The kit does not come with any kind of electrical or
mechanical governor that attempts to keep a constant Rotor RPM
(RRPM) and this is strictly a pilot operation. I am aware that
CHR has been looking into the development and sale of a governor
system for many years, but this appears to be a very low priority
for them. From what I understand, they have had the prototype for
a long time and haven't had it installed in a craft for testing.
I strongly believe that a governor
system is a safety item and should be made available to all
builders. This is even more important when one considers that
many builders are new and/or low time pilots and maintaining
manual throttle control will significantly increase their
workload. Perhaps CHR should review the various NTSB reports that
led up to the introduction of the governor system on the Robinson
R22 ... they make for a very sober reading experience.
Update: CHR first reported to the ATSB in 2004 that they were working on a governor. The 2012 Safari 400 advertisement states that it includes a governor but I have not seen one of these units.
Since RRPM is critical to safe operation, I have chosen to purchase, modify and install an electro-mechanical governor from a Robinson R22. At engine speeds above 80%, this governor attempts to control the throttle within a fixed range based on engine RPM; other controllers may be based on rotor RPM, but that's a whole other debate. I did consider designing my own governor system, but that was another project and the R22 governor already has a lot of nice features, such as rate of change, built into it. Although I may leave this switched OFF for all the initial adjustments and testing, I felt it was just as easy to install this during initial construction as there are support brackets for the actuator that need to be fabricated and welded in addition to some extra wiring and changes to the collectives. In the event that I choose not to use it permanently, it is very easy to either just unplug the governor controller or to physically remove it.
The R22 governor is an expensive addition that can sometimes be found at the various aircraft salvage yards if they are parting out an R22. If one is trying to purchase a governor/actuator, they should also try to get the MS3116F12-10SX circular connector that is attached to the rest of the craft's wiring since this connector is not an item that is stocked by a lot of the normal electronics suppliers ... my cost for new ones has been in the $40-$50 range. Note that an internal modification to the controller is required that will make this unit unuseable again in an R22!
The steps required to perform this modification were graciously made public to the Safari community and are available here in MS Word format. It should be noted that this document has been updated and the previous version contained a major error relative to input frequencies and I've noted a slight change when calibrating at the originally specified 4/rev frequencies vs. the updated 2/rev ones. There are still a few math errors and/or nomenclature confusions, but anyone performing these modifications should be double checking this anyway. Also, I have to respectfully disagree with the document's author as I firmly believe the pinouts of the connector on the Robinson documentation are correct as the connector is a well documented variant of the basic one.
I built a frequency generator to re-calibrate the controller per the conversion document and discovered that the adjustments have to be made extremely slowly in order to get consistent results. I was aiming for a setpoint of 495 RRPM and finally chose to do this by averaging both the upper and lower points where I first detected pulses occasionally being sent to the actuator; I've also verified this averaging with the point where the actuator first moves continuously. Note that in the conversion document it actually talks about a "target frequency" and the offset to where the first pulses start rather than an actual frequency. It also talks about a "whining" sound as the "target", but I found this somewhat difficult to isolate as it first goes through progressively faster "ticking" sounds to various "whining" sounds and finally full movement. I actually found it easier to connect an oscilloscope to the actuator signals and watch for the first pulses rather than try listen for them.
Update: After working with a fair number of controllers, I think I've found another anomoly ... a couple of units with the highest serial numbers I've seen appear to have been factory set with a wider deadband but a much more responsive rate-of-change. On these units, I think it's more important to re-calibrate them based on the high end of the deadband rather than the center of it, especially if using the digital tach with a high alarm horn at 502 RRPM. I now have a much more precise frequency generator and one of the things I've noticed is that the deadband appears to be wider than with the 555 frequency generator described in the conversion document. I'm attributing this to the fact that the 555 timer is not that stable and the slight "jitters" are actually triggering the governor controller sooner than a stable frequency generator will.
