Auto Pilot Preflights

Captain

Final Approach
Joined
Mar 12, 2012
Messages
8,002
Location
NOYB
Display Name

Display name:
First Officer
There's been a few threads on auto pilots lately and that's brought this topic to my mind.

I'm a big fan of preflight checks and the autopilot preflight checks are no exception. However, there is one check I hate. Every plane I've flown with an autopilot installed has the pilot ensure they can override the clutch by engaging the AP and pulling against it to the point of slipping the clutch.

To my mind this seems like a bad idea. It's wear on the parts and as long as the various disconnects work normally I don't see the point. Every AP I've flown has no fewer than 3 seperate ways to disengage the thing and each are checked in the first flight preflight. So what's the point of grinding the clutch?

I don't like damaging equipment and in this case there seems very little benefit in performing the override test. What do you all think? Am I off base?
 
I think that if the override test is written into the AFM Supplement as a required test, we don't have the option to skip it no matter what downside we think that test might have. Also, my experience with things mechanical is that if they aren't exercised, they tend to lock up, and an autopilot clutch that is locked up is not something with which I'd want to fly.
 
Grinding clutches may not be required. For the KAP-140 in the Cessna I fly the wording in the checklist is:

"FLIGHT CONTROLS - MOVE. Make sure that the autopilot can be overpowered in both pitch axis and roll axis."

Thus, I take "overpowered" to mean that if it's trying to hold wings level then I can turn the yoke, see the ailerons move, then I let go. That's as far as I take it, no need to "fight" Otto for the controls as I've verified that in an emergency I could overpower the autopilot and fly the plane, albeit with quite a workout. I see nothing that states I need to slip a clutch.
 
Last edited:
My current autopilot has a test button, so that's the only thing I check.

Back when I flew with a GFC700 I tested it before each flight (well, first flight of the day. I engaged the autopilot, tried to gently move the stick to make sure it doesn't move (without overpowering it), then pressed the CWS button and tried to move the stick that way, then disengaged the autopilot.

I don't think I have ever tried to overpower an autopilot. Just seems like I'm damaging it.
 
My current autopilot has a test button, so that's the only thing I check.

Back when I flew with a GFC700 I tested it before each flight (well, first flight of the day. I engaged the autopilot, tried to gently move the stick to make sure it doesn't move (without overpowering it), then pressed the CWS button and tried to move the stick that way, then disengaged the autopilot.

I don't think I have ever tried to overpower an autopilot. Just seems like I'm damaging it.
If that test is called for in the book, it's designed to take it without damage, and it's both risky and illegal not to do it.
 
Oh for God's sake, the manufacturers know better than you and they ask you to test it. Just shut your pie hole, stop thinking (because you're not good at that) and follow the damn instructions okay?
 
Oh for God's sake, the manufacturers know better than you and they ask you to test it.
Actually, if it's in the AFM supplement as a mandatory test, they're not asking, they're telling, and the FAA says in 91.9(a) that you must do what they said.
 
Oh for God's sake, the manufacturers know better than you and they ask you to test it. Just shut your pie hole, stop thinking (because you're not good at that) and follow the damn instructions okay?

So a guy who has zero aviation experiences comes here and tells us how to think...
 
Last edited:
Grinding clutches may not be required. For the KAP-140 in the Cessna I fly the wording in the checklist is:

"FLIGHT CONTROLS - MOVE. Make sure that the autopilot can be overpowered in both pitch axis and roll axis."

Thus, I take "overpowered" to mean that if it's trying to hold wings level then I can turn the yoke, see the ailerons move, then I let go. That's as far as I take it, no need to "fight" Otto for the controls as I've verified that in an emergency I could overpower the autopilot and fly the plane, albeit with quite a workout. I see nothing that states I need to slip a clutch.

I just looked at the a KAP140 supplement for a DA40 and it seems to contradict itself.

In the limitations sections it says "Overriding the autopilot to change pitch and roll attitude is prohibited. (Disengage or press CWS while maneuvering.)"
Then in the before taxi checklist it says "FLIGHT CONTROLS - MOVE fore, aft, left and right to verify that the autopilot clutches can be overpowered."

So if it's prohibited to overpower them in flight...why would it be safe to do that on the ground?


Here is another thought, say it's a rental aircraft and it does 4 or 5 flights per day (with different crew). Each crew will do a run-up with the autopilot test, this means the clutch will be overpowered 4-5 times per day, every day. Something tells me that clutch will not last very long.
 
Last edited:
If that test is called for in the book, it's designed to take it without damage, and it's both risky and illegal not to do it.

Perhaps....but how often do you use the exact checklist from the POH when your flying? Chances are you use a checklist that's similar to the one in the POH, but not exactly the same. So if that's illegal, it is clearly not enforced at all.
 
