GFC500 problem

[rbp]

These graphs are from my G3X logs 00:43 - 00:47, that show a climb from 2800ft to 7500 ft using the autopilot in IAS mode at 120.

The first graph is FD pitch commands versus ALT. As you can see, the altitude gain is quite smooth, but the pitch commands vary wildly, including some 10+ degree up pitch commands and one of 15, which were quite concerning during the flight. 

the second graphs is IAS versus pitch. You can see how bumpy it is in level flight from 00:47 onward

As I mentioned previously, it was a very bumpy day of moderate turbulence and windshear, so one might expect IAS to fluctuate. You can see IAS versus VSI in the second file.

 

Edited April 29, 2022 by rbp

[jetdriven]

20 hours ago, donkaye said:

If Garmin is monitoring this thread, and they have been known to do so, they are in big trouble.  Before i knew that it was illegal to change the settings on the G5 for a certified installation, I changed mine and was raked over the coals by Trek.  I quickly changed mine back.  These systems are significantly tested through all sort of tests, and, although it might seem to correct the problem under a certain set of circumstances, that doesn't mean it couldn't be unsafe under another untested set.

Bottom line, if you value your license, your insurance, and maybe your life, I wouldn't make ANY changes to your system that is not approved under the STC.

I’m not condoning it I’m just merely mentioning that I know a guy who changed the gain Settings and it fix the problem. It is fixable.  It seems that Garmin may be more Concerned about what looks proper than what actually works. It can be fixed, you do some more flight testing and you change the settings to where it doesn’t porpoise in cruise, and then you issue a service letter.

There’s a whole thread over on Beechtalk regarding a  -3° down angle in cruise in a bonanza in level flight. And we all know that airplanes really don’t fly with a nose below the horizon…., and it’s the same thing a Garmon rep gets  at all high and mighty and said you are  illegal and you are breaking the law and your plane in Unairworthy if you reset the pitch to zero in Cruise even though every Aircraft Theyve ever flown  including about six or eight jet transport airplanes also show +2 to +3 positive pitch in cruise.  And that the attitude indicator they removed also had an adjustable Horizon bar to compensate for this. Anyway, I’ve seen this regarding other issues and although I appreciate their stuff  should be documented and airworthy, the stuff ought to work especially when you’re spending several cubic yards of cash on it.  You can have both but it takes  manpower to do it.   My opinion only

[PT20J]

Garmin doesn't really explain a lot about how the GFC 500 works. But in various manuals it is described as "rate based." Rate based autopilots are usually twitchier than attitude based autopilots in my experience. I find the GFC 500 IAS mode only works well if it is engaged with the airplane trimmed and the airspeed constant and the air reasonably smooth. Otherwise it seems that the pitch response is poorly damped. VS is better, but if the air is rough, I revert to PIT. 

Skip

[bmcconnaha]

Concerning thread, as mine is in the shop getting a GFC500 installed, along wit the a g3x system.  Even more concerning is the fact that @PT20Jsays its a rate based AP.  I was under the impression it was a fully digital attitude based system. 

FWIW this is what garmin says “The revolutionary GFC 500 autopilot brings unprecedented levels of capability, reliability and affordability to light piston, fixed-wing aircraft. GFC 500 leverages our advanced attitude-based flight control technology developed for the popular GFC 700 autopilot, which drives our most advanced glass flight deck systems”

[carusoam]

Back in the day…

Process controllers got tuned by hand by entering three values… proportional, integral, derivative… (PID)

Each variable got tuned separately…

The initial set-up probably has a few interesting steps as well… zeroing things out to eliminate funny displayed numbers…

 

In the modern era…

Process controllers get auto-tuned… where the qualified technician sends it through it paces, and the controller self calculates the PID values…

These values may have been determined during the STC process…? And now being applied to all identical planes in the family….?

 

It would be interesting to see how these things get tuned to the specific plane… and how much WnB plays a role in the plane’s responses…

 

PID tuning is often an electrical engineer’s realm…

PID tuning for an airplane… would invite being a Master CFI, and Electrical engineer… at the same time…

 

Proper adjustment… the oscillations get smaller until they settle out, quickly…

Improper adjustment can result in… oscillations increasing uncontrollably…

There must be some interesting procedures for tuning these things properly…

 

The Garmin manuals for both user and installer must be gigantic…

PP thoughts only, not a controls engineer…

Best regards,

-a-

[PT20J]

1 hour ago, bmcconnaha said:

Concerning thread, as mine is in the shop getting a GFC500 installed, along wit the a g3x system.  Even more concerning is the fact that @PT20Jsays its a rate based AP.  I was under the impression it was a fully digital attitude based system. 

