RIGGING
GENERAL INFORMATION
BLADE
PREPARATION AND INSTALLATION SEQUENCE
For those owners that have not downloaded it yet, there is a
new Rigging Manual dated 12/01/03 that is available on the secure
area of the CHR website for download. I have seen a couple of
questions pop up that were answered in this document and several
that are not.
Update: Now there is also a 2008 version of this Rigging Manual
but it does not cover the main blades. This manual still has
ambiguities as it talks about checking for degrees of freedom and
then saying to set the final positions well past that. The tail
rigging section still needs work in my opinion.
The more I work with these two manuals, the more disappointed I am at them. While they may describe the ideal situation, they don't provide a consistent and reliable method that can be used on all ships. It would appear that there is enough differences in the factory assembled swashplates (main and tail rotor) that there is significant variations in rigging requirements. I have also had a chance to inspect a factory rigged craft where none of the settings appear to match those in the manuals.
I have gone through the rigging manuals and made the first pass at rigging my various controls. I then chose to enlist the services of a very experienced Safari owner to assist me in the final inspection, setting of the controls plus tracking and balancing which lead up to the first hover. His methodology is different than what is described in the manuals but I must admit that it worked extremely well. His ship also has the lowest vibration readings that I've heard about. Having gone through this experience, I consider it almost mandatory to get such assistance ... it will save a lot of time and frustration with the net effect of getting a well rigged craft in a very short period of time. I have no reservations about recommending the person who did mine ... owners can email me for contact information.
Throttle
Although the Rigging Manuals may not be perfect in this area, I highly recommend that one follows the concept when rigging the throttle. Essentially the instructions are trying to get all the linkage arms to be horizontal or vertical at the middle of the throttle range. This seems to work well. However, I've seen a craft where the rearmost link was too short (contstruction print length) and nearly horizontal at the idle position. This causes it to nearly go over center at full throttle ... more importantly, the link between the two collectives did go over center and contacts the cross tube ... not good. What I find appalling is that this was on a factory built and rigged craft where they chose not to follow their own manual.
The idle speed of a helicopter engine is set much higher than
on a fixed wing aircraft. Since there is no flywheel effect when
the sprag clutch is not engaged, it's important to keep the idle
speed high enough that the engine won't quit when doing a
practice autorotation. I don't remember seeing or hearing
anything specific for the Safari so I'll probably go with the R22
approach where the warm engine idle speed is set at 53% to 57% (i.e.
55% +/- 2%) or approximately 1500 RPM.
Update: My idle is now set to just above 50% and it appears to
work well, both for idle speed and for the main clutch engagement.
One thing I did have to do was to slightly reduce the length of
the standard spring which holds the idle adjustment screw on the
carburetor in position. Without this change, the spring would
fully compress and not allow any further movement to raise the
idle speed.
When I first started my engine I felt that the throttle was extremely sensitive due to very little movement at the grips for full travel. Looking at the linkage, the easiest way to increase the throw was to shorten the moment arm on the bellcrank below the carburetor. I machined up an insert to fit into the next inboard lightening hole in order to accomplish this and it worked well. Unfortunately we did two tests on the next startup ... the throttle sensitivity and the first trial of the governor. The governor was hunting which has since been identified as the LASAR® tach signal being a poor source for it. While the tach signal source has been corrected with a separate sensor, we also went back to the stock throttle hole in order to eliminate that as a source of the problem. At this time I haven't gone back and re-tried the throttle arm change.
Correlator
The correlator will need an adjustment that is unique to each
craft and this can only be determined after the craft is hovering
and ready for flight. Unfortunately, it is a unique combination
of the exact linkage, carb, engine,etcetera. Ideally, once the
craft is in the hover there should be no actual throttle changes
required to do an entire circuit and back to the hover. The only
recommendations I've received is that most builders have had to
use a significant spacer to accomplish this (3/4" to 1-1/4").
I've also been told that this adjustment is well worth the time
and effort to get it right as it makes the workload much easier
on the pilot. I installed a 3/4" spacer as my starting point
and will see how this works out.
Update: After looking at other installations and talking with
other owners, I decided to start with a 1" spacer. I've now
backed it off to 7/8". The reason is that anything much
longer than this will interfere with the seat's vertical rail at
full up collective and full low throttle. I'm aware this is an
unusual combination, but it is one that might be done during a
hovering auto. Besides the obvious paint chipping, a bigger
concern is that the leverage from the collective could cause the
arm to be bent or in some other way be altered. If the correlator
is not effective enough, I'm going to have to very carefully re-examine
this area.
After the initial hover testing, I've gone back to the 1" spacer plus a thick washer (1-1/16" total). I need to carefully review this area for any interference ... last time I checked the interference was only at full up collective and full low throttle which might happen during engine failure at a hover practice.
Control Rods / Swashplate
The 2003 rigging instructions call for having the swashplate level laterally and 2-1/2" or more from the pylon to the bottom of the swashplate spider (the 2008 rigging instructions call for 2-5/8"). It also calls for the swashplate to be level in the fore-aft plane, the lower knuckle to be level or slightly angled up and the upper knuckles to be at 18 degrees. This can be a pain to try set the control rod lengths while everything is flopping around so I came up with a simple plan. I used four short equal length posts between the swashplate and pylon to hold everything in it's final position while I finished fabricating the rods and adjusted everything. A bit of trial and error produced posts that were 2.76" long (note that I later lowered the three links by two turns). I have a lathe to make these kinds of posts but alternatively, four AN4-25 bolts would have been relatively close and a reasonable starting point. If a cyclic control lock that accurately centers the stick has been fabricated / installed, it then becomes a very simple task to adjust the rod ends on the three control rods.
Cyclic & Collective Control Stops
The collective control stops are relatively easy to adjust once the swashplate has been leveled. They are just set to allow 1" of vertical movement of the swashplate.