These adjustments are made with slow changes since the controller actually has circuitry to detect and respond to rate of change inputs. Those that have flown a governed R22 can probably relate to the relative smoothness of the governor unless it is dealing with a big (i.e. quick) change at which time one can actually feel the "twitch". This is also a testament to the excellent throttle correlation on the R22 whereby very little governor intervention is required. Just so I remember: CCW on the adjustment pot raises the target frequency and it took about 1/8 turn to notice any significant change.
I have now had a chance to work with several different controllers and carefully took note of their stock deadband averages using the above technique. Interestingly the averages ranged from 526-532 RRPM and their average midpoint is 529 RRPM which is JUST below the top of the green in an R22 (104% = 530 RRPM). This is basically consistent with data in the reference document, but logically seems a little on the high side. Perhaps this is because its very easy to lower RRPM, but harder to regain it. When I get some feedback from Safari users with known calibration I'll try to further explain this.
Update: I've now had some feedback on these settings and it would appear that the above averaging technique is accurate and the governor seems to hold the setpoint quite well. As a result, I'm now starting to set the setpoint a little higher for ships with the thick wall masts. Instead of targetting 495 RRPM, I'm now aiming for 499 RRPM. On one controller where I carefully tested both settings, it took 2-1/8 turns CCW on the adjustment pot to raise the setpoint by this amount. Note that this may not be viable in craft that use the stock CHR horn module or the newer digital tachs as it will cause the high warning horn to frequently trigger.
Update 2: I have received additional feedback about the operating characteristics of governors that I've adjusted in the above manner. One unit had been set on the bench for a 495 RRPM setpoint and had a deadband that appeared to go from 491 to 499 RRPM. Actual testing has indicated that it appears to hold the range of 492.5 RRPM on the lower end to just ocassionally triggering the horn (502 RRPM) on the high end and the owner is happy with this calibration as-is. I've now had one controller that operated about 2 RRPM higher than the setpoint and another that operated about 1 or 2 RRPM low when operating in a running ship. This is only a 0.4% discrepency and it would appear that there is some variation when the unit is installed with different sensors and possibly different voltages. I still think that I'll be adjusting the setpoint on my personal controller to approximately 497 RRPM. After field testing, this adjustment may need a little bit of fine tuning. Also, since my tach has an adjustable high alarm setpoint, it will be set a little higher than the factory setting in order to act more as a true alarm versus a warning.
Update 3: As more users are actually testing these in flying machines, I'm getting more feedback on the results of the bench-adjusted setpoints. The small deviations noted above are still occurring, but this could also be due to different calibrations on various horn modules and a host of other factors unique to the craft. So far, the widest deviation appears to be about 2 RRPM which I think is okay unless it is causing the high alarm horn to chirp too often.
The actual controller is approximately 5-1/4" wide by 7" long x 2" high. However, one needs to allow for the circular plug which potentially increases the clearance requirement in length to about 9-1/2". I've chosen to mount the controller on a plate under the pilot's seat in a manner similar to the solenoid board under the passenger seat. In doing so, I added six welded tabs to the frame.
N.B. This picture is during construction for sizing, not final positioning.
If the governor is being installed as a retrofit to a painted craft (i.e. one doesn't want to weld on tabs), it's possible to mount a plate to the frame under the pilot's seat using adel clamps. The .063" aluminum plate in the following picture is held in place with three adel clamps; two along the top edge and one on the lower right side just outside of this view.
One of the things I don't like about the internal connections is that the controller case is electrically connected to the power ground pin on the connector rather than a separate frame ground pin or mounting bolt. While it is extremely important that the case is grounded for EMI/RFI shielding, this introduces the possibility of creating a ground loop if the controller is mounted on a metal surface. I mounted my controller using nylon shoulder washers in order to electrically isolate the case from my aluminum mounting plate.
Accessories and Wiring
In addition to mounting the controller and actuator there are a few other items that must be installed and wired in order to complete the governor installation. Even if a builder is not planning on initially installing a governor system, I would highly recommend that they allow space in their panels for these items and preferably even pre-drill the holes. For the small cost of these parts (perhaps $25+), they may want to even pre-install them and just mark them as INOP. Alternatively, for under $1, one can get plastic "buttons" in the hardware store to fill these holes until components may be installed.
