them. I then tried a couple of battery powered electric drills and they couldn't get enough speed to cause enough friction heat. The best drill I tried was an 18V Snap-On battery drill that had both the right speed and plenty of torque with a charged battery. I used a shop bolt about 1-1/4" long that was mostly threaded, and I ground three flats on the last 1/2" or so for the chuck jaws. The rod end was then installed with a nut and the drill run in reverse to keep the nut from backing off. The threads of the rod end were placed in a vice with soft jaws and it was time to spin it up. The following sequence was observed:
- The ball first spun at high speed
- It then began to slow nearly to a stop as it heated up and the friction increased
- It then began to speed up again
- There was usually a bit of smoke at this stage
- One could see some of the tan-coloured liner exiting the slot
- It was time to quench it just as or slightly after the smoke / liner material appeared
- After quenching, some of the shredded liner was visible as hair-like pieces ... sort of looks like wet tobacco
Although I worried about ruining a whole bunch of rod ends, this technique appeared to work quite well on pretty well all of the rod ends that required it (i.e. nearly all of them). As far as I can tell, there is no apparent slop that has been introduced yet they are extremely free in rotation.
One thing I've noticed with both the supplied hardware, and on completed ships, is that the rod ends are just inserted into a tapped tube with a single stop nut to secure them in place. Throughout the rest of the kit and general aviation practice, threaded fasteners are always held by two devices such as a nut + cotterpin, tension + lockwire, tension + lock washer, tension + elastic insert nut, etc. While I quickly looked in AC43.13 and couldn't find a direct reference, the use of a threaded rod end plus stop nut would appear to be just a single tension retainer. Those familiar with the R22 would have noted that their rod ends all have the stop nut PLUS a pal nut which acts as a locking jam nut.
Since all the rod ends have a default right thread, even if the stop nuts on both ends of a control rod loosened off then the length would remain relatively constant UNLESS one of the rod ends came out of the tapped tube. If the rod ends are inserted as deeply into the tapped tubes as they should be, this should not occur. HOWEVER, there could be a slight variation in length, a slight slop due to the threads not being full depth and a rubbing on the sides of the rod end cage if the stop nut loosened. Since many of these rod ends are under the seats and not part of the pre-flight inspection, I'm adding AN356 pal nuts to at least one rod end on each tube similar to the way Robinson does it. Torque seal will also be used for visual inspection.
Since I was working on the modifications for the straight control rods, I figured this would be a good time to finally mount the engine and shroud. The intent was to make the required cutouts in the shroud and then re-install the transmission to verify their position and size. While the engine was installed I was going to drill/mount the oil cooler and work on the oil cooler duct. I had deferred working on the oil cooler duct when I [re]built the shroud as this piece is aligned both with the main shroud and the oil cooler which is attached to the frame.
After making the rough cutouts in the shroud it was time to remount the transmission; but first the clutch drum and fan had to be installed. I knew this was a tight fit to the crankshaft flange and bushings and had delayed trying to tighten it down the last 1/8". This had worked fine when assembling the shroud top, but now it was time to actually tighten it down so the transmission would fit properly. I tried all six rotational positions and gentle (plus a bit more) pressure to try seat the clutch drum; all with no joy. My next thought was to install the six prop bolts to verify their alignment and possibly use them to help snug down the clutch drum. SURPRISE ... even though the transmission box was clearly labelled O-360, it contained 3/8" bolts for an O-320 and not the required 1/2" bolts!
As the blue air cleared, I tried to understand what was preventing the clutch drum from seating. Eventually I got out the micrometer and the answer was pretty obvious. The center of the clutch drum slides over the snout of the crank and it is a relatively close tolerance, but a slide-in, fit; about .005" or less clearance as I remember. The holes in the clutch drum for the prop bolts have a relief on the backside to clear the threaded bushings on the prop flange; in my case, four of the six bushings are of the protruding style and they stick up about 3/16" above the face of the starter ring gear support (flywheel). These relief holes are also relatively close tolerance to the bushing diameter; less than .005" clearance as I remember. HOWEVER ... when I measured from the center bore to the edge of the relief holes there was an inconsistent spacing. In fact, there is a .012" variation!!!
Further checking of measurements has yielded a hypothesis. It would appear that the clutch drum was first milled / bored from the side with the lining material and also included the center bore and the prop bolt holes. The piece was then turned over and milled for the recess where the fan is attached and also the reliefs for the prop flange bushings. It would appear that the piece was offset .006" during the flipping since both the fan flange and bushing relief holes are offset a similar amount. Without the proper bolts, I don't know if the actual prop bolt holes will line up accurately with the prop flange, but preliminary measurements appear that it will. It would also appear that the stainless steel fan has been offset, but accurate measurement is beyond the tools that I had access to. I would assume that this assembly is definately out of balance.
Update: The clutch drum/fan was included with the shipment of the transmission to Florida. Although I did not receive any written or verbal feedback on the original clutch drum/fan, it was replaced with a different (new?) one when the transmission was returned. Unfortunately, the replacement clutch drum/fan is actually causing me MORE trouble than the original one!!! The new clutch drum will not slide over the snout on the engine's prop flange, let alone worrying about whether it will slide over the prop flange bushings. Preliminary quick measurements leave a lot of doubt in my mind whether it will. The new fan has a much rougher edge surface from the cutting process. It would also appear that a slightly different material may have been used for the clutch surface. On a positive note, the six "prop" bolts were replaced and I now have the correct bolts to attach it, if I can ever get it to fit.
Update 2: I have been in contact with CHR about the "new" clutch drum. It would appear that they have decided to change the machining of these such that they are now an interference fit on the crankshaft snout and it requires heating in order to make them fit. One of the requirements for this "new" style is the use of three jacking bolts that can be threaded into the drum and press against the starter ring gear support (aka flywheel) in order to facilitate removal. According to my measurements, the fan side of the drum is STILL off-center by .006" on this second unit and will require the reliefs for the prop driver bushings to be opened up. I don't know about other builders, but my preference would still be for a precision machined drum that just slid over the snout / bushings ( <= .005" clearance on all holes) without the need for a heated interference fit, jacking bolts, the need to open up the reliefs and the potential imbalance issues.
As part of the investigation, we mounted the clutch drum/fan on an unmounted PSRU so we could spin the unit while seeing if the fan was centered. It was very obvious that the fan is off-center. I would HIGHLY recommend that all builders plan on performing a dynamic balance of their engine with the clutch drum/fan installed, but no transmission. Although my engine has already been balanced, it would appear that the clutch drum/fan will now throw it out of balance and contribute to vibrations. I had thought about sending out the clutch drum/fan to be balanced by a propeller shop, but with the "new" fit it would now appear that this is not practical.
Update 3: Although I still believe the drum should be machined more accurately, this was holding up my construction progress. Since I'd already determined which side of the relief holes were off-center, I chose to open up the one side of four holes to allow clearance. This was done by GENTLY sanding them with a small drum and using the micrometer to gauge progress followed by a quick polish with a small scotchbrite drum. After a couple of trial fits and lightly snugging the prop bolts, it was apparent that center bore of the drum was slightly rubbing / galling on one side. After smoothing this out, the clutch drum is now a very tight but non-interference fit onto the engine. The clutch drum and fan will be dynamically balanced in-place once the engine is ready to be run.
Update 4: It was still bothering me to know that the heavy and large fan / clutch drum was out of round and probably out of balance. Even with a plan to do a dynamic balance, there would be a period of time running the engine where there might be a significant vibration. I finally took the fan / clutch to the local propellor shop which had the right equipment to do a proper static balance on it. I was quite surprised to find that it only required one AN960-10L washer to achieve static balance and it was located about where I'd predicted. Regardless, I'm glad that I took this step.
Installation note: The manual talks about installing the engine and snugging up the bolts in the conical mounts to align the engine to the transmission. While this is obviously logical to avoid excessive side loads on the bearing that extends into the clutch drum, it is not totally obvious how to do this. Once the transmission has been mounted and the support rods drilled, they form a relatively rigid unit with the frame without any adjustment options. Therefore any alignment issues during re-installation must be done via the engine mounting bolts and compression of the conical mounts. Once the engine is installed, if the the transmission and support rods are installed without the clutch drum, then it is relatively easy to align the engine as the lower "Glory Nut" is visible and actually within the crankshaft snout. Any misalignment is pretty obvious and can be corrected before the clutch drum is installed.
Since I was just about to paint the frame, I decided that I'd first investigate and make the changes for the straight control rods with the idea that this would give me the flexibility to either use the straight control rods or the walking beams without any significant changes. Originally I'd planned to delay the engine and transmission mounting until after the frame had been painted and thus I could leave them in place and possibly save at least one removal and installation cycle. Since the transmission is much easier to install than the engine, I decided to first install it without the engine and make the changes to the firewall and the support tubes; if all went well then I'd mount the engine and make the required changes to the shroud. Time to finally remove the transmission from its crate.
Installing the transmission and mast without the engine is pretty easy and was accomplished with a couple of people rather than any kind of hoist. I did notice at least one place on the rear support where I'll do a little grinding to gain a bit more clearance. Installing temporary bolts showed a very tight fit, but not unreasonable. Measuring and marking the firewall and support tubes for their clearance cuts was also pretty straight forward. I strung a piece of nylon coated fishing wire, instead of a control rod, from the swashplate to the arm at the rear end of the cyclic tube. This line just touches the transmission support tube and by moving the swashplate and cyclic through their full range of motion I could mark the exact location and width of the interference. Measurements from this line also gave the initial location for the slot through the firewall.