I found that setting the cyclic stops was an iterative process as I tried to find the optimum height of the swashplate that gave maximum movement without binding. While I could get the lateral movement that the manual(s) call for, I had trouble getting the fore-aft settings. In the rearwards direction, the cyclic lock interferes with the cyclic stick before it reaches the full 9 degrees. With the rear stop set just before the cyclic touches the lock, the lock still interferes at the extreme lateral positions. However, full aft and full side cyclic is not an operational position that I can envision! In the full forward direction there is a binding in the swashplate at full down collective before I could obtain the specified 12 degrees of swash movement. This goes away as soon as the collective is lifted a bit. In talking with several builders of operational Safari's, they stated that 12° was not required and one builder ended up with only 7° in all directions. I set my forward collective stop so that it is just starting to get to the interference point at full down collective ... again, a flight condition that I would not envision. After setting the stops, my final swash angles were:
Lateral: 9° left and right
Forward: 9.6° (12° in 2008 Rigging Manual)
Rearward: 7.7° (9° in 2008 Rigging Manual)
After going through the manuals and doing the above, it was interesting that the person helping me has a different way of setting the controls when rigging ships. The first major change is that he set the neutral cyclic position such that the swash had 2° of forward tilt and 2° of right tilt using the control rods. There were also some other changes that I need to verify. Again, we were playing right on the edge of the interference that appears to be caused by the spherical bearing in the swash assembly. From what I wrote down and have with me, the final results based on cyclic stick angles were:
Forward: 20° (swash angle 9°)
Rearward: 13° (swash angle 5°)
Left: 25°
Right: 18°
Update: After doing some more hover testing it was decided to remove the 2° of forward tilt in the swash at neutral. Both myself and an experienced Safari pilot felt that there was not enough back cyclic available when hovering. Part of this may be due to the cyclic interfering with the seats rather than the metal of the seat supports. Regardless, it is an aspect that must be considered if contemplating this change.
The various stops for the cyclic (8) and collective (2) each use a bolt and a jam nut to set the stop and lock it into position after adjustment. This kind of arrangement does not provide the two levels of securing that is normally used with aircraft bolts. This same style of stop has been known to loosen and result in several incidents on Rotorway machines. One could possibly use a toothed lock washer under the jam nut or removeable thread locker (normally blue). I used blue thread locker on these stops ... first they were adjusted and then the bolt and jam nut were carefully backed off to apply the thread locker followed by final installation and adjustment.
Lead / Lag Adjustment
From an engineering / maintenance standpoint, I really don't like the whole lead/lag approach used on the Safari. This is described as opening five of the six holes in the cheek plates that go into the main grips and using the sixth bolt as a pivot point. Any removal of the grip bolts (i.e. to remove a blade) will require that one go through the whole lead/lag process all over again. Removing just one blade might not be as bad since one might be able to do the CofG over the center point during re-alignment. If the attachment from the blades (i.e. cheek plates) to the main grips were precision holes/bolts and the lead/lag was performed by using the bolts on the blades (or some other mechanism) then it would be a trivial process to remove / re-install a blade. More importantly, one could try different blades simply by getting a second set of cheek plates without having to continually redo the lead/lag process.
Before actually opening up the holes in the blade grips, I'd
highly recommend that one do an initial lead/lag check. There is
a slight amount of movement available with just the standard
holes / bolts and if this is sufficient then there's no need to
possibly screw up the cheek plates and/or grips and make oversize
holes. I'm also aware that one builder received cheek plates
which had some undersize holes on the grip ends. Before cementing
these plates to the blades, all of the hole sizes should first be
checked ... it's a lot easier to work on these parts separately
on the bench before they've been attached semi-permanently to the
blades.
Reference: I've been told that it's normal to be able to get
about 1" of lead / lag at the blade tips without opening up
any of the holes. This is accomplished by disconnecting the pitch
links and rotating the blade(s) such that just their weight will
move the position in the cheek plates due to the tolerance of the
bolts and holes.
After looking at the lead / lag technique some more, in the event that one cannot obtain enough lag with the described technique I think there is another alternative to try before opening up the grip holes. The "pivot bolt" shown in the Rigging Manual is the outermost trailing bolt. Assuming everything has been drilled accurately at the factory, the maximum amount of pivoting on this bolt is based on the clearance of the bolt furthest from it. If there is not enough movement, then one could try pivoting on the center bolt beside it which could allow nearly twice the movement as this would rely on the clearance of two bolts but a shorter pivot arm distance.
My blades were first "strung" and then also checked
with a laser off the head. Both techniques worked well and there
was very little difference, if any, in the result.
Update: The blades were re-checked after
about 10 hours of engine time as we were trying to track down a
source of unexpected vibration. One blade was leaded by ~3/8"
while the other was still perfectly aligned. I don't know if this
was due to a shift during operation or the result of trailering
the craft.
Pitch Links
The tracking manual indicates that the blades should be set for 1/2° of negative pitch. I've been told that it is better to set this to 1° negative as it will provide for better autorotations. One also wants to be very careful if they do any tracking adjustments so as not to remove this negative pitch.
One also needs to remember that these have two rod ends with opposite threads and be very careful when loosening and tightening the check nuts. A felt marker gives a reference as to where the previous setting was. Each link should have some kind of a marking to designate which blade it is paired with ... if one ever takes these links off they want to replace them onto their respective blade.
Dynamic Balancing
Obviously this is way down the road for me, but I have been doing a little research into this critical issue, the equipment required and the methodology. So far I have not been able to find good general information on the procedures The documentation for the Safari can be very confusing as it describes various techniques and not an orderly methodology. Confusing the issue even more is that various experienced Safari personnel and owners use different techniques. There is some good information available for balancing Rotorway Execs and also some very knowledgable people.