A 2 Amp breaker needs to be added to the power distribution bus. If one is using a relatively stock layout, I would assume that this is placed on the switch console alongside the other breakers, switches and/or fuses. Since there is an on/off switch that should be used during startup and shutdown, it is reasonable to consider just using a separate fuse/breaker rather than a breaker/switch as supplied with the Safari instrument kit.
A small reliable SPST on/off switch needs to be added
somewhere that's conveniently accessable in flight. I got used to
the position in an R22 so I'll be mounting it on the end of the
collective. If the builder already has the stock starter switch
mounted there, then they'll either have to modify this area or
find another location such as on the cyclic grip. In my case,
I've added a small "box" onto the end of the pilot's
collective which houses both a starter switch and the governor
on/off switch. I've also left a little room on here for future
expansion in case I decide to add other switches.
Update: By using the ultra mini toggle switches, it is possible to mount one of these in the Flagman grip between and slightly above the two pushbuttons. Besides drilling the hole, this will also require the removal of a bit of the internal webbing that is cast into the plastic.
A Governor-Off indicator lamp should be added to the main instrument panel near the other warning indicators. While the dimmable MS25041-xx press-to-test lamps are nice and also consistent with the indicators supplied in the Safari kit, a simple 12-volt indicator lamp is all that's really required since the governor switch should be OFF before startup and the lamp illuminated. This lamp should use a relay to connect it in a manner similar to the RHC wiring diagram. I have heard from a couple of builders who are confused about the relay and why it is used in the lamp circuit. I believe this was installed to act as a signal inverter, a fail-safe indicator and also to conform to other lamps whereby the ground is switched rather than switching the positive voltage. Many of the mini 12 volt SPDT relays should work with a GE-330 lamp as it only requires about 80 ma. of current. One builder had suggested using a SPDT switch for the "Governor ON" function and using the second throw (i.e. OFF) to power the lamp; that will work, *BUT* it will not properly illuminate the lamp if the breaker is in the OFF position. I've chosen to build a small solid-state relay substitue that actually sits on a small bracket attached to an MS25041-xx lamp. This device should be more reliable than a mechanical relay, but more importantly it also has an optional input for the LASAR® enunciator signal. Thus if either the governor is switched OFF (switch or breaker) or the LASAR® enunciator is ON (i.e. LASAR® is OFF which implies the tach signal is not available) then the Governor-Off lamp is lit. Just to clarify, I'm installing both a LASAR® enunciator and a Governor-Off lamp.
Updated: There are lots of ways to use electro-mechanical relays; these are some UNTESTED ideas.
From various other snippets of information I've received, it would also appear that it is very important to be extremely careful when doing all of the ship's wiring and try to harness any sources of EMI/RFI in order to avoid any interaction with the controller. I tend to go overboard on the shielding, ground loop avoidance and separation of wire bundles, but that's because I've had to try track down this kind of problem in the past and I know how difficult it is to find and correct the source.
Using the default wiring from the Robinson documention, ClockWise rotation on the actuator (looking at the actuator arm and nut) appears to attempt to decrease RPM while CCW attempts to raise it. It is EXTREMELY CRITICAL to verify an installation before attempting to actually engage the governor in flight. One suggestion MIGHT BE to manually raise the RPM to approximately 90% (~2500 RPM) on the ground with the collective FULLY DOWN and then BRIEFLY engage the governor while watching the tach. If the tach goes up then the governor can be further tested with progressively longer engagements. However, if the tach goes down then the actuator is wired backwards; this simply requires the two wires to the actuator to be swapped since it is a reversible device. The stock actuator comes with a polarized connector that mates to an RHC part # B263-33 housing with two B263-2 sockets. These are actually a relatively standard electronics connector and I have some of these housings / sockets available. It is possible to just crimp on the sockets and then connect them to the actuator pins until the correct orientation is determined; the housing can then be installed over the sockets.