Since everything was easily accessible, I decided to do an initial trial fit of the four pylon support rods. Accurately drilling these holes is not a step that I was looking forward to, but I figured that I could get a good handle on it by seeing how they looked when positioned in place. This revealed two problems:
For anyone who wants to check whether they have a similar rotated pylon, its relatively easy and obvious with the transmission mounted on the frame. If a long straight-edge is placed along the side edge of the pylon, it should be square with the frame in the fore-aft plane. If the straight-edge is placed along the front or back edge of the pylon, it should be parallel with the firewall. Also, the lower swashplate attach points that connect to the rods that go to the end of the cyclic tubes should be parallel to the firewall. These discrepencies are actually quite visually noticeable without tools when you know what to look for.
I must admit that CHR has been very willing to try rectify this issue once they acknowledged what the problem was, since they have seen it before. However, this is the kind of issue that I sincerely believe should have a service bulletin issued. This misalignment can cause a lot of frustration for the first-time builder, but more importantly, will result in a misalignment of the swashplate unless corrected. It's my understanding that various components are manufactured in batches which implies that there may be other transmissions / masts that were shipped with the same problem. There are several potential solutions to my situation:
Update: The transmission was shipped to Florida for a warranty repair. I had no desire to re-install new pylon support rods at a later date and I was already having enough trouble trying to get them close to lining up. It turned out that the upper casting for the main transmission had been drilled incorrectly for the four bolts that attach the mast tube. I know it took a lot of effort and shipping costs on CHR's part, but the upper casting was replaced. While all this effort is highly appreciated and it now appears to have cured the rotational alignment issues, I am worried that the main shaft now has detectable scratch marks on it just above the swash driver. I assume this was from the multiple removals/installations of the upper bearing. Also the gear train appears to be much "stiffer" than it was initially.
The four pylon support rods were shipped back to Ear Falls to check both their alignment (i.e. twist) on their jig and because the forks in the large ends were too narrow to slide over the pylon on the mast. When the rods were returned, it appears that the forks have been opened up some and do in fact slide over the pylon now. There is still some question whether they will slide over deeply enough, but I'll have to wait until I'm ready to actually install and drill them.
Status: I've tried again to align the engine, transmission and pylon support rods. The more I tried to get the engine and transmission aligned and measured, the more it became obvious that I'll probably have to use shims both under the engine and on the transmission mounts. While I can get some engine movement by snugging the bolts as described in the manual, I'm not comfortable with really overtightening the conical bushings. It would appear that a starting point for shims will be at least 1/16" under the front left mount and 1/32" under the rear left and front right mounts. I'm not sure how much the transmission needs to be shimmed upwards, but it appears that the right side needs at least 1/16" more than the left side.
This whole shimming and measuring process is time consuming and it seems that just as I thought I got it, I found a different direction/angle to measure which then turned out to have a discrepency. Originally my idea was to get this whole shimming issue and the pylon support rods drilled before painting and final installation. However, this is slowing things down and I'm not sure if it will still be accurate after painting and re-assembly. I finally decided to skip this process until after everything has been painted and its being put together for the final fit and assembly. Perhaps this will cause some extra effort repairing paint chips, but I think that's the best sequence and method.
Update: Now that the frame is painted and the firewall is permanently attached, I'm trying to align the transmission and pylon rods again. This is turning out to be almost as painful as the first and second times I tried to do it. The transmission does not align perfectly with the mounting holes in the frame and even though I used bracing, this may have been due to my welding on the mounts for the straight control rods. The pylon struts still have problems even though they were sent back to CHR to be double-checked on their jig. One of the aft struts is 5/16" longer than the other one and causing a lot of problems. The options were to mill a lot of material out of the fork, cut and re-weld it shorter or just make a new strut. I finally contacted CHR and they agreed to fabricate and send out a new strut based on my measurements. I'll openly give credit where it is due and say that this was done in a very pleasant manner and with extremely quick turnaround (1/2 day) ... thank you.
It would appear that CHR's welding jig does not make any allowance for paint on a completed craft and the lower forks on these rods were bent and welded for exactly 1/8" or just slightly under. Luckily, I found that I had some 1/8" rough 4130 that was actually about .140" and these forks could be driven onto it to slightly expand them. All of the upper forks were also slightly too narrow for the spider on the mast. A piece of 1/4" sheet helped in the opening and bending of these forks but a couple of them still had problems at the deepest part of the forks since they have to be mounted so far onto the spider. I finally had to use the edge of a cutoff wheel and a file to slightly grind these spots.
Final Installation:
After re-reading the manual and talking with CHR personnel, it
appears that the recommended installation method is to mount the
engine then just torque the conical bushings to level the engine
in both planes The transmission is then bolted in place and the
pylon rods installed so the transmission lines up with the engine.
I have a lot of problems with this for multiple reasons:
- This does not align the mast with the frame and there can be unwanted offsets. In my case this could have been something like 0.6 degrees to the right and about 0.4 degrees to the the rear.
- There is no reference to making sure that the clutch shoes on the transmission are parallel to the clutch drum. In my case, the recommended procedure resulted in a very noticeable mis-alignment that I suspect would cause rapid wear whereas I spent the time and effort to get the two halves of the clutch parallel in both planes.
Final position (should be checked in both planes)
- The conical bushings are designed as vibration isolators and the only way they can properly do their job is if they're equally and properly torqued. Since they're quite stiff, it takes a large torque differential to make a noticeable difference to the engine's angle and the unequal torque is potentially going to introduce a new vibration.
- I ended up using .063" shims under the two front mounts of my engine. Without them, it would have required about 2-1/2 more turns of the mounting nuts or about 25% more compression on two conical bushings.
- If there has been any heat warping of the frame during welding (some is guaranteed), one doesn't know whether this is in the engine mount, the transmission mount or some other area. Note that during welding, the factory jig aligns the transmission mounts to the engine mounts. What happens if the engine mounts aren't perfectly square to the frame to start with?
Interestingly, the construction manual which talks about welding up the transmission mount also states that the main rotor shaft should be parallel to the frame in both planes. Since my transmission mounts were already welded in place, the new recommended procedure makes no reference to main rotor shaft alignment. Since I consider the mast alignment to be a critical apsect of construction, I chose to try align things a little differently:
This sounds like a relatively straightforward procedure, but it caused me no end of grief. It seemed like everytime I would make a change and go back to check the results, there would be a totally unexplained movement somewhere else. I know that part of the issue is dealing with a heavy motor on "rubber" bushings, but there were still some very unpredictable changes. I'm also aware that once the engine is running, there will be some movement as it finds it's natural resting place. After spending several build sessions just to set the engine and drill the four holes in the pylon rods / pylon, I've come to a simple conclusion ... it would be impossible on my craft to have the mast perfectly aligned and simultaneously have the clutch drum perfectly aligned with the bottom of the transmission with relatively equal torques on the conical bushings. I can accomplish this on the lateral plane, but can only come close in the fore/aft plane. The only way this could be done is if I moved the mounting holes for the transmission. I suspect there might be a very slight alignment error in the factory jig and/or they didn't normalize this area after TIG welding. Because the engine bolt holes and the transmission mount holes are pre-drilled, they must be perfectly aligned in order to align the rest of the drivetrain.
My final configuration was to use a .063" shim under each of the front conical bushings, a .080" shim under the rear transmission mount and .125" shims under the right mounts. Although these are noticeable if one looks at them, I really don't care since the objective is for accuracy and functionality. There is roughly a one turn difference on the front-rear engine bolts since the .063" shim wasn't quite perfect ... probably .080" would have been better but this is time consuming and painful process to make new shims and then remove / re-install the engine bolts.
I ran into a second major issue when I was trying to do the final drilling for the pylon rods. I had finally gotten the mast into acceptable alignment in addition to the clutch drum area. The transmission was lifted, the pylon rods were inserted and the transmission was re-bolted to it's mounts. The mast angle changed by 0.3° ... $%^@#& !!! After a lot of trial and error, it seemed that the two front pylon rods had a slightly wrong twist in them and were binding just before the final seating position. This was corrected and re-tested ... there was still an alignment problem. Turns out the front left pylon rod was just bottoming out when the transmission bolts were tightened ... grrrrr! Since there was only a slight interference, a cutoff wheel was used to slightly grind the welded tube end at the bottom of the fork.
Once the craft is running, I may need to come back and re-visit this area. Without the shim under the right transmission mount, the mast would be angled to the right. With the rotation direction of our blades, this is the correct direction to try reduce translating tendancy or tail rotor drift. It will also interact with the rolling tendancy. The downside of any changes in this area is that it will require new pylon rods to be built ... perhaps adjustable ones could be used while trying to sort all of this out. I am also aware that some owners have purposely changed the swashplate angle when the stick is neutral so as to reduce the in-flight forces.
Note that the instructions talk about lifting the transmission / mast in place and/or using a hoist to move it. I would HIGHLY recommend that the builder figure out some way to attach the transmission to a hoist. During the final alignment stage and drilling the pylon rods, the transmission / mast will be repeatedly lifted and lowered an inch or two. This would be a very painful process if it was done manually and I'm sure that something could easily get bumped out of alignment.