Before going further, it should be clarified that there is both a static balance and a dynamic balance required. The more I read, the more it appears that the static balance is extremely important to resolve before attempting any dynamic balancing. If the static balance has not been fully solved then it will be impossible to obtain a good dynamic balance at anything other than one specific RRPM and forward speed. Everything in this process is a compromise, but starting without a good static balance is just asking for trouble.
As far as dynamic balancers go, I'll just add some links and observations as I don't have enough background information to make an informed opinion. These are expensive units to purchase but there is also the option of possibly making arrangements with CHR to perform this function or using a commercial helicopter company to do it. At this time I also don't know what a recommended interval would be to re-balance the ship, other than when one feels a higher than expected level of vibration. If one was to purchase their own unit, they should also make sure they have the appropriate accelerometers, sensors, cables, calibrators etc.
There are several other manufacturers of balancers (such as ACES) as well as various military surplus units. I have not seen any references to using these on amateur-built machines so I can't comment any further except to say that it would appear that a user of this type of equipment could quite possibly be totally on their own when using such equipment. It would also appear that Pro-Drive is distributing the DynaVibe balancer ... I have no further information on this unit.
One very interesting thing I noticed at Sun-n-Fun '08 was the G3 Engine Monitor from Insight Avionics. Besides the obvious engine monitoring and logging, the unique thing about this instrument is that it has vibration measurement and analysis built into it. In conversation with their personnel, it would appear that other people with helicopters had also noticed this. Since the president of this company also flies helicopters, they were very enthusiastic about exploring the options of creating a helicopter vibration monitor based around their technology.
Tracking
CHR's website now has a press release dated March 30, 2005
indicating that they have a new product available to assist in
performing blade tracking. In essence, its a pair of high
intensity LEDs of different colors, one for each blade end, which
will allow for setting the blade tracking without the use of flag
sticks or strobe lights. They are listing the price as $90US for
a pair of these units for a two bladed system. At this time, I
have not seen the physical units. I happened to be in contact
with another supplier and got a pair of tracking lights that go
under the logo of "E-Z Trak". They seem to be a bit
smaller, lighter, cost less, use only two cells instead of three
and have a feature where the centrifugal force from the blade
motion can be used to turn the LEDs on/off. Once I've had a
chance to use them then I'll be able to report on their
effectiveness. Another option for tracking lights is available
from Rotorway
Australia and supposedly they work very well in daylight.
Update: The E-Z Trak lights were tried on a
relatively bright day and the results were less than stellar ...
in fact they weren't useable and we chose to immediately remove
them. I've received information that the RW Australia lights are
much brighter and still useable in bright sunlight.
Note that because of reflection the battery tubes look longer than they actually are.
One interesting aspect of in-flight tracking is the rotational position of the blades relative to the fuselage. Because of the pilot's sight picture, it is normal to look at tracking lights in the fore-aft position. However, it is also important to look at the lights while the blades are in the sideways position (especially right side) where the effects of forward speed will be most noticeable.
Reference: I've been told that one turn of the pitch links is equal to a 1/2" change in the tip path.
Tail Rotor Rigging
The first problem I ran into with the tail rotor rigging is with the mounts for the cable. The rear mount was just useable after some bending but the front mount was too far away from the pedals and did not allow the cable assembly to be connected. Rather than cut and re-weld a painted frame, I ended up changing from a female to a male rod end and making a spacer.
In order to get the "slight tilt forward"
of the pedals, mine are essentially centered vertically over the
forward pedal tube and I think that's about as far forward as
they can go without changing the pedal stops. For the left pedal
stop (on the right side of the craft) I ended up using an AN4-6A
bolt with a lock nut whereas on the right pedal stop I ended up
using a fully threaded AN4H-4A bolt with safety wire and no lock
nut.
Update: After final rigging, both of my pedal stops are now using
a fully threaded AN4H-4A bolt.
I have read both the 2003 and the 2008 versions of the Rigging Manual about how to set up the tail rotor cable. In short, I do NOT agree with the method that is described. The described methodology assumes that the tail rotor assembly links have been accurately and consistently pre-adjusted at the factory ... based on the quality of work I've seen, I do not accept this assumption and my assembly could not be rigged per the described procedure. There are no dimensions, reference criteria or angles specified. Every other helicopter rigging manual that I've looked at gives the blade angles at full left/right pedal positions and other critical dimensions that are required to rig the tail rotor blades. That is the only way I know to consistently and reliably rig tail rotor blades.
I am just in the process of collecting actual angles and dimensions from owners of running Safari's. I would welcome and appreciate input from all owners who can provide this data. While the factory may not want to publish this information, I think it is required and will try to tabulate the values I have available. As delivered, my tail rotor had -2.1 to +31.5 degrees relative to the main mast without the cable attached and -2.2 to +24.1 with the cable attached. That is before reducing the range slightly when adjusting the pedal stops.
Information I've received about
other craft:
-4° to +31° - Currently flying
-5.4° to +33.3° - Factory rigged craft with no allowance for
cable overtravel
-7.9° to +24.4° - Set per Rigging Manual(s)
~ -3° to ~ +27° - Currently flying
During the final pre-hover rigging we looked at this area again. By various adjustments of the links, we were able to gain considerably more movement of the blades. Using the same measurement technique as before, I now have about -5° to +34° of movement. It should be noted that due to the Delta-3 hinge one must be very careful when adjusting these links to keep the blades at the same angle. I made a jig and used thin shims to lock the spindle in place when doing my testing. I also used a piece of angle stock temporarily tie-wrapped to the pedals to lock them in place.
Interesting field report: I have received information from another builder whose tail rotor blades measured 28° and 30° relative to the mast. He then changed this to be 33° for both blades. As a result, he reports a very significant difference and its much nicer to fly. The one thing I forget to ask him is how he determined "relative to the mast" ... the delta pin arrangement makes it somewhat difficult to get a true starting point. I believe his results may have been more from the increase in maximum pitch.