Controller "Tach" Input
The main input to the controller is an engine RPM (i.e. tach type) signal with two pulses per revolution. In the R22, the S4LSC-204T mag on their engine has an extra set of breaker points that are used to supply a 2/rev pulse train to both the governor and the engine side of the rotor/engine dual tach instrument. It appears that the governor controller can be directly connected to this mag-supplied signal without an R22 tach or any other form of external pull-up or signal conditioning. I have tested this condition using a FET driver without a pull-up resistor and it appears to work fine. After some more testing and investigation, it would appear that both the tach and the governor controller have pull-up resistors and either one can be used without requiring the other one.
At first, it appeared that I would probably have to perform
additional interfacing considerations since the LASAR® ignition
/ magnetos do not have the same breaker point outputs as those
used on R22s. Although I could easily add electronics inside the
governor controller box, my objective is to do this outside the
box such that a single calibrated controller could be used with
any type of ignition system. It would appear that the LASAR®
tach drive signal (brown wire) has a compatible pulse train (2
pulses per engine revolution), but the signal may need to be
level shifted as the LASAR® specification calls for the low
level signal as a maximum of 1.5 volts which I believe is too
high for the governor input. While working with another builder
who has installed a governor with the LASAR® system, he has
found that wiring the LASAR® tach drive signal directly to the
governor controller has worked fine for him. The main negative of
this scheme is that this tach drive signal is not available when
the LASAR® system is turned off and
one should ensure that the "Governor Off" lamp is lit
when the LASAR® enunciator is on indicating backup mode. For
now, this is the basic scheme that I'll be using.
Update: The later versions of the LASAR® installation manual clarify that they use a four pole magnet rather than the more traditional two pole magnet used in most magnetos. This implies that a vent hole sensor should provide the 2 pulse per engine revolution signal that the governor controller requires. Once my craft is operational, I'll be testing this combination. The obvious advantage is that the governor system will then be independant of and not rely upon the LASAR® system being turned on and functioning correctly.
Update 2: This combination of a LASAR® system and a vent hole sensor has now been field tested and works extremely well. I believe this is the best way to connect the governor system to a craft that has LASAR® mags ... I've ordered the sensor so that I can upgrade my system.
Update 3: I have now tested the governor in my LASAR-equipped craft with both the brown wire tach signal and a vent hole sensor. My conclusion is simple: USE THE VENT HOLE SENSOR!!! The brown wire tach signal does work but it appears that there is some noise on the signal ... I didn't scope it but it showed up as a hunting governor and one can see the MP gauge fluctuating quite a lot as the governor moves the throttle, even on the ground with the collective full down. With the vent hole sensor installed, this fluctuating settled right down and there were no obvious issues. The solution is simple!
Of course there are possible cases where neither the R22 mag input or LASAR® tach signal are available (i.e. dual Lightspeed ignition, E-Mags, etc.). I have looked into interfacing to the Lightspeed units, and from the documentation and bench testing it appears that wiring pin 6 of the Plasma input connector directly to the governor controller's input should work fine, but with the same limitation as described above for the LASAR® system. This still needs to be confirmed on an actual installation and I am in contact with one builder who has this configuration. I also need to check whether there is a spare Hall-effect output on the sensor that could be used in the case where the builder installs one standard mag and only one Lightspeed unit. The E-Mag documentation states that it has a digital "tack" signal and the manufacturer has confirmed that it's output should be suitable for use with the governor ... I have no direct experience with this option.
Update: The Lightspeed unit cannot
be directly wired to the controller as it appears that the pulse
width is too narrow and requires either a pulse stretcher or an
internal change to the controller. I've
developed a pulse stretcher that seems to cure the problem, both
on the bench and in flight testing. This appears to work very
well but I have chosen not to fabricate a PCB (printed circuit
board) at this time. These prototype circuits are very time
consuming for me to fabricate so if anyone is expecting me to
build one for them then they should check with me first ... I
really don't enjoy building them and unless there's more of a
demand I won't be creating a PCB which would greatly speed their
Update 2: The pulse stretcher that I originally designed works fine on Plasma II and Plasma II+ systems. However, I have recently had a negative experience with a certain revision of the Plasma III where the manufacturer has confirmed that the pulse shape does not match the documentation. While I believe I have a simple solution for this, it has not been field tested. Although the pulse stretcher is the cheaper method to interface, I now recommend using a vent hole sensor and a frequency doubler when a combination of one Lightspeed / one magneto system is installed. This keeps the governor totally independant of the Lightspeed unit.