I agree with Captain - it seems, somehow, wrong to overpower the clutches.

I agree with Ron - the AFMS requires you do it.

I am going to write an avionics guy I know, see what he has to say about the whole question.
 
I just looked at the a KAP140 supplement for a DA40 and it seems to contradict itself.

In the limitations sections it says "Overriding the autopilot to change pitch and roll attitude is prohibited. (Disengage or press CWS while maneuvering.)"
Then in the before taxi checklist it says "FLIGHT CONTROLS - MOVE fore, aft, left and right to verify that the autopilot clutches can be overpowered."

So if it's prohibited to overpower them in flight...why would it be safe to do that on the ground?

It's entirely reasonable that an AFM might tell you to check that you can still control the aircraft in an emergency, while prohibiting that emergency override during normal operation of the aircraft.

It's also entirely reasonable that the wear conditions on the autopilot mechanisms are worse when they're present for an extended period than if they are only present briefly. For example, imagine a system where ~5 sec of override creates negligible stress/wear on the system, but >15 sec of continuous override starts to cause degradation.

The manufacturer is not required to provide all the rationale behind their instructions, but they really do generally think about such things and often such instructions are based on quite solid data. Ignore them at your risk.
 
Oh for God's sake, the manufacturers know better than you and they ask you to test it. Just shut your pie hole, stop thinking (because you're not good at that) and follow the damn instructions okay?

Interesting position ya got there.
 
I let the auto pilot run it's own check,then cycle it through on the ground.
 
I just do the test it asks for in the flight manual. I haven't seen any issues or heard of anyone able to actually wear them out. IIRC the clutches are electro magnetic with a small friction plate when power is applied to the system. My guess is that it would take more cycle to wear that friction plate out then you could ever put on it by just doing the pre-flight checks. Did find an interesting article about these clutches not disengaging even with power off. IF you cant over ride them it can be a problem.

http://www.ntsb.gov/doclib/recletters/1986/A86_132_134.pdf
 
I just do the test it asks for in the flight manual. I haven't seen any issues or heard of anyone able to actually wear them out. IIRC the clutches are electro magnetic with a small friction plate when power is applied to the system. My guess is that it would take more cycle to wear that friction plate out then you could ever put on it by just doing the pre-flight checks. Did find an interesting article about these clutches not disengaging even with power off. IF you cant over ride them it can be a problem.

http://www.ntsb.gov/doclib/recletters/1986/A86_132_134.pdf

Thanks for the link
 
If that test is called for in the book, it's designed to take it without damage,

Agreed

and it's both risky and illegal not to do it.

Ehh... most list it as a limitation for use. So if you don't use the autopilot, no preflight check is required.
 
Last edited:
Ever drive a vehicle with a manual transmission? You slip that clutch a lot more each time you start from stop than a 1//4 turn of the yoke to make sure it will slip. And it takes tens of thousands of starts to wear the clutch in your car. So, if your auto pilot clutch is designed to be tested by slipping every few hours, I wouldn't feel the least bit bad about doing it. (Unless, of course, the autopilot is a jerk.)

Do you feel bad about using the brakes?
 
Do you feel bad about using the brakes?

I really do hate it when someone pulls out in front of me, cuts me off or otherwise forces me to stomp on the brakes. Hate putting all that wear on the brakes and tires. Seriously, that's something I think about.
 
Ever drive a vehicle with a manual transmission? You slip that clutch a lot more each time you start from stop than a 1//4 turn of the yoke to make sure it will slip. And it takes tens of thousands of starts to wear the clutch in your car. So, if your auto pilot clutch is designed to be tested by slipping every few hours, I wouldn't feel the least bit bad about doing it. (Unless, of course, the autopilot is a jerk.)

Do you feel bad about using the brakes?

As I understand it the wear surface of an autopilot clutch is not much like a clutch on a car. I believe my Century has two sandpaper covered disks that are pressed together to make the clutch. I really don't want to be slipping that assembly very often...and I do know at least four ways to "kill" the autopilot...
 
I never use the darn thing so it's good to give it some work with a ground check.

The Century I ground check:

1. Turn on Century I Master Switch.
2. Pull Turn Command Control “out” and rotate left and right. Note wheel movement
in proper direction.
3. If Tracker is installed, with Omni Converter operating, rotate the Turn Command
Control to the detent position and push “in” to turn the Tracker “on”. Rotate the
Omni Bearing Selector to move the deviation needle left and right. Notice that
the control wheel moves in the correct direction,
4. While taxiing, with the Turn Command Control in the detent, note that wheel
movement is opposite the direction of turn.
5. Check the aileron over-ride in each direction.​
6. Depress wheel interrupt switch and note servo disengagement.
 