One thing I'd noticed is that the outer links have a LOT of movement which allows the slider cross to rotate through a fairly large range. This doesn't effect the pitch setting but I can see how it's not required and would quickly lead to wear patterns in the slider cross. The person assisting me with the rigging agreed fully and told me how he usually adds some kind of fiber washer in here to act as a stop and prevent wear. Since we couldn't readily find the appropriate size of fiber washer, I made up washers from UHMW sheet material approximately 1/32" thick. This definitely restricted the link movement a bit and got rid of the metal-on-metal contact.
Yellow lines point to washers
Before trying to balance the tail rotor, I think it's very important to check the blade lead/lag. In the Owner's Area of CHR's website there are drawings for "Tailrotor blade beam" and "Tailrotor blade jig Body" which create a tool that can perform this function. Although the tailrotor assembly comes pre-assembled from the factory, we discovered that one of my blades was significantly leaded which caused us a LOT of balancing problems. More importantly, since there is no lead / lag adjuster it caused us even more problems trying to get rid of the leading. How this blade got past any form of quality control when it was manufactured is a different and much better question.
Weight and Balance
I have heard from several builders with the O-360 engine who performed their W&B and found their ships were slightly tail heavy and had to make changes like relocating the battery to under the instrument pod. Until I get to that stage, I can't make first-hand comments but this seems to be a recurrent problem that requires extra dead weight (i.e. bigger batteries up front or shot bags) to correct. In the meantime, I've included this information so other builders can be very careful not to add additional weight behind the CG. I do have an idea for a light weight horizontal stabilizer that would make a significant difference due to the long moment arm.
The procedure for calculating the weight and balance is contained within the Pilot Operating Handbook. Like most of the Safari documentation, this is subject to interpretation and is confusing. The datum line (i.e. reference) is listed on page 15 as being "91 inches (231,1 cm) forward of Aft Skid leg (station 91)". Interesting, but according to the plans, the rear legs are just in front of station 76. It's also not clear to me whether this is referring to the centerline of the skid leg or the forward face of the leg ... since the sockets are 2" diameter and one is only dealing with a 4" total CG range (2" at gross weight), the distinction is rather important.
I thought that perhaps the cargo hook longitudinal calculation could clarify the above since it's hardpoint is fixed along the rear edge of the tube at station 52 (i.e. station 52-3/8). No such luck as the closest I could infer would still be 5/8" off of the cited references. The more I thought about this, the more I realized that the skid legs are not the best place to measure the CG from ... it's quite possible that there is a slight twist in them from the welding and bolting operation. Instead, I chose to use the center of the tube that makes up the rearmost frame (STA. 76). This should be a repeatable location on all ships. With the main mast vertical, a plumb bob was dropped from the frame and the position moved 3/8" to compensate for the diameter of the tube (7/8").
Perhaps I'm brain dead as I'm writing this, but the lateral CG calculation on page 34 of the POH is beyond my comprehension: "The lateral CG position equals: (R scale reading - L scale reading) x 1/2 Distance between skids - (R scale reading + L scale reading)." Besides the fact that this appears to be an erroneous equation, it is using the skids as a reference ... obviously this assumes perfect welding and alignment such that the skids are exactly equally spaced from the mast centerline.
Update: I found the equation that CHR copied for their POH and they obviously made a copying error ... the second minus sign should have been a division sign. At least in the original document the authors refer to "rotor centerline" for the datum reference ... a lot more reasonable and accurate.
N.B. The CG and weighing process should be done with the craft as close to flight ready as possible. I included flight manuals, oil, etc. but no doors, fuel or ballast weight for my initial checks.
As to how to measure the CG location: I chose to try a method that Rick Reese discussed ... basically putting the skids on a piece of tubing and rolling the craft forwards / backwards until the balance point was achieved. This worked extremely well and is much simpler that what is described in the POH ... thanks for the tip Rick. I used a couple of small pieces of 1/8" steel plate to protect the skids and a piece of 1" heavy wall tubing under the plate. The craft was jacked on the wheels to locate these parts and a mark was put on some masking tape once the balance point had been determined.
Once the CG had been marked on each skid, a long T-bar was placed across the two skids and a measurement was taken from the middle of the bar between the skids to the plumb bob off the rear frame to give the CG location. Using the bar this way compensates for any misalignment in the two pieces of tubing relative to the frame. In my case, the measurement was 12-1/8" or 78-7/8" relative to the datum (91" ahead of rear skid leg) and without the doors installed. Measurements I've heard from other builders are: 78-1/8", 78.6", 78", and 78.5 or 79". Obviously my craft is tail heavy relative to most.
The next step was to weight the craft. After zeroing the scales, we put one scale under each skid under the CG mark. The empty weight including flight manuals, oil, etc. but no doors, fuel or ballast weight was 1081 pounds. This is about 40 pounds heavier than other crafts that I've heard weights on. So where did this weight come from? I can think of several places where a couple of pounds here and there were added: bear paws, a painter who layed the paint onto the boom and frame VERY heavy, paint on interior hidden panels, confor foam in the seats, a REAL strobe power supply, COM plus GPS/COM, extra adel clamps to make everything secure, tip weights in the blades, etc. etc. The moral is pretty obvious ... try to be conscious of weight buildup at all stages of the build process.