Note three primed pylon rods and the one new rod
Tip: When trying to get this alignment finalized, I found it was easiest to remove the clutch shoes as they just get in the way and can potentially spring out (be careful to keep track of the actual springs). Note that each of the shoes has markings that can be recorded for later re-assembly. I also found it easier to do the initial engine to transmission alignment with the clutch drum removed which allows one to see and measure the transmission's lower bearing alignment to the crankshaft snout. My thoughts were that if I can get this alignment very close then there won't be excessive side loads on the transmission's bearing which fits into the clutch drum.
One thing I'd caution readers about is to not make the same mistake that I did. When I did the final install of the transmission I used a gang of the large SCEET tube clamps to hold the shoes in position so everything could be lowered into place. HOWEVER, when tightening the band of clamps I only tightened one of the screws which caused an uneven pressure on the shoes and springs. The result is that one of the springs cocked over and although there was a slight movement of the shoe it was not consistent with the others. During final rigging and checkout we had to lift the blade and transmission assembly in order to replace the spring ... basically a very time consuming task just to get at the spring.
The cocked over spring after removal
One of the things that we did when re-installing the shoes and springs was to file a spring width notch into the shoes where the springs rest. This helps to keep the springs centered and also helps to prevent them from cocking over during installation.
Tip 2: It is absolutely crucial that the pylon rods are not binding in any way when they are in their final position and being aligned for drilling. At most, a very slight pressure should allow them to move a bit on the pylon. Here are some of the issues I had to deal with:
Tip 3: The deepest part of the upper forks in the pylon rods have a gap on each side of the weld on the end of actual rod that's about 1/16" wide on each side. It is nearly impossible to spray paint into this area without getting runs on the end of the forks and thus there may be unprotected steel in these areas. One solution might be to apply a bit of JB-Weld or other filler into this area after the rods have been fully adjusted and just before painting them. Of course I didn't think of this till I was actually spraying the rods and it was too late. I used a small brush to get some more paint into this area to try get full coverage.
One issue we found during the final rigging and balancing is that my pylon did not have a Heli-Coil® in it for the balancer's accelerometer. Although the hole was properly tapped for it, this step had been omitted. Luckily I had already obtained a kit of these and had one available. This can be easily checked by seeing whether an AN4 bolt will thread into this hole.
The transmission has a vent line that exits from the top and according to the instructions it is looped up and then down ... sort of an inverted U shape. In conversation with an owner of a running craft, his transmission finds a natural level that is approximately 1/3 of the way up the sight glass. I am aware that there is actually an oil mist within a running transmission and that an oil separator has been used on at least one Safari to try reduce oil streaking. I've noted that the Kiwi ships appear to be using a longer vertical line that ends at the fuel tank cross-brace and appears to have some kind of funnel / cap attached to it which acts as the fill point. I think I'll try that approach and see how it works as it should also make the re-filling process quite simple.
The combined vent/filler in the above photo does work but there is a limitation. I forgot that there is a fairly heavy oil being introduced to the system via a relatively small hose. If one tries to rapidly fill the system (i.e. fill the cup) there will be an air bubble that needs to escape and it can be messy. This can be overcome by just filling the system relatively slowly. The advantage of this system is that there is just a simple dzus fastener to undo and no need to remove adel clamps when filling the transmission. What's not obvious in the above picture is that there is a lip under the cap and there are two small notches in it which act as the vents.
Another owner has chosen to add a small aluminum expansion chamber (~ 1-3/4" x 4") that has pipe threads top and bottom. The bottom is threaded into the transmission vent hole (AN911 fitting) and the top end has the stock hose fitting ... it looks to me like the builder used AN867-2 weldable flanges (available from Wicks) for the end caps. The hose is routed up and then looped down where it eventually terminates below the cab. He reports that he has never seen any oil drips out of the bottom of the hose.
The lower drain hose is exactly that ... a long hose with a
supplied plug held in place with a band clamp which are both
removed to drain the oil. This is somewhat inconvenient and a
waste of non-circulating gear oil trapped in the hose. Once my
transmission is fully seated and no longer producing chips, I'll
consider replacing this fitting and hose with a Curtis drain
valve (note that chips can cut the O-ring at the base of these
valves and also prevent proper sealing).
Update: I reread the instructions that came with the transmission
and it appears that CHR considers this hose to be used both for
draining and for re-filling the transmission oil. The drain part
is obvious but to use it as a fill line means that the lower adel
clamp will need to be removed and the hose bent upwards so the
end of it is at least level with the sight glass ... much above
this level and one will get an air bubble that distorts the
reading. I'll be using the vent fitting as the fill point.
I chose to leave the factory plug and shipping oil in the transmission during the installation. The one negative of this is that there is limited access to this area when the plug is changed to the hose fitting ... in fact it is quite awkward to work on. It probably would have been easier to change this fitting (with a temporary cap) just before final installation of the transmission. I ended up using a small margarine container wedged against the fan and transmission while I worked the plug out. The hose fitting was then installed and the hose added before removing the container. Luckily I only got a couple of drops of gear oil that missed the container ... actually drips from the tools I was using.
Most machines that I've looked at have the drain hose attached to the firewall with an adel clamp. One needs to first remove the adel clamp in order to remove the band clamp and plug when changing the oil. I chose to mount the lower adel clamp on a short spacer (3/8") in order to be able to remove the band clamp and plug without having to first remove the adel clamp. It appears that I can simply mount a temporary can under the hose end using safety wire over the spacer and then remove the plug when changing the gear oil ... simple with no mess, no fuss. Note that one has to be quite careful when drilling the hole for the upper adel clamp in the following picture as there is a frame member in close proximity.
For now, I'm leaving my chip detector plug out of the transmission. It is my understanding that the plug shipped with the transmission has a strong magnet attached to it which is actually more effective at attracting any chips, albeit without an indicator lamp. Once the magnetic plug no longer shows any chips (perhaps 50-100 hours), the chip detector will be installed and connected to the indicator lamp.
Observation: When I removed the factory installed shipping plugs in my transmission to install the hose fittings and temperature sender, I had to spend a lot of time trying to remove the various residual threads of the teflon tape the factory used to install the NPT plugs. Throughout the instructions / manuals there are also various references to using teflon tape. I sincerely wish the factory would remove these references and actually follow normal aviation practices. There are a lot of good options for thread lubricants / sealants and in my opinion, teflon tape should NEVER be used around aircraft.
CLUTCHES
I have had the opportunity to look at a newer Safari and one
thing I noticed is a change in the clutch assembly. The first
thing I noticed is there are now castellated nuts sticking up
through holes in the shield plates. Upon closer inspection one
can see there is also a change in the inner diameter of the
shields and changes to the races. I assume there is also a change
to the bearing(s) but it is not obvious whether a different sprag
clutch is being used. While it is apparent that there is now more
environmental protection for this assembly (a good thing), this
kind of change really disturbs me. I have never seen any
documentation about the change on this critical part and one of
the few detailed diagrams (AP 108) is now out of date. Should
owners of the older style clutch consider upgrading? Is this just
a one off change or a newer and permanent way they'll be
delivered? Are there any other advantages to the new clutch
assembly? I have absolutely no idea since the factory has chosen
to just quietly introduce this change without documentation
changes or notifying owners about it. Needless to say, I'm really
not impressed with how this kind of upgrade is being handled ...
in fact, I'm really quite ticked about it.
Update: I've been informed by CHR that the
style of bearing on the red frame below is actually the newer
version ... it's not clear why the green frame below which was
delivered six years later has an older style of bearing and
shield plates.
Old style on the left and new style on the right
SWASHPLATE
One of the things I noticed is that there is several degrees of play in the upper swashplate ring. The more I looked at this, the more I realize that it is due to the knuckle bushings, both the lower single knuckle (four bushings) and the upper links (eight bushings). It would appear to me that there is both end play and the bolt holes are slightly too large which allows twisting motion. It would also appear to me that there is an inconsistency in the shoulder widths on these bushings but at this time I haven't removed them for actual measurements. This slop in the swashplate makes me wonder whether it might be allowing the 2/rev vibrations to be magnified.
I believe these bushings are made from Delrin AF® and while there is a thermal expansion of this material, more importantly for my climate is that there is also thermal contraction at lower temperatures. I've obtained some stock that I will be using to either just machine shim washers or actual bushings. Ironically, these are the same bushings that are described in the Rigging Manual (page 13) as an item to be inspected after 50 hours of use and replaced if wear is apparent ... in my case, the slop is obvious in the as-delivered bushings.
The main rotor head comes pre-assembled and is basically a bolt-in part ... supposedly. Internally, there are three ball bearing on each side that make up the main feathering axis bearing and take the thrust load of the blades. After reading various articles, it would be very interesting to see what would be the effect of replacing these ball bearings with an elastomeric bearing. The general concept of the articles was that due to the very small axial movement of ball bearings in this application, they are prone to fretting and wearing. Also, due to the high centrifugal forces, ball bearings actually have a much higher friction torque than would be expected. Since the feathering motion occurs (and reverses) twice per revolution, this can induce a 2/rev vibration. Although it's probably not a cheap solution, perhaps this is the ultimate cure for the Safari's 2/rev issue. Certainly it sounds like it would reduce the vibrations. One negative of elastomeric bearings is that they have a calendar life and need periodic inspection since they do deteriorate with time and use. Once I get the actual bearing part numbers, another alternative might be to research whether there is an alternate ball bearing that is higher precision and/or better designed for lower friction torque in this application.
