GFC500 problem

[PT20J]

1 hour ago, Ragsf15e said:

I am surprised how hard the servo has to work in the middle.

What happened initially is that I dialed the speed back from ~140 KIAS to ~95 KIAS and the AP overcorrected and then about the time it was beginning to settle down, I increased the power which caused it to overcorrect again.

[carusoam]

Some things stand out….

1) IAS climbs have always been more challenging to get right…

2) Does Big G give a specification on what we should reasonably expect?

3) What sensor is being used to maintain the IAS… one supplied by Big G?  Or something that has been in the plane for decades, not intended to be a control sensor?

4) the process controllers are getting faster and more powerful… are the sensors being updated to match?

5) When specifying a control sensor… there is a response time and accuracy discussion to go with that…

6) When comparing to other APs… the computer speed is also an important detail… like our personal computers measured in mhz…

 

PP thoughts only, not a controls engineer…

Best regards,

-a- 

[PT20J]

@rbp it might be interesting to look at the AP Pitch Torque plot during your flight and compare it to mine. If the torque isn’t a lot higher that might eliminate control friction. 

[PT20J]

I had my plane in the shop this week and asked the original installer to check the pitch servo cable tension. What he told me was interesting. He said that the turnbuckle Garmin uses can only be adjusted in increments of a full turn. During the original installation, he set the tension to the lower limit because another turn was too tight. This week he found it was below the low limit and he was able to take up a turn and adjust it to the high limit. Afterwards when I flew the airplane it did not oscillate significantly in IAS mode with large changes in the IAS target setting or power changes. 

So, perhaps the autopilot is very sensitive to cable tension and setting to the high limit might improve things. The measurement is a bit complicated as I believe the Garmin limits are specified at 70 deg F and corrections have to be calculated for the actual ambient temperature.

Skip

[jamesm]

sure would have been nice if Garmin had the installer replace the original bell crank with one with an extra horn for the autopilot to use push-rod type control rather than cable.

[PT20J]

17 minutes ago, jamesm said:

sure would have been nice if Garmin had the installer replace the original bell crank with one with an extra horn for the autopilot to use push-rod type control rather than cable.

Garmin pretty much copied the Bendix-King installation.

[EricJ]

23 minutes ago, jamesm said:

sure would have been nice if Garmin had the installer replace the original bell crank with one with an extra horn for the autopilot to use push-rod type control rather than cable.

The existing Mooney pushrods and belcranks have accomodations for bolt-on pushrod servo connections, and these were used by Century and S-Tec autopilots.   There's no need to change anything to do it that way, the factory provided direct access natively.   It's a bit of a kludge (imho) to add the cable controls and pulleys and bridles, but King did it that way and Garmin copied it, even though Garmin has pushrod-style servos, too.   It was definitely a missed opportunity and it sounds like it would have avoided some of the existing problems.

[rbp]

On 2/11/2023 at 11:26 PM, PT20J said:

I had my plane in the shop this week and asked the original installer to check the pitch servo cable tension. What he told me was interesting. He said that the turnbuckle Garmin uses can only be adjusted in increments of a full turn. During the original installation, he set the tension to the lower limit because another turn was too tight. This week he found it was below the low limit and he was able to take up a turn and adjust it to the high limit. Afterwards when I flew the airplane it did not oscillate significantly in IAS mode with large changes in the IAS target setting or power changes. 

So, perhaps the autopilot is very sensitive to cable tension and setting to the high limit might improve things. The measurement is a bit complicated as I believe the Garmin limits are specified at 70 deg F and corrections have to be calculated for the actual ambient temperature.

Skip

One of the items on the Garmin troubleshooting. page is the tensions. they want them at the high end of the range. Mine were adjusted, but it didn't help thing. Then again, my plane is a newer than yours Skip, so that may be why 

[rbp]

OK, this is the same flight as I posted above, but it includes both the AP Pitch Torque % and AP Pitch Deg

Lotta work going on 

[rbp]

It seems like the airframe is now completely up to Garmin specs.

The only thing left to check is that there are no blockages in the Pitot system, as per G.

this will require pulling some panels off

>>>I hate to have you inspect the entire length of the pitot system, but I really don't have any other solutions at this point. The last log was showing some really odd behavior with the air speed data fluctuating and the only reason we're needing to verify everything is based off of history. We had another airframe in TX which had very similar behavior and the dealer insisted that it was good, only after inspecting it very closely did we find a small mud dobber wasp obstruction and after clearing it the problems cleared up.

[PeytonM]

I hope you guys are having some success. 160 hours flying with the equipment in 10 months and it still oscillates. Garmin visited several months ago and everything checked out. I have no timeframe on a resolution. I have just about given up. Bummed. 

[rbp]

1 hour ago, PeytonM said:

I hope you guys are having some success. 160 hours flying with the equipment in 10 months and it still oscillates. Garmin visited several months ago and everything checked out. I have no timeframe on a resolution. I have just about given up. Bummed. 

oscillates in which modes?

[hais]

On 2/11/2023 at 8:26 PM, PT20J said:

I had my plane in the shop this week and asked the original installer to check the pitch servo cable tension. What he told me was interesting. He said that the turnbuckle Garmin uses can only be adjusted in increments of a full turn. During the original installation, he set the tension to the lower limit because another turn was too tight. This week he found it was below the low limit and he was able to take up a turn and adjust it to the high limit. Afterwards when I flew the airplane it did not oscillate significantly in IAS mode with large changes in the IAS target setting or power changes. 