E-Mag update: I've now worked with an owner who has E-Mags installed. The E-Mag was configured as two pulses per revolution and open-collector output but failed to operate successfully. After a lot of testing, it appears that the E-Mag "tack" signal pulse width is too narrow since the addition of a pulse stretcher made the combination work fine. Unless the E-Mag firmware is upgraded to lengthen the pulses, it will be necessary to add a pulse stretcher between the E-mag "tack" signal and the governor controller.
As an alternative for the situation where there is only standard mags and no electronic ignition, I've thought about a couple of alternatives but haven't actually tested them. The ideal solution would be to find a magneto vent hole sensor that will deliver two pulses per engine revolution. I have looked at a few of these, but their skimpy documentation seems to indicate that they only provide one pulse per engine revolution and thus would require a frequency doubler. If anyone has the details on a vent hole sensor that would provide a 2/rev signal, then I'd really appreciate receiving that information.
Update: I have delivered a governor system using a vent hole
sensor to a Safari owner for testing. This system included a
prototype frequency doubler and I'm awaiting feedback on this
Update 2: The oringinal frequency doubler had an issue that the owner failed to detect and/or give me feedback about. As a result, I have created a revised version of this module. I now have experience with a couple of systems that have been installed with a vent hole sensor and the frequency doubler. It would appear that this combination is working well and I have had no reports of problems with the updated modules.
I am aware of at least one builder who added a second set of points into his standard mag and also another builder who is having an AME (A&P) do this for him. This is probably the best solution since it uses a proven technique using standard components, but it does require the mag to be removed and then re-timed.
Another simple option might be to use a dual tachometer drive (similar to Spruce's part # D-778-T) and use one side for the actual tach drive and the other side to drive a Westach 303DH2T Hall effect sensor. Although the literature states that this is a 5 volt device, I've checked with Westach and this sensor can be run directly off 12 volts. However, since the tach drive turns at 1/2 engine RPM and the 303DH2T sensor gives 2 pulses per rotation, this would require a frequency doubler or a geared adapter. The UMA Instruments part # 1A3-4 sensor should give the required 2 pulses per rev since it actually generates 4 pulses per rev of the sensor. I've also thought about using a gear tooth sensor detecting the bolt heads of the tail rotor driveshaft coupler as a tach pickup/input. This is a simple mechanical/electrical hookup and by switching two of the bolts to non-ferrous stainless, one should get the equivalent of two pulses per engine RPM; *BUT* this is actually measuring Rotor RPM after the clutches, not engine RPM. This definitely has some downsides. A better alternative may be to add a pair of targets (magnetic or ferrous depending upon sensor) in the starter ring gear support and use a gear tooth or hall effect sensor to detect them and provide a tach signal. It would probably be a wise move to choose magnetic targets and install them per the Lightspeed requirements and thus have the option of a future upgrade.
I've been working with one builder who was awaiting LASAR® mags and only had standard mags available. He chose to add two targets onto the clutch shield plates and then have a sensor mounted to detect their passage. Three inch long targets appeared to give too short a signal for the sensor and controller combination that he is using, but five inch long targets seem to work quite reliably. By my calculations, four inch targets would probably also work, depending upon the sensor. If using this technique, the builder should try to match the weight of the targets in order to prevent new vibrations.
Actuator & Linkage
There seems to be some confusion about the mounting location and linkage for the actuator. The original released photos of the actuator on a Safari showed the actuator mounted on one of the lower frame tubes and the linkage going directly to the horn at the front of the throttle rod that goes through the firewall. Another placement I have now seen is on the collective cross-tube such that the linkage goes to the right throttle horn ahead of the correlator. The more I've thought about this, the more I believe the actuator should be ahead of the correlator as this will prevent excessive movement in the actuator's slip clutch, prevent increased drag on the collective and prevent the correlator and actuator from opposing each other. The actuator essentially imitates a manual input when placed ahead of the correlator and this is the location where it is installed on the R22. Another pictures of an installation like this is available here.