The blade grips contain oil to lubricate the bearings. I am aware of a couple of builders who have replaced the oil with ATF (Automatic Transmission Fluid) which is thinner in viscosity. While one owner is very happy with the results, another owner perceived that the thinner ATF slightly increased vibrations and he went back to the thicker oil which seems to have a bit of a dampening effect.
Within the yoke caps are two head bushings (part # 11 on
drawing AP107) and I'm aware of one builder who had an issue with
their clearance. Luckily he was at the factory for initial
rigging and they went through several sets of bushings trying to
find a pair that gave the proper fit. I'm not sure whether this
was a ganging of tolerances or just lack of tolerance, but each
pair of bushings gave a different fit. Regardless, this makes me
really question why initial shipping of this pre-assembled unit
was done with bushings that weren't deemed as appropriate when
the rigging was later performed by the same company.
I'm now also aware of another builder who carefully checked his
yoke and discovered that the wall thickness on his bushings was
not consistent. Although the yoke caps were machined accurately,
the off-axis bushings effectively moved the teeter pin out of
alignment.
One of the things I noted when installing the yoke caps is that my head angle plates will require some machining. The area where the washers for the four bolts seat was actually onto the curved inner radius of the head angle plates. Thus if one is to try snug these bolts down, the washers would be at an angle and one would not be getting accurate torque readings. These were milled flat before final assembly.
It also turns out that both my yoke and yoke caps were not milled accurately. There are slots in the yoke and tangs on the yoke caps that lock into them. These are quite close tolerance as these pieces take the entire main rotor head load and need to hold everything in place without any movement. It turned out that both the slots and the tangs were slightly off center which means that the teeter pin would be rotating in a small circle rather than precisely above the main shaft. Luckily another builder an hour away still had his transmission in the box and they were milled accurately. We swapped parts and he is awaiting replacement parts from the factory ... thanks John, it really was appreciated! I would highly recommend that a builder carefully check these parts with a micrometer before installing them and trying to balance the main rotor. This misalignment would have made it very hard to get a good dynamic balance.
Update: The next Safari that the person helping me with my balancing worked on had exactly the same problem with the yoke and caps. Since they didn't have an option of readily available parts, he had to use weights in order to try compensate. In the end it took about 21 grams (~ 3/4 ounce) on one tip in order to get it in the ball park.
We also found that my main rotor drive pin (i.e. teeter pin) was slightly long and did not allow for a smooth teetering action. The person helping me with the final rigging and balance chose to take a small amount of material off each end and it now teeters very freely.
I am aware of at least four different types of main blades that have been used on the Safari:
What really makes the above confusing to me as a relative newcomer to the dynamics of rotorcraft is when I read about the Waitman blades and their description of the advantages of rounded tips. Also, the Robinson R44 Raven II has switched from square tips to rounded tips in order to reduce noise.
I decided that I wanted to add a stripe on my blades for
increased visibility. Since I was also going to add tip weights
and the one bolt is in the area to be painted, I decided to do
these tasks together. The first thing I did was to mark the
stripe out on one blade (10" wide and 10" from the tip)
and added the edging tape for painting. This distance will leave
enough clearance from the paint (~2") if I later decide that
I have to add trim tabs. I then marked the other blade, but
something didn't look quite right. I double checked the mark by
measuring off the blade strap holes. It turns out there is just
under 1/2" difference in the two blades at the trailing edge
... this was triple checked and then quadruple checked by laying
the blades on top of each other. I'm aware that the round tips on
these blades needed to be fabricated by hand (not sure why as I
haven't seen the molds), but this is a bit extreme!
Update: This difference in blade length has now also been
confirmed by measuring from the head shift bolts with the blades
mounted.
Tape is the exact same distance
from the outer blade strap hole
For those that have the blades with rounded leading edge at
the tips, they should note that the Owners Area of CHR's website
has a document about adding tip weights to these blades. From the
reports I've received, it would appear that this change has made
a very significant reduction in the vibration levels on some
ships.
Update '08: I double checked with the factory whether they are
still recommending this change for the swept tip blades ... the
answer can be abbreviated to one word: YES.
Time to mark the holes for the tip weight modification. Hmmm ... how to handle the different blade lengths? First I verified that the center of mass for the two blades was roughly equal (it was within 1/16" spanwise). Since the two bolts in each blade will obviously effect their balance, I decided it was best to try match their position relative to the blade strap holes. Although the instruction measurements are relative to the blade's tip, I basically split the difference between my two blades. This was double checked to verify that the marks were exactly the same distance from the strap holes. Although the procedure for drilling and tapping the holes is pretty straightforward, I can't say that I enjoyed it ... the whole time I was doing it, I was thinking about how a mistake could cost me a set of expensive blades. Oh well, it's now done.
I had heard about another builder who had a plugged vent hole after performing this change. I installed a piece of 1/8" nylon tube about 12" long into the holes before adding the epoxy. Sure enough, one of the tubes was glued in place and could not be removed ... without the tube I probably would have had some epoxy running out the hole. Why does this happen? I can only speculate that the spars on some blades were not completely sealed during their fabrication.
Getting the epoxy and shot into the blades proved interesting. The instructions are somewhat ambiguous and it sounds like they may have layed the blades back horizontal to install the shot but that would allow the initial epoxy to move. I chose to keep the blades vertical and this required a unique pouring mechanism for the 1/4" hole. A cup with a short straw hot glued to the bottom corner worked okay but it did require the use of a "stir stick" to keep the shot moving through the straw. The 60cc syringes that we use have a catheter tip that is too big to fit in the hole. A piece of roughened 1/8" tube hot glued into the end worked well to overcome this but it did result in a slow injection ... probably a good thing.
With the expoxy and shot installed, it was time to add the remaining two screws ... these were added with red Loctite and a bit of JB Weld to make sure the counter sink was fully filled. Actually the JB Weld was also used to fill in the slot in the screws. With the bolts in place, the stripe on the blades was then painted. I briefly considered whether to paint the full underside of the blades black so as not to reflect the strobe or ground lights during night operations. I decided this was not worth the effort and that any night flying I do will be a real exception with a high probability that it's only a matter of making it back home just after twilight. Unlike most certified machines, the underside of my blades will remaining in white gel coat (at least for the time being). I have also mentioned to the factory that they may want to investigate the use of white gel coat on the top for heat resistance on composite materials and black gel coat on the bottom for strobe reflection.
Before the final assembly and static balance we had to decide how to handle the unequal length blades. After much contemplation, we decided that it probably was important to the aerodynamics to have both blades as closely matched in length as possible. The blades were stacked and the outline of the shorter blade was transferred to the longer blade. The longer blade was then sanded to shape and cross-section before spraying it with epoxy paint.
I have an unconfirmed second-hand report from someone who has tried both the round tip blades and the square tip blades. It seems that they feel the new square tip blades are noisier and may require more power ... interesting contradiction to the factory statements. I'm trying to find more details on this setup and comparison.
I am aware that many of the ships from New Zealand are adding trim tabs to their blades. This makes a lot of sense, especially for the older round-tip blades, since it allows for a lot more flexibility when doing the tracking for hover versus forwards flight. The March 2007 issue of Experimental Helo has an article describing the implementation and use of these tabs on Safari blades. Note that there is a typo error in the article with reference to the adhesive that was used ... it is actually Plexus MA550 not MS550.
As to whether to add the tabs to my blades before balancing them ... we decided that it would be appropriate. Although the original article describes tabs top and bottom joined at the trailing edge, mine were done in the currently recommended way of just a single piece. These are normally applied to the top of the blades but I think I had a jet-lagged installer from the Southern Hemisphere ... mine are on the bottom. Although this is not as visually appealing and may have a slight aerodynamic penalty, it's probably stronger as the aerodynamic loads will be pushing the main part of the tab into the bonded area rather than trying to pull it off.
As a double check, after installation we checked the angle of my blades both at the root and at the tip. As it turned out, both my blades have a slight twist in them but in opposite directions: the blue one increases in pitch 1° from the root to the tip while the red one decreases 1° from the root to the tip. I'm fully expecting that these blades will require the trim tabs to compensate for different tracking at hover versus forward flight. As to why this might have happened ... it's been my experience that molded composite parts should be left in the mold much longer than what the resin specifications state. Although the resins may have "kicked" and achieved most of their strength during the stated times, I have seen cases where they're still slightly "green" (i.e. soft) and will easily warp during the final curing process. Leaving parts in the mold as long as possible (even for weeks) will help prevent this kind of warping.
I have absolutely no further information on these blades, but
I ran across this site on the web from Composite Blades and
they appear to offer a set of main blades specifically for the
Safari. Ironically their picture of the mounting system for the
Safari blades appears to be for an R-22 so I wonder if they've
actually tested them on a Safari.
Ooops ... looks like their website is gone. I don't know if
they're still around but if anyone has more information they'd
like to share then I'll update this.