So, perhaps the autopilot is very sensitive to cable tension and setting to the high limit might improve things. The measurement is a bit complicated as I believe the Garmin limits are specified at 70 deg F and corrections have to be calculated for the actual ambient temperature.

Skip

Is there a requirement for minimum reaction time? I think a slow control loop - one with a stronger damping - wouldn't be sensitive to cable tension. 

[PT20J]

9 hours ago, PeytonM said:

I hope you guys are having some success. 160 hours flying with the equipment in 10 months and it still oscillates. Garmin visited several months ago and everything checked out. I have no timeframe on a resolution. I have just about given up. Bummed. 

Can you confirm your airplane model is an M20J? (The J was the last model with trim bungees; subsequent models had a bobweight and variable down spring. Not sure if that makes a difference or not.)

Which modes does it oscillate in: PIT, ALT, VS, IAS?

Thanks,

Skip

[PT20J]

43 minutes ago, hais said:

Is there a requirement for minimum reaction time? I think a slow control loop - one with a stronger damping - wouldn't be sensitive to cable tension. 

Loose tension would look to the servo system like lost motion in the control system which is difficult to compensate because the system is essentially open loop around the trim point. I had mine checked because during annual inspection it just felt loose to me. Perhaps the tension spec is too low. 

[PT20J]

On 2/16/2023 at 12:26 PM, rbp said:

OK, this is the same flight as I posted above, but it includes both the AP Pitch Torque % and AP Pitch Deg

Lotta work going on 

Yes, that's different from mine. In my case, it oscillated a little initially and then settled down. Tightening the cable tamed the initial oscillation. But yours never settles down and the peak to peak oscillations are greater than mine.

It cannot be the software since the same software works in other airplanes. Since all the servos are the same mechanically, I wonder if it's possible to swap the trim and pitch servo as a test to eliminate any pitch servo issue. Might have to reload the software. 

 

 

[rbp]

26 minutes ago, PT20J said:

Yes, that's different from mine. In my case, it oscillated a little initially and then settled down. Tightening the cable tamed the initial oscillation. But yours never settles down and the peak to peak oscillations are greater than mine.

It cannot be the software since the same software works in other airplanes. Since all the servos are the same mechanically, I wonder if it's possible to swap the trim and pitch servo as a test to eliminate any pitch servo issue. Might have to reload the software. 

 

 

interesting idea! 

[PT20J]

3 hours ago, rbp said:

interesting idea! 

If you have a yaw damper it might be easier to swap that servo with the pitch servo.

[201Mooniac]

In all the problems listed, I haven't seen the yaw damper be mentioned.  I've had issues with mine since it was installed and no resolution as of yet.  My issue is that it just doesn't really keep the ball centered.  It is fine in straight and level flight and does work in turbulence but it barely keeps the ball centered in AP driven turns and certainly doesn't come close in climbs.  It is even occasionally out a little in decent.  I did have an issue for a short while where it would oscillate but tightening the cable tension seems to have fixed that.  Anyone else see a yaw damper issue?

[Mooney in Oz]

2 hours ago, 201Mooniac said:

In all the problems listed, I haven't seen the yaw damper be mentioned.  I've had issues with mine since it was installed and no resolution as of yet.  My issue is that it just doesn't really keep the ball centered.  It is fine in straight and level flight and does work in turbulence but it barely keeps the ball centered in AP driven turns and certainly doesn't come close in climbs.  It is even occasionally out a little in decent.  I did have an issue for a short while where it would oscillate but tightening the cable tension seems to have fixed that.  Anyone else see a yaw damper issue?

The yaw damper doesn’t provide trim function. Its purpose is to dampen yaw (tail wagging) caused by environmental factors such as turbulence.

[LANCECASPER]

10 hours ago, Mooney in Oz said:

The yaw damper doesn’t provide trim function. Its purpose is to dampen yaw (tail wagging) caused by environmental factors such as turbulence.

Exactly. A Yaw Damper is not  rudder trim.

Using it on a long climb without rudder trim or keeping some right rudder in would take a servo designed to make small adjustments in cruise flight and keep it at max workload for an extended time during the climb. (edited)

[PT20J]

1 hour ago, LANCECASPER said:

Exactly. A Yaw Damper is not  rudder trim.

Using it on a long climb would take a servo designed to make small adjustments in cruise flight and keep it at max workload for an extended time during the climb.

There was a post by Garmin on BeechTalk explaining that the damper is not intended to replace rudder trim.

However, Garmin user documentation claims that it will keep the ball centered during climbs and coordinate turns. 

Some have reported that it does, and some have reported that it doesn’t.

 

[donkaye]

4 minutes ago, PT20J said:

There was a post by Garmin on BeechTalk explaining that the damper is not intended to replace rudder trim.

However, Garmin user documentation claims that it will keep the ball centered during climbs and coordinate turns. 

Some have reported that it does, and some have reported that it doesn’t.

 

It's purpose is to damp Yaw.  I use it in an extended climb to do just that.  On level off I turn it off, trim the rudder with the rudder trim switch, then turn it back on.

[201Mooniac]

Yes, I understand it is not intended to replace rudder trim however I was told by Joey F from Garmin that in an M20J it should keep the ball centered, even in most climbs.  As I mentioned, it won't keep the ball centered even in level turns and that the documentation does state it should do as you also mention.  I also on occasion find the ball a quarter width out of center in straight and level flight.  The installer has leveled and re-calibrated but nothing has changed.  It might be an airframe issue I guess but it seems to fly with the ball centered with the AP off.

[rbp]

my technique is that after changing power settings, i turn off the YD, center the ball with the rudder trim, and re-engage the YD. the larger the power change, the more rudder trim adjustment