The mounting technique pictured above appears to work well when retro-fitting an actuator to a completed ship and this has been done by several builders. The mounting tabs are welded to a piece of half-round split tubing which is then bolted to the collective pivot tube. Since I haven't done this, I have no opinion as to whether there is any significant weakening of the pivot tube due to the holes for the two bolts. The linkage goes from the actuator arm to a tab that is welded to the bottom of the copilot's collective just in front of the stock arm. One thing to be very careful about if using the above-referenced pictures for the actuator mounting location is that the linkage will probably interfere with the stock routing of the seat belts.
I looked at all the various permutations of where to mount the
actuator and decided to try something very different; namely on
top of the collective cross-tube and connected to the throttle
arm on the pilot's collective. The advantage is that there is
absolutely no interference with seat belts; the disadvantage is
that one has to carefully position the actuator so it clears the
throttle inter-connect linkage (I used about 1/2" clearance
at the top at 1/2 throttle). Also, any slop in the two throttle
inter-connect rod-ends will be seen by both the actuator and in
the pilot's throttle grip. I welded a plug into the largest
lightening hole at the bottom of the pilot's throttle arm and
drilled a couple of 3/16" holes in it which is where my
linkage will bolt into. This can also be accomplished with a
simple flanged plug and washer that are held in place by the bolt
that goes through the rod-end. I had thought about whether it
might be viable to run the actuator linkage directly from the
actuator arm to a tab welded or clamped onto the inter-connect
rod. Unfortunately, due to the relatively long arms on the
throttles, this would highly desensitize the actuator and there
would probably not be enough throw unless one added an extension
to the actuator's arm.
Note that there is a significant angle to the rod between the actuator and throttle arm when using this arrangement. The full range of motion needs to be checked and it may be necessary to use cap screw washers as spacers similar to what is done when rigging the tail rotor control linkage.
Update: Another builder has also chosen this same mounting location on a retrofit installation but used the half-round split tubing base since his control system had already been completed and installed. This craft is now flying with the governor engaged and the owner is totally happy with the mechanics of the installation since the system is working extremely well and caused no interference issues with seat belts or other items.
More actuators mounted with the link to the pilot's collective
It will be necessary to fabricate some mounting tabs for the actuator. I used .063" 4130 and I've tried to attach a scan of my basic pattern ... this should be double checked due to re-sizing errors of pictures. In the R22, the actuator is mounted to the tabs using a NAS1351-4-28P screw with a MS21042L4 nut plus two AN503-8-4 screws. The linkage to the actuator's arm is in the sequence: NAS6603-7 bolt, arm, AN960-10 washer, Heim HM-3M rod end, AN960-10L washer, MS21042L3 nut, pal nut. After this assembly, all the nuts are on the same side of the arm.
Actual width is ~ 1-7/8" and the shaded area was removed
As received and noted under Collectives in the Controls page, the stock collective tubes have a fixed (well temperature variable) amount of friction built into them as a result of drag and grease between the two rotating tubes. In my opinion, this is too much resistance when a governor is installed and must be altered such that there is MINIMAL friction without the governor actuator linkage connected. The other area of concern with throttle friction is the throttle torque tube that runs from the cab to just below the carb which is just a greased sleeve bearing on earlier ships; see the notes on the throttle linkage in the Upgrades section of the Frame page where the factory appears to have changed these bearings to rod ends which should work better with the governor. Once the governor actuator linkage is installed, the actuator has a built-in slip clutch that will provide the required friction to maintain the throttle position. In the event one is pre-planning for a future governor installation or later removes the actuator, some other form of friction device will definitely need to be installed. One idea I was given but have not followed up on is to investigate the various after-market motorcycle throttle friction devices.