Cheek Plates
When it came to mounting the cheek plates to the blades, I first took a real careful look at them. Although they are ferrous, it appears that they have had some kind of a liquid (i.e. dipped) based coating applied to them for corrosion protection. I am also aware that one builder commented to me that he had seen rust developing on his. More importantly, I noticed that the edges of my plates had not had their sharp corners knocked off and there some scratches that needed to be removed with a Scotch-Brite pad.
The sharp edges were lightly filed and the scratches were buffed out. Luckily the scratches were on the plates that will be on top the blades and mainly in compression rather than tension. The cheek plates were then glued together with 3M77 contact cement ... my feeling is that with all the bolts that go through these plates and with their relatively tight holes, there really isn't an issue of the cement doing much more than holding the plates in alignment until the bolts are cinched down. The other irony with the cementing process is that the cheek plates I received had not been flat surfaced after welding on the reinforcing angles. Thus the outermost plates had a crowning on their inside surface (i.e. the glued side) due to heat penetration from welding.
Since the above preparation process removed some of the surface coating, I decided it was best to paint these parts. The side that goes against the blade and grips was just left in primer and the exposed sides were given a light topcoat. Since it is important that any cracking in these parts be immediately spotted, these paint coats were applied very thin and primarily for corrosion protection rather than for cosmetic looks.
When I first tried to test install the bolts into the cheek plates before painting, I discovered that all of the holes were very tight ... actually to the point where I didn't want to force them and cause scratching. This was reconfirmed when they were ready to mount them permanently. I ended up having to ream all of the holes (20 * 4) ... luckily I had both a 5/16" reamer and a 3/8" reamer available. The first mounting of the cheek plates was done using temporary nuts on the 5/16" bolts. I was rushing to get the lades installed before the final inspection and knew that I'd be removing them to glue the cheekplates to the blades and to perform the final static balance.
Mounting the Blade Assembly
The blades and cheek plates were first mounted to the main spindle while everything was on the floor ... note that this will require some propping of the blade ends due to the pre-coning machined into the spindle. Afterwards, four people were involved to install the blade assembly into the yoke ... one on the end of each blade to support it and prevent them from rotating and one on each side of the yoke to verify alignment. It was definitely a very tight fit to try get the yoke caps into the yoke. I would highly recommend that one takes their time to get perfect alignment before even thinking about trying to tap the yoke caps into place.
This process was actually performed twice ... once for the inspection and again when the various operations such as tabs, leading edge tape and static balance had been performed. It was actually much easier the second time after the main rotor drive pin had been slightly shortened.
Note that there are two tapped holes on the top of the main rotor head box. These can be used to temporarily attach a piece of angle iron (or two pieces) which can then be used to lift the blade assembly. By using some kind of a hoist, the entire blade assembly could be installed by one person alone.
Blade Tape
Because the main blades are of a composite construction with a gel coat, their surface (especially the leading edge) will not hold up well in either a dirty environment (i.e. sand) or in rain. Less than five minutes running in rain can totally remove the gelcoat from the leading edges. To protect the blades, a leading edge tape is used. The Information Package makes a reference to stainless steel tape but I have yet to find a good source for this or hear from owners who are using it.
It appears that the tape of choice is Aeroshield 2604 which is a polyurethane tape that is 0.35 mm (~.014") thick and is available from Aircraft Spruce in clear or black and 2", 3", 4" and 6" widths. While the clear tape will yellow when exposed to UV, I don't think I'd be comfortable putting black tape and the ensuing heat buildup on a composite structure. This tape is VERY expensive and I still need to do some more research on the recommended width and the length of the leading edge it needs to be applied to. From the old Safari forum, it would appear that 4" wide is the width of choice but it will not wrap around the ends of the curved blades.
Update: It would appear that Aircraft Spruce is no longer carrying Aeroshield 2604 so the search is on for an alternative. Aviall carries 3M 8671 which is the same thickness as Aeroshield so I might give that a try. I've noted that some owners are using 2" tape and some are using 4". While the 4" is harder to apply (especially on curved tips), I think aerodynamically its best to move the seam line a bit further back from the sharp leading edge.
Sometimes one finds things like this tape that is used for different applications. Digging around on the web, I noticed that ISC Racers Tape packages a "Surface Guard / Helicopter Tape" product which is available from such outlets as FindTape. I have not seen this product or done further investigation, but suspect that the 14 mil product might be useable on the Safari blades.
Update: Another Safari owner has found a unique source of 4" clear 14 mil tape that is not supposed to yellow. Initial tests are very positive and I obtained some of this tape from him for my initial installation. This was first installed while the blades were on the bench and before lifting the blade assembly onto the yoke ... much easier than doing it on a ladder. Since I have the old swept tip blades, we chose to taper the last bit of the tape down to a 2" width in order to make it easier to wrap around the tips.
I have read on various forums some condemnations of blade tape. In particular, the Rotorway crowd talks about how it robs a noticeable amount of MP if it is applied. They attribute this to the lip at the trailing edge of the tape and that it disturbs the airflow (boundary layer as I remember) on their non-symmetrical blades. I talked to one Safari owner who made a direct comparison while he was re-applying blade tape. First the blades were well cleaned and then an exteneded hover test was performed to get a baseline on the MP requirement for that day. Then the blade tape was applied and he immediately did another hover test within about 1/2 hour of the previous test. The result that there was no perceptible differences in the MP requirements.
Another Safari owner did a test with 2" "thick" tape and 4" "thin" tape. He reported that he had no problems with increased power requirements or increased vibration with the 4" tape but did have problems with the 2" tape ... unfortunately he didn't clarify whether he thought that was due to the thickness or width of the tape. He also didn't clarify his relative terms of "thick" and "thin".
For reference purposes, here's a link to some of 3M's protective tapes.
Tidbit: I notice that 3M has a protective kit available for heavy helicopters; part # 8999K11. What's interesting about this is that it uses both an 8" black tape and a 2" sacrificial tape in front of that. Also, they apply the wide tape with 5" on top and 3" on the bottom ... sort of opposite of what I would think would be logical when one considers blade angle of attack.
Tail Rotor Drive Shaft
When I first tried to fit the driveshaft into the pillow blocks, I found that only one of them would slide over the shaft, one was close to sliding over and the other four definitely would not slide over. A quick check with a 5/8" reamer indicated that all of the bores were already loose on it. Looking at the bearing specifications indicated that the bore tolerance was +0.0006/-0.0000" and their shaft tolerance was +0.0000/-.0005". The supplied 4130 driveshaft was .625" to .6255" but I'm not sure how round it was, whereas the bearing bores were ~.628" ... hmmm, since the bores didn't have the oxide coating, it makes me wonder if CHR had already opened up the bores but not consistently or perfectly round. Regardless, I had to open up the center part of the bore to allow the shaft to slide over. The reamer didn't seem to take off the high spot and since I didn't have a very small flapper wheel, this was done using a fine sanding drum on a Dremel tool on the high spots to obtain just enough clearance (thanks for the idea Rick).
From the above, one can see the problem with driveshaft preparation. Before this trial, I had already removed most of the mill scale but eventually the shaft will require some kind of protective coating (i.e. paint) which will only increase it's diameter and cause more problems. If the shaft is painted, this will be the time to practice applying thin coats. I'm still debating whether I might try something like Policoat or Dry Film as a protection since they're typically applied only about .001" thick or less.
The next task was to use shims to do a rough alignment of the
shaft so I could cut it to length and weld the doubler on the
tail rotor end. First I added the taper lock coupler onto the
front of the shaft so I could temporarily mount the Omega coupler.
Simple ... well it was after I discovered that the slot for the
key in the driveshaft hadn't been milled to a consistent depth (about
.005" high at one end). Instead of trying to alter the
slot's depth, I took that amount of material off the one end of
the key.
Tip: If the tail rotor driveshaft is slid all the way forward to
where it meets the taper lock coupler on the transmission, it is
extrememly easy to perfectly align the two 5/8" shafts and
to eliminate any undue flexing of the Omega coupler.
Update: I've talked to another builder who used this technique
and later had opportunity to use a factory jig to verify the
alignment. He was quite surprised to find the factory jig and
dial indicator showed that there was a significant misalignment.
Tip 2: The star coupler is bolted to the driveshaft however the
taper lock coupler allows fore-aft movement before tightening and
can be used to make slight changes in the driveshaft length when
it is installed.
Driveshaft slid forward - bad flash shadow but you can see the alignment with the transmission output shaft
The Construction Manual refers to using a 2-1/8" spacer to emulate the Omega coupler and to align the shafts. I found that to be a bit too much ... 2-1/16" is probably closer. I just used the coupler to set the length, being careful to make sure that it was in it's default position rather than stretched or compressed.
N.B. The instructions that come with the Omega coupler refer to capscrews that should be torqued to 200 in.lbs. These are not the bolts that are supplied in the kit and I have talked to one builder who twisted the head off one of the supplied bolts as he tried to meet that torque specification. I'm not sure of the exact bolt type in the kit, but I'm going with the AC43.13 spec of 40-50 (max 75) in.lbs. for a 1/4-20 UNC bolt ... also note that Lycoming specifies 96 in.lbs. for a lubricated 1/4-20 bolt. As far as I know, the normal capscrews are not safety wired so we do get a bit of extra protection.