Ballast Weight
This is one feature of the Safari (and also the Rotorways) that really bothers me. The simple question for the kit manufacturers is "Why does no current certified two-place helicopter require a moveable ballast weight?". Obviously there is a design issue that is being corrected with a Rube Goldberg solution and we have to waste 1-2 horsepower just to allow for it. The biggest negative I see is that sooner or later the pilot will have the weight in the wrong position and will encounter handling difficulties. A good example is when one lets a passenger out and then wants to continue flying without shutting down ... oops, the ballast weight must be moved first. I know I have done this type of thing many times in an R22 where an instructor either got out or in. This is one item that the pilot must pay particular attention to when they first lift off and do their hover CofG check. It's also noteworthy that at least one Rotorway owner has added sensors to the ballast weight holders and under the passengers seat along with the electronics to indicate an error when the weight is in the wrong position.
I would like to make this item as lightweight as possible and before filling it with lead I want to have a full weight and balance report for my unique craft that includes all fuel levels and the full range of weights for passengers.
Although one can purchase new lead metal material, it can usually be found at a friendly tire shop as removed tire weights. Alternatively, dive shops normally have a source of lead for use in weight belts. Since lead melts at a very low temperature, it can be melted with some kind of torch or I've even seen it being done on a Coleman stove ... note that whatever "pan" is used for this should not be used for foodstuff afterwards. If one is doing this melting indoors they'll want to use a well-ventilated area and will also want to use a good respirator ... my experience has been that it is quite noxious and should be done outdoors.
One issue about the ballast weight is whether to fill it with lead before or after the holes for the locking pin has been drilled. I chose to pre-drill the holes and verify that they lined up with all three of my mounting locations. The question then became how to make sure that the molten lead didn't drip out the pin holes ... I chose to use a piece of brass tubing (available at hobby shops) and a scrap shop bolt to prevent this. However, I forgot to allow for the contraction of the lead as it cooled. Because the upper part of the lead was captured by the brass tube, this resulted in about a 1/32" gap between the lead and the bottom of the tube. Since I chose to fill the main tube just enough to cover the brass tube, I'll probably try remelting the top part of the lead around the brass tube to allow it to re-settle. If I was doing this again, I would fill it to just below the brass tube, let it cool, re-melt the exposed part of the lead and then top it up with more lead.
Horizontal Stabilizer
I don't remember seeing any reference in the rigging
instructions for the initial angle of the horizontal stab. I do
remember hearing that 7° leading edge up was a good starting
point. I have been in contact with one builder who has been doing
some flying with a high-time instructor and they chose to change
the stabilizer from +7° back to horizontal. This got rid of some
of the nose-down attitude in cruise flight and in addition to
picking up significant cruise speed, they also felt that it made
the ship easier to handle in rough air. After even more testing
and flying, they have now set the angle to -5° (ie. leading edge
down) and feel that it is even better, although they have lost a
couple of MPH in cruise speed. I'm sure that each Safari is
slightly different, but a 12° change from the starting point is
significant and other owners may want to consider whether they
should consider this adjustment as part of their flight rigging.
Update: I've now heard from a couple of other builders who have
experimented with the angle on their horizontal stab. It would
appear that the +7° position (roughly
the angle along the top of the tail guard hoop) definitely
results in a nose-down attitude in cruise and that a closer to
horizontal position is preferred. The downside is that it will
require more forward cyclic in cruise to attain the same speed.
Contrary to a previous comment, another builder found that the
more nose down attitude resulted in much more of a tendancy for
the craft to want to cruise at a higher speed. What's probably
more relevant is that this attitude can be achieved by either the
horizontal stab position or by the CofG position ... given the
choice, I believe a CofG shift is safer than using aerodynamics.
For now, I've set the angle at 7° until the craft is actually
into forward flight where I can experiment more with the setting.
Using a modelling wing incidence jig to set the angle
One concern I have heard about adjusting horizontal stabilizers is the effect it has if one ever encounters an engine failure. Raising the leading edge tends to lower the nose in cruise flight and reduce the amount of forward cyclic that is required. This balance is achieved because of aerodynamics and the forward thrust from the rotor system. However, if this thrust is suddenly removed (i.e. engine failure) there is a whole new set of forces at play and it is possible for the craft to tuck or do some other drastic reaction unless there is an immediate cyclic input. It is my understanding that if one is playing with horizontal stabilizer angles then it should be done in small incremements and thoroughly tested at each step. The testing in cruise flight with sufficient altitude should include progessively faster throttle roll-offs leading to a throttle chop to determine if there are any major reactions. Also note that there will be a different reaction when flying solo versus dual.
I've been told that the Safari has a mild hunting tendancy in cruise flight when compared to the R22. For those that are into aerodynamics and engineering, they may want to read about and consider the testing of a Gurney flap. I am aware of a JetExec where these were installed and were the cure for a porpoising condition in cruise flight. In this case, after installation they allowed for a very significant increase in cruise speed within the smooth handling envelope. An introductory article on these can be found here (Part 1) and here (Part 2). If anyone has done any testing with these, I would appreciate receiving any information about the results.
2/Rev Vibration
There has been a lot of talk about how many ships are experiencing a 2/rev vibration that seems to be accentuated when the fuel level gets below 1/2 tanks. From what I gather, the vertical component is most noticeable at the front of the passenger side floorboard which is the cabin area farthest from the mast and has the least triangulation. There have been several modifications that different builders have tried with varying degrees of success. From what I gather, no one has fully identified the source of the vibration and how to totally eliminate it before it enters the frame.
One of the most dramatic 2/rev vertical eductions has been achieved by Claude Lescure who has been experimenting with a tuned resonator. The installation of this device has caused this vibration to go from the 0.9-1.3 ips range with less than 1/2 tanks to the 0.1-0.2 range. This is a VERY significant reduction and I'd like to thank him for sharing this information.
I'm waiting to hear of significant reductions in the 2/rev lateral vibrations. Of even more value would be if someone can fully identify the source within the mast and come up with a relatively simple change that could eliminate, or at least reduce, both of the 2/rev vibration components.