It should be noted that the tension for the slip clutch in the actuator can be changed by replacing a pair of journals and I'm aware of at least five different thicknesses of journals that can be used to obtain the specified 12-13 inch-pounds of rotational friction on the actuator arm. Unfortunately, I am not at liberty to reproduce here the documents that this information was obtained from. Also, I would be interested in getting feedback from any Safari builder who has measured this friction and can tell me both the arm length on their throttle linkage and whether the specified friction is appropriate for the Safari.
After talking to several owners who have installed governors,
it seems there are various opinions about the amount of force
required to manually over-ride the slip clutch. A couple of them
have chosen to back off the retaining nut on the clutch thus
reducing the friction but creating a delicate balancing act
between loose enough for manual over-ride and strong enough to
overcome the various frictions within the throttle linkage.
Another builder using inserts in the center hole of the
collective's throttle arm, feels that while it is quite stiff it
is acceptable and provides solid governor operation. I have
several holes available in my collective's throttle arm and I
will be experimenting with various moments.
Update: I am confortable with my setup using the stock tension adjustment on the actuator and a linkage to center of the largest lightening hole in the pilot's throttle. Yes there is a lot of friction but not so much that it can't be over-ridden without the use of excessive force. More importantly, I have not detected any indication that the slip clutch is slipping on the very fine initial movements. I would highly recommend that one try using the stock tension before trying to adjust it. In the event that one does choose to adjust the slip clutch retaining nut, there are a couple of cautions:
1) This adjustment is extremely sensitive. In my testing, if I can detect any movement of the nut then I can measure a change in the slip clutch friction.
2) This nut has a lot of torque applied to it. Since there is also a gear train attached to this shaft, any adjustment to the nut means that the hex end of the shaft should also be held with a wrench or vice in order to prevent the wrench's torque from being transmitted to the gear train.
I have been asked how to electrically test whether the actuator works without using a special test box or a completed ship with a running engine. It would appear that the actuator is essentially just a reversible geared electric motor and both the speed and direction issues are dealt with in the controller. Thus one should be able to just connect the actuator to a 12V battery to test it; I have done it with mine to verify that this works and can see no ill effects.
I also remember someone commenting on the old forum about the fact that both throttle arms and their associated linkages are on one side of the torque tube and thus tend to try move the linkage via gravity. A counter-weight could eliminate this, but I think I'd review the whole linkage design before resorting to this solution. I did briefly look at this when I converted my throttle torque tube from sleeve bearings to rod ends, but the simplest routing changes appeared to have interference issues. A spring could also be used to neutralize this force, but then one has to consider what effect it would have if it breaks, etc. The use of 3/16" rod ends vs. the CHR-supplied 1/4" rod ends would reduce their weight, but I haven't tested this or investigated any strength related issues.
After installing a governor, the pilot will need to be consciously aware of not over-gripping the throttle. For those with experience in an R22 or other ship with a governor, this will be natural. However, if one has only flown something like a Bell 47 or Safari without a governor then it is possible that they can grip the throttle too tightly and override the governor due to the slip clutch in the actuator.
It should be noted that Robinson issued a Safety Notice that the governor can mask carb ice. In an icing condition, the governor may slowly be adding throttle to compensate for the loss of power and the pilot may not detect this. The Safari has a carb temperature gauge and one should include it in their visual scan. Carb heat should be applied as appropriate to keep the temperature out of the icing range.
It is reasonable that there may be some movement of the throttle by the governor as one transitions from ground idle to a hover and back. However, if there are "twitches" from the governor during other phases, such as hover through takeoff, it can be indicative that the correlator has not been properly adjusted. Ideally, if the correlator is properly adjusted there should be no actions by the governor and it should merely be a watchdog function. This also greatly simplifies governor-off operation, whether for training purposes or due to an actual failure.
I believe that the governor should NOT be turned on during startup or shutdown and I would suggest a modified version of the R22 procedures. During startup, one should wait until after all the checks have been performed, including the sprag clutch needle split, and one is ready to throttle up for flight. The throttle is then manually raised to just below normal RPM, perhaps 95%, and the governor is then turned on. One can then feel the governor advance the throttle the last little bit and also see this on the tach to verify its operation. If the governor is turned on before throttling up, it will attempt to take over once the engine RPM reaches approximately 80% and it is possible that there might be a sudden unexpected engine surge that could possibly lead to liftoff RRPM or a rotor overspeed. I do not know of a factual case where this happened, but I use the above sequence and watch the tach carefully as the governor is engaged.