The Lovejoy (i.e. spider) coupler (actually a Boston Gear FC15) that goes on the driveshaft caused me some problems as the bore was somewhat undersized. I first ran the reamer through it and there was definitely material being removed. It was still undersize and I decided to mount it in the lathe to do an inside bore. A quick power-off check with the reamer inserted showed that the bore was somewhat angled. ??? I have no idea how this happened but the lathe got it pretty true ... also interesting that I didn't want to take off too much material and there was still one high spot that I then sanded off. I regret that I didn't put the micrometer on this piece and also check it in the lathe before I started ... I don't know if this was the manufacturer's tolerance, the bore was initially ovalled or if CHR had performed work on the bore (I believe they drilled the setscrew out for the through bolt and spot faced the resulting holes). In hindsight, what I don't understand is that the manufacturer's tolerance is +.001"/-.000" and the reamer had to remove material to be inserted ... unfortunately the exact part number had been removed when the part was faced for the bolt. I also had a bit of a problem with the facing (just under 1/2") that had been done to provide a flat surface for the bolt's washers ... the AN960-4 washers I received with my kit were mostly about 0.506" and wouldn't seat properly. I checked around the hangar and luckily found some AN960-4 washers from a different batch (or manufacturer) that were 0.497" and although they were a very tight fit, they worked.
Update: During final installation I found that my driveshaft had ~.015" runout right in front of the spider coupler. After triple checking this, I decided it was too much and needed to be removed. The first trick I tried was to rotate the coupler 180° on the driveshaft ... luckily this removed the visible runout. I assume it was a ganging of tolerances from opening up the coupler plus removing the topcoat from the coupler area of the shaft that had caused the problem.
After cutting the driveshaft to length, the doubler was welded in and both ends of the shaft were welded shut (the forward end had only been fuse welded at the factory for the doubler and not closed out). The bolt hole for the star coupler was then drilled. Each builder probably has their own way to try do match drilling like this but the objective is to try get matched holes without having to resort to ovalling of the hole ... easier said than done. I used a combination of a V-block, undersize drills and reaming to get the job done ... luck was on my side this time and there was no need to oval a hole.
There would have been quite a delay for me to have Dry Film coated onto my driveshaft so I decided to just go ahead and paint it. The other concern I had is that steel needs to be prepped just before coating and I have no knowledge of the sandblasting outfit that the local coater uses ... there's only one sandblast company in town that I know for a fact is familiar with prepping 4130 for aircraft use. I ended up putting on just enough primer to get an even coat (just past the transparent stage) and then a light topcoat followed by a "just wet" topcoat. A heavier wet final coat would have given a better glossy wet look to the driveshaft but I was more concerned about just getting full coverage with a minimum of buildup. Final dimension of the shaft with paint was ~0.6305" for a paint buildup of ~0.005" or 2 times ~0.0025" which is right at (or slightly under) the lower end of the manufacturer's recommendation for primer plus topcoat.
Many builders choose to paint the pillow blocks for the tail rotor drive shaft. Looking at the manufacturer's data, the races are already specially coated with a black oxide for corrosion resistance. Note that the inner races appear to be honed or sanded and they will definitely need something like paint or Corrosion-X applied to them in order to prevent rust. If one chooses to paint the inner races then I'd delay painting the bores until after the driveshaft has been painted and the races have been checked for final clearance. I chose to apply a thin coat of Perma-Slik G to the bores along with some Corrosion-X. Another alternative might be a bit of sealant on each side but one needs to consider how the bearings will be removed if they ever need to be replaced.
After painting the driveshaft, the bores on all of my pillow
blocks needed to be slightly opened up due to the ~0.005"
thicker shaft. This time I was looking for a more accurate way of
opening up the bores while using the tools I had available. After
a bit of head scratching, I found a 2" long x 1/2"
diameter sanding drum with a worn out disk and no replacements. I
wrapped it with squarely cut sticky back 120 grit sandpaper until
it would barely fit into the bores. A couple of quick passes of
the die grinder using a honing motion then opened up the bores
just enough to slide onto the shaft.
Tip: Fresh sandpaper is cheap and cuts quite quickly. It's best
to go slowly and do lots of checking. If the same piece of
sandpaper is used on multiple bores then be careful when
switching to fresh sandpaper as the next bore will go much
quicker than the last one.
Tip 2: Don't go at the bores for too long at a time as would
happen with very fine or worn out sandpaper. This technique will
cause the bores to heat up significantly and the lubrication can
be thinned and start to run out.
I had planned to install a Teletemp on each of the pillow blocks and there is a nice square pad on the top of each one. However, it should be noted that these are isolated pillow blocks and heat does not transfer from the races to the frame. While there is a 1/2" wide curved projection of the bores, the Teletemps that I have are 3/4" wide. I chose to cut down the Teletemps and will have to wait and see how well their adhesive will hold on the curved surface.
Once I was ready to do the final installation of the pillow blocks and couplers, the first thing I did was to slide all the pillow blocks into their rough position. I then added the rear spider connector to the driveshaft but just before doing so I used a small syringe with a bit of heatshrink tubing to add about 4 cc's of Tubeseal into the driveshaft ... probably 1 or 2 cc's would have been more appropriate. I then proceeded to align the pilow blocks into their final position. The Construction Manual refers to adding a small spacer shim inside the spider coupler to allow for end play after assembly ... I realize that the shim needs to be physically "small", but this gives no data as to it's thickness. I used a 0.15" shim as it "looked about right".
CHR uses a dust cover over the tail rotor star coupler and I find no reference to this in any of the documentation. On factory built craft they use a piece of hose over the coupler that is then held in place with a band clamp (actually two for balance). While this may serve the purpose as a dust cover, it also makes it impossible to inspect the coupler during a DI (Daily Inspection) unless one actually moves the band clamp and hose.
I chose to try clear heatshrink tubing as a dust shield to allow for easy visual inspection of the spider.
It's my understanding that newer tail rotor driveshafts now
have a slot and key for the spider coupler whereas mine simply
relies on the single AN4 bolt. Although it was reamed and a very
tight fit, in only a couple of hours of running the hole has
enlarged and started to oval. One of the recommendations to me
was to take this assembly apart and use a lot of permanent
Loctite under the bolt side in order to try permanently lock that
side of the coupler in place on the shaft. Obviously I'll have to
do something like that ... sooner rather than later.
Update: I've now applied the Loctite. Also
I checked the setscrew in the rear part of the coupler and found
it to be on the loose side. It was re-installed using blue
Loctite.
Tail Rotor Gearbox
I finally removed the tail rotor from it's shipping crate although it had been visually inspected upon receipt. I immediately detected that one of the bolts that held it to the crate was very stiff to remove (ratchet clicked the whole way) and another was quite stiff for about the first three or four turns. The first thing I did after removal from the crate was to double check this as these four bolts are the only thing that hold this critical part to the frame. I guess I hadn't thought about it before, but the holes actually have Heli-Coil® inserts in them as they use fine bolts into the casting rather than the more normal course bolts for this type of application. It appears that the one with just a few tight threads had not been tapped deeply enough before installing the Heli-Coil® insert or that it's tang hadn't been cleanly removed. More importantly, it would appear that the stiff one was either not installed deeply enough, it had been cross-threaded when mounted in the shipping crate or the bolt had been bottomed out ... in any case, the result was that the end of the coil stuck out and did not form a properly threaded insert. I chose to replace it as this is a critical bolt and the damage to the shipping bolt was obvious ... it was both gouged beyond the end of the threads and the threads were rounded. A waste of several hours of my time and $28 due to careless assembly and/or crating at the factory.
First I removed the old Heli-Coil® by carefully unwinding it towards the center of the bore using hemostats. I had been given a couple of 1/4-28 x 3/8" inserts by a distributor who didn't have a basic install kit. I tried to install these without the use of a pre-winder but just succeeded to break the tangs off when they were about half way in. I then made the long trip to the industrial area on the far side of town and got the proper basic install kit. Close inspection of the hole and very careful use of the tap indicated that the top part of the threads had been damaged when the original bolt had been bottomed out during crating. After cleaning up the hole, it was a trivial matter to use the pre-winder to install the replacement coil.
The pitch control on my tail rotor was quite stiff and difficult to move. Although this assembly had been put together by CHR and fully cotter-pinned and lock wired (including a backwards one), they bypassed the step of making sure that the rod ends were first loosened up. Upon disassembly, the rod ends I received were extremely stiff and could not be moved with my fingers unless a bolt was put through the hole for leverage. Obviously I chose to loosen them up the same way as was required for the other rod ends supplied with the kit ... the result is the pitch control went from a very stiff assembly to one that is now very smooth and easy to move. I have no idea why CHR chooses to avoid reading / following their own sketchy instructions and ships units that appear ready to use when in reality they first have to be dis-assembled.
What was more shocking is that one of the bolts holding a rod end in the slider cross was severely over-torqued. Two formally trained aviation people happened to be talking to me as I was removing this bolt and both were amazed at the amount of force I had to apply to remove the nut ... I'm guessing at well over double the limit for an AN4 bolt. Needless to say, the bolt and nut immediately went in the garbage so they would never be re-used. Further investigation revealed that the two AN960-4L washers between the rod end and the inside "U" of the slider cross were thinner than normal ... 0.022" versus the normal 0.030". Instead of changing the washers, the person at CHR assembling the tail rotor chose to just apply more torque to the nut/bolt causing the slider cross to bend and make up the .016" difference. I now have very serious concerns about each and every bolt that CHR has installed on the tail rotor and transmission assemblies. A torque wrench was obviously not used and their "calibrated fingers" are in need of a major readjustment.