One builder that I'm in contact with says that his 2/rev vibrations are now nearly constant at all fuel levels. In order to better his weight and balance, his battery (Concorde) is mounted at the front of the craft below the instrument pod. He has also chosen to go with a single cross-brace and a spring arrangement between his two fuel tanks.
It is also interesting to note that I've been following a Rotorway forum and one of the things that has been discussed is the use of composite vs. metal blades. The Rotorway metal blades are non-symmetrical and the reflex on the trailing edge is adjusted for cruise tracking among other things. The consensus of contributors seems to be that the composite blades can be harder to find the sweet spot and that there will always be trade-offs, especially if there are no trim tabs. Their hypothesis is that this also contributes to the Safari 2/rev vibration as it cannot be easily trimmed out via actual blade tweaking. Note that the March 2007 issue of Experimental Helo has an article describing the implementation and use of trim tabs on Safari blades.
There is another interesting hypothesis that has
been floating around with regards to the 2/rev vibrations. If one
looks at the Young's modulus for titanium versus various grades
of steel, it is apparent that titanium is more flexible. Since
the main mast and spindle are both made of titanium, they could
be allowing a bit of flexing. Based on their leverage, even a
very small amount of flexing would cause significant vibration to
be transmitted to the frame. I'm curious if anyone has ever tried
a main shaft made of a high grade steel and what the comparative
results might be. Regardless, this just reinforces the importance
of getting a good track and balance on the blades.
I am now aware of an owner who is rebuilding his transmission and
has investigated changing the main shaft to 4140 steel. The
primary catalyst for this was when he found that no relief radius
had been machined into the steps and thus could cause stress
risers. Since this machine has flown with a titanium shaft it
will be very interesting to hear what his opinion is on the steel
shaft if he goes ahead with that change.
Trim Systems
I was thinking about the implementation of a trim system and also reading through the old forum posts. I have installed a spare hat switch to allow for the future addition of an electric trim system. Although I have a good idea of how I'm going to implement this, it hasn't been built yet so obviously there are no pictures available and I'm not planning to publicly post any details until I've at least done a prototype implementation.
In the old posts there were two interesting references to lateral trim systems. One was the use of a servo actuator that pulled on a rubber strap and the other described a change to the rigging by changing the angle of the swashplate so it is leaning down towards the pilot and also leaning forwards. I do not have first-hand knowledge of these changes and it's only included here for reference.
During the initial hovering we added a spring to compensate for the side force on the cyclic. This will probably be removed at some time for a more permanent trim system but for now it gives a very neutral feeling cyclic.
Yarn Yaw Indicator
The cheapest instrument on the craft is a piece
of yarn on the front center of the bubble that is held in place
with a piece of tape similar to what is used on Robbies. When
learning to fly, I was originally much more inclined to use the
ball as a yaw indicator. However, with increased flight time and
different situations I learned that the yarn is much more
sensitive and can be used to indicate things that other
instruments don't show. The only lesson with the "string"
learned the hard way is not to pay much attention to it if flying
in the rain as it tends to stick.
Trivia: As far as I know, the yarn is on the Robbie's Minimum
Equipment List (MEL) which means that technically one can't fly
the craft unless it is present.
Miscellaneous
For those that like to read more about blade balancing, there is some interesting information here.
BLADE PREPARATION AND INSTALLATION SEQUENCE
There are several sources of information for this critical process. Since some of it is contradictory and since each of the factory personnel has a different method of accomplishing this task, it makes it extremely confusing for the owner. I've tried to go through both the written material, the experience of other owners and [in]direct comments from factory personnel in order to come up with the sequence that I will be following. I realize that this creates yet another procedure, but this is for my own reference and I'm choosing to make it publicly available ... it's also still a work in progress. If anyone can find fault, simplification or other comments on my sequence then I'd like to hear this and will update this section as appropriate.
- Blade Inspection
I have no idea why the Rigging Manual talks about changing the profile of the blades. With the price of these blades and their construction method, this is something that I would fully expect the factory to have performed. To my way of thinking, if the blades need to be re-profiled then there is a major manufacturing problem.
Update: I actually did reprofile the tips on my swept-tip blades as there was a considerable difference both in the blade length and their contour. Hopefully this should not be a problem with square tip blades.One thing to double check at this stage is the root thickness where the cheek plates attach. I am aware of one builder's blades which weren't exactly the same and this caused problems when trying to attach one of the blades to the grip. If the root is too thick then it should be a simple matter to remove a bit of the gelcoat using a flat sanding block. If the root is too thin then it will need to be built up a bit using some kind of shim material, possibly composites.
- Static Balance - General
From what I've gathered, I believe the static balance is an extremely important step ... if the blade system is not accurately static balanced then dynamic balancing must compensate for both the static imbalance plus any dynamic imbalance. This is a relatively straightforward process and there is nothing mysterious about it. While some others dispute it, I also believe another critical point is to try get the center of mass for the two blades as close as possible before moving ahead to dynamic balancing.
I think the biggest problem in this step is to obtain accurate weights. Most of the affordable scales I've seen in this weight range only have a 0.1 pound resolution (~ 5-1/2 AN970-4 large washers) ... that can cause a LOT of discrepency when one is trying to balance blades. Much higher resolution scales (i.e. 0.005 pound or a 3/16" thick elastic stop nut) are available, but they are quite pricey and probably not available to most builders.
Because of scale inaccuracies, I believe that individual blade weight is just an interesting number but not a reliable starting point. Thus a different technique is required. Of course, it there are gross differences in individual blade weights then it is indicative of a manufacting error and the factory should be consulted.
Within the accuracy of the scale I used, my blades appear to be relatively close. They were both right on the edge of where the last digit on the scale was changing ... a small weight would make both of them stabilize on the higher digit. Basically close enough to proceed.
- Mark Rotational Center on Teeter Block
Since this is going to be required, one may as well get this out of the way before any mounting of other parts. Note that it is very important to make this mark as accurate as possible as any misalignment will cause extra work later on.