During shutdown, I would suggest turning the governor off just BEFORE reducing the normal operating RPM for cool down and then manually reducing the throttle to it's cool down setting. This way if the selected cool down RPM is still within the governor's operating range, or advances to within this range, it will not attempt to suddenly and unexpectedly raise the throttle setting.
I've now modified several of these governors and provided them to other owners. So far the owners have been 100% unanimous in their praise for the system and how much easier and more enjoyable it makes the craft to fly. Instead of things like keeping one eye glued to the tach or keeping the revs below top-of-the-green in order to avoid an overspeed, the governors have been keeping the RRPM right at top-of-the-green and peak power without any need for manual over-ride. There has been some rumours that this system has too much lag in it and allows for delays and/or significant changes in RRPM. All of my first-hand reports of installed systems indicate absolutely no such behaviour and the owners are totally convinced that it does a much better job of RRPM management than they could perform manually. They are also unanimous in their opinion that it was money well spent to improve their craft and would not consider removing it.
I've also heard a concern that the use of the governor could mask a condition whereby maximum power has been obtained by the governor without the pilot's knowledge and there is no reserve power available for landing. While this is possible, I believe that this condition would be the direct result of poor piloting technique, which includes a failure to detect carb ice as noted above. During the startup sequence, one should check the OAT and altimeter which should give a good indication of the prevailing density altitude. During the initial liftoff and even in cruise flight, the MP gauge will give a good indication of how much power is required under the existing conditions. If there is any doubt about reserve power during a landing, the pilot should perform a dummy run with a power check to verify the power requirements. If there is not sufficient reserve power available, then it makes no difference whether there is a governor installed or not ... full throttle is full throttle and the options are exactly the same.
Although I have obtained many of these systems in the past and made them available to other builders for their use, this is something that I no longer do since it takes a lot of time to locate quality units at an affordable price. Anyone looking for a system can start with the various salvage yards, AME's (or A&P's) and in general keep a lookout for various roll-overs. Note that any used parts must be carefully evaluated as to whether they are re-useable and suitable for conversion. Any RHC service center should be able to order overhauled parts, but these will be considerably more expensive and there is the question of whether they'll do it for a non-RHC machine. As for any of the systems that I've obtained in the past for others ... I'll continue to try my best to provide any support that is required.
Just to clarify and make this point perfectly clear: Unless a great deal on one of these systems were to drop into my lap, I have no intentions or desires to try procure these systems for sale to others. IF a Safari owner does manage to obtain one, I will work with them to evaluate it and modify it. I do have the equipment and experience to perform the conversion but I refuse to spend a lot of time trying to track these systems down ... it's simply not worth my time. It's also not really worth my time to perform these conversions when I calculate the hourly rate I receive, BUT I feel very strongly that this is a safety item. Until there is a reasonably priced governor option from CHR I will continue to assist builders with these.
As to where a person might start looking to obtain one of these systems ... besides Robinson dealers and A&P's there are the various aircraft salvage yards. In the past, I've used Dodson and Wentworth plus a few others that I'm not as keen to recommend. There are usually advertisements in Trade-a-Plane and Barnstormers for people that are parting out machines as well. The ideal source is to find someone who has bought an entire R22 for salvage because of the engine or avionics and is willing to sell some of the other parts at a reasonable price. So what is an appropriate price? That's a question with results that are all over the map. At the time of this writing, new parts from RHC are listed at $3,050 and overhauled parts from RHC (without core) are listed at $1,920. Unless a seller is willing to guarantee the parts (and you have faith in their guarantee), I don't think they're worth a lot more than the core value of $820 for the pair. I would think that anything under $1,000 for a useable system is a good deal.
<NEXT> Preparation and Paint
<BACK> Controls and
Last updated: June 19, 2013