Because of the chip detector, it is important to have a good electrical bond between the tail rotor gearbox and the frame. It would be easy to add a stud centered at the rear of the gearbox mount for both this and the tail light but by the time I remembered it, the gearbox and driveline had already been installed. Instead of removing the tail rotor assembly, I chose to remove the paint under the forward right attach bolt (least strain) and add an MS35333-74 stainless star washer between the bolt and frame using lots of DC4 to prevent corrosion. An ohmmeter had confirmed that there was no electrical bond before this change but there was a good bond afterwards.
My tailrotor output shaft was machined very roughly and it would appear that there was no final pass on it when it was machined. Running my finger nail on it there was a very obvious washboard texture to it. I could also hear the delrin bearing squeaking across it when moving the blade angles ... sort of like a finger sliding on a guitar string. The person helping me with the final rigging had seen this many times and chose to correct this before it wore out the bearing. Progressively finer grits of wet paper were used to remove the ridging.
Tail Rotor Blades
The original tail rotor blades were made with aluminum skins while the current ones have a stainless steel skin. I also understand that the original style blades are NOT RECOMMENDED anymore. I'm also aware that composite blades were tested many years ago by CHR but as far as I know they are not available. While these blades would be expensive, there are many potential benefits.
It was reported on the old Safari forum that a user with aluminum blades weighed them at 380 grams each versus 790 grams for the new stainless steel ones. These lighter blades had a lot better tail rotor authority and resulted in lower MP readings in the hover.
South Pacific Home Rotors has now developed tail rotors with titanium blades that are said to be 700 grams lighter, quieter and result in better tail rotor control. The Sept. 2007 issue of Experimental Helo has an article describing the evolution of these blades and they certainly appear to be very beneficial. I will be seriously investigating whether to upgrade to these blades as I feel that quick tail response is very beneficial in gusty winds. The reduced weight will certainly help to reduce the tail heavy tendancies of this craft.
I am also aware from the old Safari forum that one builder/owner chose to add endcaps to the tail rotor. His comment was that the noise was significantly reduced, but I have no further details. Interestingly, I noticed a lot of noise when the tail rotor was installed on the frame and spun without the driveshaft connected. It would appear to me that there is a lot of bearing or gear noise that is being transmitted and/or resonated through the blades. With my tail blades powered, they're extremely noisy. In fact, the people helping me balance the craft stated that they're the noisiest blades they've heard. One of the factors may be that these are earlier blades and there is a plate on the inboard end of the blades that fits outside of the blade airfoil. This piece is not very accurate and there is a gap where it meets the blade material. perhaps some kind of filler will reduce the noise but there is also the risk that it will be flung off and result in an imbalance.
Another problem with my tail blades is that one of them was delivered in a leaded position which caused us a lot of problem when trying to dynamically balance them. There is no lead/lag adjustment for the tail rotors and they rely upon the accuracy of the three holes that are drilled at the factory. Unfortunately mine were not drilled precisely. We tried to slightly ream the holes and then lock the blades in place with red loctite ... this worked for the alignment but after a few balancing runs it was obvious that the blade had shifted again. Eventually we got a couple of small .010" stainless shims (about 3/16" x 3/8") strategically placed against the blade inside the hole and those seem to have held it in place for now. Not what I would consider a real good solution but the best we could do under the circumstances. Hmmm ... I wonder if this is considered to be a warranty replaceable part?
There has been some discussion as to whether the tail rotor blades should be painted and/or leading edge tape applied. I think at this point that I'll go along with the majority and just leave them as plain stainless until I get a more definitive answer or requirement.
It is my understanding that the "China weights" on
the tail rotor make for lowered control forces when changing the
pitch of the blades. On my tail rotor there were three AN970
washers on each "hat" whereas looking at a newer craft
there are four washers in each of these. I haven't followed up
with the factory about this and don't know what the impact will
be.
Update: I need to do some research on these weights. I find that
my right pedal requires a LOT of pedal
pressure. In fact, I initially felt that I was on the pedal stop
most of the time. I've found that at this time I have to keep my
heels off the floor in order to consciously think about applying
enough pressure on the right pedal.
Trivia: I am aware that Jim McCutchen (the original developer of the Waitman / CHR main blades) is developing a new set of tail blades with a flexible hub for the Mosquito, Rotorway and Safari. I have no further information nor do I know when he plans to start testing them. As to the rumour that he is also working on new style main blades ... I have no further information.
Final Pillow Block Shimming
Before beginning this process, the pillow block mounts should be checked both for any twist in them and for any longitudinal alignment error. If required, the mounts can be "tweaked" using a pair of parallel jaw grips ... in my case, four of the six mounts required significant adjustment. The general instructions for the pillow blocks indicate that the bearing and shaft should be mounted within 2°. I also chose to pre-drill a #40 pilot hole in the the pillow blocks for the the roll pins that will eventually be added ... I'm using 1/8" pins but prefer to initially drill these holes undersize before drilling to final size. A small corner jig mounted to a drill press made it very easy to set the position of these twelve holes in the corner of the pillow blocks.
I was asked what material I intended to use when shimming the
pillow blocks on the tail rotor drive shaft. Aluminum will
probably work, especially if one has lots of different
thicknesses, but the problem becomes how to finish it for
corrosion protection without changing it's thickness. Probably
just alodine and possibly paint along the edge should be OK.
Steel would probably work, but there's the issue of both weight
and corrosion protection. I managed to pick up some 6" x 12"
thin sheets (.010", .018" & .028") of 430
alloy stainless steel from my local hobby shop in their K&S
Metals display. Hopefully I'll be able to make the
appropriate thickness of shims either directly or from some
combination of these without having to worry about corrosion
protection. If the stack becomes too thick, I'll probably add
some lightening holes in the various pieces or use a combination
of the stainless and thicker aluminum shims.
Update: Although this was a good thought, the use of the
stainless stock would have required a lot
of pieces and extra weight. I decided to go mostly with alodined
aluminum as I had the material available in various sizes from .250"
to .015" but I still ended up making three shims from the .010"
stainless. The other choice was whether to start with thick
aluminum (i.e. 1/8" to 1/4") or stacks of thinner
pieces. On the two blocks closest to the transmission I started
with .250" shims and on the others I used various stacks
from .080" and thinner. For reference, my final shim stacks
(front to back) ended up as: .330", .279", .199",
.147", .084" and .028".
Some of the shims I prepared ... a mix of .015, .030, .063, .080
and .250"
The top shim closest to the transmission is somewhat different on my craft and has a bracket welded onto it. For various reasons, I wanted to mount a gearbox sensor that gives me the tailrotor driveshaft RPM at two pulses per revolution. By using two of the supplied steel bolts and two stainless bolts I should get this. Since the supplied steel bolts were of the six hole variety with a depressed center, I chose to slightly offset the sensor so it would be getting uninterrupted steel exposure ... this also made the mount a little narrower which was good. Note that the use of this kind of sensor should be carefully planned so there is no way the sensor or it's wire can get caught on the driveshaft.
Initial testing of sensor mount
It is quite normal for the pillow block closest to the transmission to require quite a thick shim. I have heard from two builders who had problems with the rearmost pillow block mount being slightly too high. One builder chose to add a shim under the tail rotor gearbox which then allowed the driveline to be aligned. The other builder chose to slightly shave the rearmost pillow block thus lowering it. During the alignment of my driveshaft, the gearbox did not need to be raised and the rearmost pillow block actually needed a thin shim (.028").
It would seem logical to align the setscrews on the pillow blocks at different angles, both to maintain balance and to prevent vibration from possibly loosening multiple setscrews. According to the manufacturer's instructions, the setscrews should be aligned at either end of a shaft. The recommended tightening sequence is to first tighten one setscrew to 1/2 of it's final torque, tighten the second setscrew to it's final torque and then go back and tighten the first setscrew to final torque. Although it's not referenced in the construction manual, the bearing specifications call for 66-85 in-lbs of torque on the setscrews.
The Construction Manual describes how to align and shim the driveshaft using calipers referenced to music wire. I'm also aware of builders that have used a laser level to accomplish this. Perhaps it's my aging eyes (I now need reading glasses) but I have a bit of a problem getting really accurate readings with a laser. There's also the issue of making sure that one has an accurate centerline for each of the pillow blocks. I'm using the caliper method. However, I don't like the idea of the very heavy weights that the manual recommends when using music wire. Instead, I chose to use thin braided steel fishing leader material ... it's relatively flexible and doesn't require nearly the same tension to keep it perfectly straight. The manual calls for an accuracy of 1/32" on the shims but I found it was relatively easy to get about 0.005" if a variety of shim thicknesses are available.
Per the Construction manual, I pinned each of the pillow blocks using two roll pins ... 1/2" pins worked on the rearmost four blocks and 5/8" pins just worked on the forward two. I had been given some McMaster 93740A209 pins by another builder and these worked extremely well. For the longer pins, I used some standard split pins from an assortment kit at the hangar and these were much more difficult to install.