The Rigging Manual describes this operation and while I fully agree that the mark is required, I'm not impressed with the technique. It relies on the accuracy of the builder to measure and scribe lines followed by an exact center punch. The reference for these lines is a bevel and I've seen numerous other places in the kit where the machining tolerance left a lot to be desired. Even a .010" error in this mark will introduce a significant error to a perfectly "strung" set of blades. I believe that this reference mark is something that should be made during the machining operation on the teeter block as soon as it has been squared and centered.
Swept Tip Blades
Owners of the older swept-tip blades should determine at this point whether they are going to add the tip-weight modification. The instructions are available for this modification on CHR's website and it is imperative to assume that equal weight (bolts, epoxy and shot) will be added to to both blades. However, there is no guarantee that this weight will be at exactly the same position spanwise due to the internal construction of the swept tips.
The instructions show the bolt holes relative to the blade tips but since my blades are unequal length, I averaged the position. I also made sure that the bolts are exactly the same distance from the outer strap hole.
If there is a difference in the CofG of the blades, a slight change in the tip weights may be able to compensate for this. Thus I think the first step of adding the tip weights is to get accurate CofG marks (in pencil) and compare them. If they are very close then one can proceed with adding equal tip weights. If they're not, then one can consider adding slightly different amounts of tip weight to try remedy this ... I'm not going into details here as one also needs to consider total blade weight.
The as-delivered CofG on my blades was within 1/16" spanwise and within the accuracy of my marks chordwise.
- Painting
If one is planning to paint the blades in any way (either the whole blades or stripes) then it should be done at this stage so that the weight of the paint is accounted for when trying to match the blades.
The other thing to do at this time is to clearly and indelibly mark the two blades in order to distinguish them (usually referred to as Master and Slave but also as Target and Blank in CHR parlance). I plan on using a small bit of red and blue paint on the roots to accomplish this. The main grips should also be likewise marked so that dis/re-assembly always results in identical pairing.
- Dry Mount Cheek Plates to Blades
The first thing to check here is that the holes in the cheek plates are the proper diameter ... I'm aware of one builder who had a couple of holes that were considerably undersize and all of my holes needed to be slightly reamed. Another thing the builder can do at this point is weigh all the cheek pieces if they have an accurate gram scale. Any discrepency in the parts can be used to help compensate for a spanwise imbalance in the blades or to just cancel out any weight difference in the various cheek plate parts. The various parts should be marked so they can be consistently re-assembled and then dry mounted to the blades.
Before moving ahead to static balancing, the blades with the cheek plates should be trial fitted to the grip. Any interference with the butt of the blade must be resolved by removing material from the root of the blade and this will alter the static balancing.
Mark Blade's CofG
The Rigging Manual describes how to determine the blade's CofG using a straight edge. At this point I would only mark the CofG in pencil. What the manual doesn't talk about is how to resolve any differences. Since it's usually not adviseable to remove material from the blades, the resolution is normally to add weight either at the tip, root or both on the lighter blade. Spanwise imbalance is relatively easy to calulate and compensate for, but a chordwise discrepency will be more difficult. The calculations are basically the same as used in aircraft weight and balance calculations; i.e. weight and moment arms.
This is the time to make any weight changes that are necessary to try get a matched pair of blades, both in total weight and CofG.
- Complete Static Balance
The Rigging Manual describes how to mount the complete blade assembly on parallels so that it will pivot on the teeter pin. This is the final and most important step of the static balance to ensure that the entire assembly is balanced spanwise. I would hope that if the individual blades have been carefully matched then there will be very little adjustment required at this step. The adjustment should only take a very small amount of weight at the light blade's tip. If it takes more than just a small screw then one might want to look at using weight at both the tip and the root in order to preserve the blade's CofG. One also needs to try maintain the chordwise CofG when adding any weight.
- Final Mark of Blade's CofG
While it is another step, the cheek plates can be removed from the grips and the final CofG determined. This is the time to permanently mark the actual point on the blades and hopefully this is the identical spot on both blades.
- Blade Tape
Although it's probably better to wait until the entire balancing process is complete, I plan to install the initial blade tape at this point. While any discrepency in the weight / position of the two tapes will cause a balance offset in the initial setup, I think it will be much easier to do the first application of the tape on the bench rather than on hung blades with access via a ladder. Since this is before dynamic balancing and on the bench, care should be taken to try use exactly the same amount of tape on each blade and position it at exactly the same place.
- Wet Mount Cheek plates to Blades
This is described in the Rigging Manual. Note that the manual calls for 20 foot pounds (240 in-lbs) of torque ... I need to double check the type of bolts and nuts provided, but this is way above the AC43.13 guidelines of 100-140 (max 225) in lbs specified for a dry AN5 bolt with castellated nut.
Update: The bolts supplied in my kit were the Supertanium type which are much stronger than AN5's.
- Install Blades onto Main Grips
If doing this on the floor or bench, it should be noted that the main spindle has a pre-cone angle in it. This will require one or both blades to be propped up while trying to assemble the cheek plates onto the grips.
Note that until the lead / lag adjustment is complete there will probably be a need to loosen / tighten five of the six nuts.
- Weigh Complete Blade Assembly
If one doesn't have a scale with the range to measure the whole blade assembly (~100 pounds) then the grip assembly's weight could be taken at this point along with the weight of the individual blades. In this case, a scale with 0.1 pound accuracy is fine and the reason for getting the total weight is to create an alternative to head shifts ... more later on.
- Mount Blade Assembly
Next up is to mount the blade assembly (two blades plus teeter block and yoke caps) into the yoke ... finally it will all look like a helicopter. I've been forewarned that one has to be very careful at this point as the blades are free to rotate without the pitch links attached. This is definitely not a one person job unless one has a some kind of a hoist mechanism to hold the entire blade assembly as it is being installed into the yoke. Safety wire between the pitch horns and the head shift bolts will help prevent rotation.