I really do hate it when someone pulls out in front of me, cuts me off or otherwise forces me to stomp on the brakes. Hate putting all that wear on the brakes and tires. Seriously, that's something I think about.
:rofl:
In the car I try to avoid using the brakes because I am converting hard earned kinetic energy into waste heat. But in the airplane it doesn't bother me a bit to stand on them when I am doing stop and go's. :dunno:
 
Ehh... most list it as a limitation for use. So if you don't use the autopilot, no preflight check is required.
Point taken. Then let me say instead, "It's both risky and illegal not to do it if you intend to use the autopilot."
 
Ever drive a vehicle with a manual transmission? You slip that clutch a lot more each time you start from stop than a 1//4 turn of the yoke to make sure it will slip. And it takes tens of thousands of starts to wear the clutch in your car. So, if your auto pilot clutch is designed to be tested by slipping every few hours, I wouldn't feel the least bit bad about doing it. (Unless, of course, the autopilot is a jerk.)

Do you feel bad about using the brakes?

As a matter of fact I do. Well, I don't feel BAD...but I do consciously try to limit their use. My plane has $15K brake assemblies and also has reverse. So I have other means to slow down and really expensive brakes.
 
Ever drive a vehicle with a manual transmission? You slip that clutch a lot more each time you start from stop than a 1//4 turn of the yoke to make sure it will slip. And it takes tens of thousands of starts to wear the clutch in your car. So, if your auto pilot clutch is designed to be tested by slipping every few hours, I wouldn't feel the least bit bad about doing it. (Unless, of course, the autopilot is a jerk.)

Do you feel bad about using the brakes?

The clutch in the car is used a lot more often than the one in the autopilot. So the clutch in the autopilot uses the same concept but I assume it's smaller and won't last nearly as long if used as often as your car clutch.

Regarding the brakes, they are much cheaper to replace than a clutch.
 
Did find an interesting article about these clutches not disengaging even with power off. IF you cant over ride them it can be a problem.

http://www.ntsb.gov/doclib/recletters/1986/A86_132_134.pdf
Ah yes, the Bendix MD-4 autopilot and its 'magnetic particle' clutch. A widow-maker if ever there was one. That report has quite a list of accidents under similar circumstances. No wonder the FAA actually acted on the NTSB's recommendations--the evidence was overwhelming. So what then happens when the fleet has complied with all ADs and SBs, and yet another similar accident occurs? They blame the pilot, of course. Like this one who crashed a mile or two from my house:
EDIT: I guess it didn't have the AD--didn't see that on my first read. The crash occurred just two days after the expiration date of the AD.

For several years after, I daily drove or jogged or both past the pile of wreckage that was moved to the airport near my hangar. Not much there that resembled an airplane, just torn and jagged brightly painted metal.

The same autopilot was installed in earlier Learjets, of which I logged a fair amount of time. Never had any issues with them, myself, but friends of mine did. Ironically, the one case of a Learjet autopilot that wouldn't disengage for me happened in a later model equipped with the 'improved' DC torquer style of servo. Heavy rainwater froze at FL 390 and when the captain clicked off the autopilot near the middle marker--it stayed engaged. He landed using nothing but electric trim. As we cleared the runway I turned off the stall warning system and the control stick released and fell forward to the stop by the weight of the elevator. Hmmmm... a lesson duly noted. :idea:

dtuuri
 
Last edited:
Mine has a test switch, I do that till it's done.

Mission complete.

Sounds like an S-TEC, right? Those are the ones I've seen with the Test switch.

But there is a helluva test sequence in the manual, and it's something I do not do: It is seventy-seven steps, to be done in the plane with the engine running per the manual!

IMO, as long as it'll disconnect, that's enough. It's not like I turn into a passenger when I engage the autopilot - I'm still paying attention to what's going on, and if there's any sign of trouble, I'll hit the red button.
 
Sounds like an S-TEC, right? Those are the ones I've seen with the Test switch.

But there is a helluva test sequence in the manual, and it's something I do not do: It is seventy-seven steps, to be done in the plane with the engine running per the manual!

IMO, as long as it'll disconnect, that's enough. It's not like I turn into a passenger when I engage the autopilot - I'm still paying attention to what's going on, and if there's any sign of trouble, I'll hit the red button.

STEC-50
 
So is there any other autopilot besides the KAP140 that recommend that you overpower it on the ground?
 
Oh for God's sake, the manufacturers know better than you and they ask you to test it. Just shut your pie hole, stop thinking (because you're not good at that) and follow the damn instructions okay?

Sorry, just read this again and got ****ed...again.

What a prick.
 
Sorry, just read this again and got ****ed...again.

What a prick.

Looks like he got banned, so we don't have to deal with him anymore.
 
Last edited:
Back
Top