One of the things I like about many of the commercial helicopters is the use of Teletemp indicators to record if a bearing is generating excess heat which is an early indicator of potential failure. I have heard that immediately after shutdown, many Safari owners just use their hand to feel each of the pillow blocks and the transmissions to see if there is excess heat. I obtained some Teletemp 110-2 recording indicators (140°F to 190°F or 60°C to 88°C) and will be using these on the pillow blocks and on the transmissions. I believe it is much better to use a calibrated indicator that is independant of weather and human subjective evaluation. These indicators will be checked during post-flight inspection and will be part of the Daily Inspection procedure.
Update: The above indicators appear to work well for all places except the main transmission. The transmission temperature is red-lined on the gauge at 210°F, not because of oil breakdown or metallurgy but because of the temperature fit used during manufacturing. During extensive hovering in hot weather (especially during break-in) it is possible to hit this redline. In order to be effective indicators, these Teletemps should be a higher range on the transmission ... I assume the 110-3 indicators (180°F to 230°F or 82°C to 110°C) should work fine.
Cabin Controls
My kit came with two T-handle control cables that were to be used on the carb heat and mixture controls. These are the type that require a quarter turn to unlock them in order to allow actual cable movement. I also required a third cable for cabin heat since I'm located in colder climes. Personally, I do not like these kind of controls that were supplied ... they may be the cheapest ones but they take up a lot of physical space due to the T-handle, I've seen some that were not smooth acting and there is no orderable colour option for easy identification. I chose to replace all three control cables with A-700 button lock controls. These are available with red (mixture) or black (carb/cabin heat) knobs and require the operator to press the button before trying to move the control ... a simple one handed operation using your thumb. I find these control cables to be very smooth and simple to operate. The 96" (8') length was more than adequate for my installation. I'll probably also use a mixture guard similar to the one used in the R22 ... basically just a piece of clear plastic tube that slides over the the knob and prevents inadvertent selection / movement.
I noticed that another builder had chosen to cover his control cables with heatshrink tubing. This will help prevent any kind of chafing, oil or dirt penetration etc. For the extremely small weight penalty that this entails, I decided to also use this technique on the parts of the cable that are outside of the cabin. Thanks for the idea Mike!
It is relatively easy to route the carb and cabin heat control cables to avoid most obstacles ... took me a bit of head scratching though. The mixture cable has to be run either behind or in front of the frame tube near the carb ... I chose to run it behind since this gives a straighter run to the adapter plate. I've now noticed that after about five hours of engine time that the cable is a bit stiffer than it was initially. I'm attributing this to the effect of heat on the cable liner since the cable runs relatively close to the exhaust at one point (about 3/4") and is also in the airflow exiting the shroud. It's not obvious how this could be converted to a solid link to a bellcrank but that might be one solution. The simpler one will probably be to re-route the cable in front of the frame member as the heat from engine cooling should be less than the radiant heat from the exhaust.
Carb Heat
I am aware that at least one builder has experimented with an automatic carb heat system that is actuated via a solenoid. I had also thought about some way to either apply automatic carb heat or at least to have an electrically actuated system based on a switch at the end of the collective. For those that have trained / flown an R22, the use of carb heat should be an automatic reflex. By mounting the carb heat control on the left side beside / below the collective, I believe this is an acceptable solution that avoids any unnecessary complication and is easily reached in flight. It also avoids any unexpected abrupt change in power that could occur with an automatic system.
Rotor Brake
One feature that I like on the R22 is the rotor brake. While we tend to let the blades just about stop on their own before using the brake, it does come in handy to align the blades. It is also handy to allow stopping the blades quicker in high wind conditions. The design of the Safari's transmission / mast does not allow for a simple installation of a brake on either the main input shaft or main rotor shaft. This leaves the tail rotor driveshaft as the obvious place. However, I do have a concern about placing too much of a load on the tail rotor driveshaft and I think one would need to consider just making a friction device rather than a true brake. In the meantime, one should be VERY CAREFUL if they're just grabbing the tail rotor driveshaft by hand in order to slow things down.
I am now aware of a builder that has tried adapting a brake to the tailrotor output shaft on the main transmission. Although it adds a bit of weight to the machine, he indicated that it works quite well.
Cargo Hook
As previously mentioned, the factory supplied frame has a hardpoint for mounting a hook but none are available as a factory option. I have found references to several hook kits for the R22 with a 400 pound rating, but these are very expensive (in the $3,000+ range). Although one might be able to come up with a Rube Goldberg KISS type hook, in the interest of safety one should carefully investigate the regulations and the need for a dual release (normally one electric and one mechanical). Even if one is unsure whether they'll use a hook, by thinking ahead during construction they can leave space for the release switch/breaker and also consider the mounting for a mechanical release. I chose to install the mechanical release and pre-wire the electricals during basic construction and to defer the physical fabrication and mounting of the actual hook until after the basic construction is complete.
Although I don't have a direct requirement for a hook, I had been thinking that it might be interesting to have one so I could eventually practice slinging techniques. I'd been on the lookout for one of these at the right price and I finally found one. I'm not sure of the original source, but it is marked as being rated for 1,000 pounds and from other pictures I've seen, I'm led to believe it may be from a Hiller 12E. I'll definately have to do some fabrication to get a reliable swivel mount with limit stops but I think it can be done relatively easily. For now, this is one of my lowest priority tasks and will probably not even be seriously investigated until after the machine is flying.
In order to meet the regulations, a cargo hook requires two forms of release and in most cases this is done with a combination of an electrical release and a mechanical release. The electrical release is relatively easy to implement and I've used the small finger button on my cyclic grips to implement this (note that due to the current, one will probably need a relay). The mechanical release caused me a bit more head scratching. I've seen this implemented on the Safari both as a pull handle on the side of the instrument pod and as a release lever on the cyclic (similar to a bike brake handle with the open end up). From a safety standpoint, I don't like the pod position since the reason one is usually pickling a load is because of an emergency and this would require either reaching across with the left hand or switching hands on the cyclic. While the lever on the cyclic probably works well and is relatively easy to implement, I don't particularly like it from an aesthetic standpoint. I decided to implement a handle mounted just below the collective that can be pulled with the left hand.
Yes, the "tower" is purposely slanted to give more clearance for the bellcrank
After playing around with various locations, I found the one that I was happy with, but it did cause more head scratching as to how to implement the linkage. Because of the tight bend radius, I felt there was too much friction on a cable that went straight down from the handle and then tried to make a right angle turn to the rear. Obviously I could have gone through the floorboard and right throught the exterior skin at this point. While this would have been easy to implement, the exit hole would not be accessible under the floorboard for future maintenance. I chose to mount a bellcrank bearing on the middle rib such that pulling the handle then results in a pulling action on the bellcrank arm inside the rear compartment. The pull cable can then be routed within the rear compartment and exit in the desired location through the belly pan in the rear compartment and out to the actual hook.
I've seen two primary ways of mounting a hook: a four-point cable suspension and a dual pivot (fore-aft and sideways). This allows the weight of the load to be directly transmitted to the attachment point and prevents the ring at the top of the line from having to slide around on the hook. With the hook mounted in the fore-aft plane, I think the sideways pivot is more important since a fore-aft motion of the line actually becomes just a rotation of the line's ring on the hook. A sideways motion on a fixed hook would be a rubbing action which could lead to failure. I've also seen a hook mounted sideways and this was considered acceptable so long as the hook had a good pivot. Also note that pivoting hooks may need bumper stops to limit the amount of travel ... some form of a bungee cord can also be used to prevent excessive motion when no line/load is attached.
Eventually I'll also need to make up a proper line. Although
certified lines are available, they're expensive for just the
occasional practice use and I'm thinking that I'll probably just
acquire some rock climbing rope and make up my own line. At this
time, I'm assuming that the top end will be just a basic loop
with some kind of anti-chafe ring on it. On the bottom end there
needs to be some kind of hook and weight to keep the line
relatively straight even if there's no load attached. I found the
following new hook on eBay for $10 and I've noticed that the
seller has run his ad many times so I assume he had a batch of
them. In addition to the hook keeper, there is a built-in swivel
to prevent line tangles.
Update: I know a builder who had a couple of short lines built
for him out of 1/4" cable and he was very impressed with
them. The price was considerably less than I would have
anticipated.
One thing I haven't spent a lot of time thinking about is line lengths. Originally I'd been thinking about long lines since that is what I've typically seen in use. The negatives of these, as I see it, is that one is in the middle of the H-V curve and its required to stick your head out the door to see the load (or to get a special bubble door). I know of a couple of Safari builders who are using short lines (8+ foot) and this got me thinking a bit more about it. The advantage of a short line is that one can do the pickup and still be in ground effect. Also, a good convex mirror on the skid should allow the pilot to be able to see both the hook and the load.
I was given the following link for a cargo hook which carries a much lower price than most of the certified ones that I've seen. I have absolutely no additional information or experience with this unit, but I've included it here for reference. I was also given this link for a non-certified hook available in the US. Again, I have no personal experience but I know one Safari owner who has installed one. The electrical connection is straight-forward but they use a MATE-N-LOK style connector rather than the more common circular connector ... I'd recommend replacing it. The mechanical release took a bit of fabrication since the hook doesn't use the more standard mechanical cable and manual trip arrangement. Initial tests with an 8' short line went very well.
Some of the certified suppliers are:
<NEXT> Preparation and Paint
Last updated: August 01, 2009