I've done this with four people (one on each end of the blade assembly and one on each side of the yoke) and this worked quite well even without locking the pitch horns down. Note that one needs to be VERY careful to have the yoke caps perfectly aligned and squared before trying to drop or tap them into the main yoke.
- Initial Head Shift
At this point one is trying to get the head perfectly centered with the head's center of rotation as starting point. This can be done either with feeler gauges or the tool described in the Owner's Area of CHR's website. Unless the head has been accurately centered one will be fighting it during dynamic balancing. Note that one should also double check the yoke and caps to verify that they are accurately milled and perfectly centered.
- Lead / Lag
Various methods are described in the Rigging Manual on how to perform this. Besides a stringing operation, the other methods basically all rely on the use of a laser pointer referenced squarely to the feathering axis that is used to measure any lead or lag in each blade. I'm also aware of what's supposed to be a simpler jig to hold the laser pointer than what is shown in the manual ... I'm trying to get more information and a picture of this.
Note that the manual calls for 50 foot pounds (600 in-lbs) of torque ... I need to double check the type of bolts and nuts provided, but this is way above the AC43.13 guidelines of 160-190 (max 390) in-lb specified for a dry AN6 bolt with castellated nut.
Update: These are Supertanium bolts which are much stronger than normal AN bolts.While some of the dynamic documentation talks about lagging a blade, I would like to keep the blades as much in the zero-zero position as possible. While the lagging may compensate for an inaccurate airfoil, if it used to compenstate for weight or head shift errors then it will also move the center of lift.
- Tail Rotor Dynamic Balance
Before getting too carried away with the main rotor tracking, balancing and testing, it's time to perform a dynamic balance on the tail rotor. With the main rotor in close static balance, it is extremely important to try get the best possible dynamic balance on the tail rotor. An out of balance tail rotor can cause major damage ... until one sees the results of an out of balance tail rotor it's hard to believe.
Because of my leaded blade and all our attempts at trying to correct it, I saw a lot of variation in the vibration levels we saw. The final reading we obtained at this time was 0.13 ips at 223°.
- Hover Track
It's then time to start on the dynamic rigging on the main blades ... first up is to track the blades. This can be done either with the flagging method just before getting light on the skids or via tracking lights. I think the tracking lights are much safer and have the advantage that they can be used by the pilot alone without an assistant. We used the flagging method on my blades and it is also very effective, especially if done by those familiar with the method.
Because the balancer was hooked up, it is interesting to see the relationship between tracking and balancing. As the track came in closer, the vibration also lessened until the final track adjustment.
- Dynamic Balance in Hover
With the blades in track, it's time to perform the initial dynamic balancing. After doing the hover tracking, my main rotor came in at 0.14 ips at 63°. It was decided to leave it here during the initial hovering ... there will surely be some adjustments made that will impact this once the tracking and balancing is performed for forward flight.
- Hover Testing
For new ships, this is the time where the twenty hours of hover testing and transmission seating is performed. Obviously one can continue to work on the dynamic balance, but the interaction with cruise track / balance can't be determined for quite a while.
- Cruise Track
- Cruise Dynamic Balance
- Head Shift
The objective here is to try get the CofG of both blades to be perfectly in line with the head's center of rotation. This is where the "stringing" operation comes in. If any discrepency is noted, there are two options to correct it:
1) Head Shift
I think this is probably a good technique to use if the initial "stringing" indicates an alignment error. However, it is also dependant upon the accuracy of the various marks. While it's probably a good thing to do for the first pass, I don't plan on doing this repeatedly.2) Balance Weight
Although we normally think of shifting the head to get perfect alignment, the same effect can be achieved by adding weight to the bolt that goes through the angle plates above the yoke caps. By knowing the weight of the blade assembly, the distance from the the center of rotation to where the weight will be added and the offset of the string, one can accurately calculate the amount of weight required. Note that the offset measurement needs to be quite accurate ... .010" is very significant.
- Current Status
I trailered my craft to a different location to do some more hover testing and hopefully the forward flight track and balance. One of the things we found is that it seemed that there were increased vibrations from what we remembered. After doing some more investigation, it was determined that one blade had leaded ~3/8". This was corrected and more dynamic balancing in a hover was performed.
It was found that although the initial readings had been 0.14 ips a few months earlier, we were now in the 1.15 to 1.2 range! Hmmm ... After a bit more double checking, everything looked the same as when we intitially rigged the blades. Adding / removing weight to a blade seemed to have very little effect. We then performed a .005" head shift and the vibration readings were cut in half. An additional .003" head shift got the readings down to about 0.4 ips which is still considerably higher than what we had a few months earlier.
This still needs more investigation. The first step will be to redo the tracking. If that is okay or has little effect then we'll look at possibly lagging a blade rather than doing more head shifts.
- Miscellaneous
I waffled about whether to add trim tabs to my blades before starting the whole balancing ritual. In the end, I discovered that very few blades have really required these (I've heard numbers like two sets) so I decided not to install them initially. If I have a significant cruise track or balance issue then I'm prepared to go back and install tabs.
Update: After discussion and contemplation, I did add the trim tabs as adding them after the initial balancing causes one to start all over again. I can see that these will be required since my blades are not uniformly straight ... in fact the twist is opposite in the two blades which will surely cause some issues in forward flight.I have not seen proper polar charts that have been developed for the Safari. Most production helicopters (and even the Rotorway) have polar charts such that when one has dynamic data (ips and degrees), they clearly show whether to lead/lag a blade or exactly how much weight to add and to which blade. A real polar chart would go a long way to removing the mystique surrounding the balancing operation.
<NEXT> Required Items
Last updated: September 18, 2009