FWIW this is what garmin says “The revolutionary GFC 500 autopilot brings unprecedented levels of capability, reliability and affordability to light piston, fixed-wing aircraft. GFC 500 leverages our advanced attitude-based flight control technology developed for the popular GFC 700 autopilot, which drives our most advanced glass flight deck systems”

I probably overstated it. It’s probably a hybrid using pitch and rate info. In some modes it does seem to exhibit some of the lag and oversensitivity characteristic of fully rate based systems.  But, it is a very capable autopilot. You just have to know it’s quirks — like it will only work in ROL or HDG lateral modes with no valid GPS signal, so if they’re jamming GPS it won’t track a VOR or ILS.

[rbp]

if I am seeing +10 degree pitch attitudes in vertical modes, then they are not limited enough! 

I will also add that the YD was extremely effective in the bumps on weds and thurs flights. I noticed that when I disconnected the YD, there was a lot more yawing motion, so i test flew with it on and with it off, and there was a noticeable difference. back seat pax will probably notice it even more

[Ragsf15e]

1 hour ago, rbp said:

if I am seeing +10 degree pitch attitudes in vertical modes, then they are not limited enough! 

I will also add that the YD was extremely effective in the bumps on weds and thurs flights. I noticed that when I disconnected the YD, there was a lot more yawing motion, so i test flew with it on and with it off, and there was a noticeable difference. back seat pax will probably notice it even more

Did you try climbing out in IAS mode but without the AP engaged (hand fly) and then see what the flight director does, especially with minor airspeed deviation?  No way you should see 10 degrees, but if the Fd is all over the place, that would narrow down the problem.

[Tom 4536]

12 hours ago, PT20J said:

But, it is a very capable autopilot. You just have to know it’s quirks — like it will only work in ROL or HDG lateral modes with no valid GPS signal, so if they’re jamming GPS it won’t track a VOR or ILS.

That was a shock to me. I recently had a Garmin 255 installed to have VOR capability if/when GPS is jammed. My GFC will track either of my VOR (430 or 255) signals when the GPS signal isn't jammed. Are you saying the AP will not follow the VOR when GPS is jammed or am I mis-interpretting what you said. Thank you for your valued input.

[PT20J]

43 minutes ago, Tom 4536 said:

That was a shock to me. I recently had a Garmin 255 installed to have VOR capability if/when GPS is jammed. My GFC will track either of my VOR (430 or 255) signals when the GPS signal isn't jammed. Are you saying the AP will not follow the VOR when GPS is jammed or am I mis-interpretting what you said. Thank you for your valued input.

The GFC 500 uses “GPS aiding” for lateral navigation. In a G3X installation, it can use the G3X internal GPS if the primary (GTN, GNS) navigator fails. But, if all GPS capability is lost, NAV and APR modes are unavailable even if a VHF navigation source is selected. I verified this with Garmin and also tested it by shutting down the GTN and G3X. You can, of course, still hand fly any approaches, and use HDG for en route navigation.

[rbp]

1 hour ago, Ragsf15e said:

Did you try climbing out in IAS mode but without the AP engaged (hand fly) and then see what the flight director does, especially with minor airspeed deviation?  No way you should see 10 degrees, but if the Fd is all over the place, that would narrow down the problem.

If you look at the graphs I posted above you will see the flight Director pitch commands

[PT20J]

1 hour ago, rbp said:

If you look at the graphs I posted above you will see the flight Director pitch commands

Interesting test result. Looks like a pretty wild ride! If you get another day of atmospheric discontent, try VS and PIT mode to see what gives a better ride. I’ve not been in bumps that bad, but I haven’t been impressed with IAS mode when it’s bumpy. 

I don’t know enough about the GFC 500 design to know its limits, but I think most autopilots have a point at which large pitch excursions will cause a disconnect. I would be concerned about the autopilot giving up and disconnecting in a high pitch up or down attitude, and possibly out of trim. This is theoretical — I don’t know if it will do that.

Skip

[Ragsf15e]

2 hours ago, rbp said:

If you look at the graphs I posted above you will see the flight Director pitch commands

Yeah, I saw that.  My thought is to take the ap servos completely out of the equation though and see if the fd acts any different.  Maybe it gives the same commands, maybe not.  It would be interesting to know 

[carusoam]

Do we have any idea at how many cycles per second this hardware works at?

And if there is a difference in one mode vs. another?

The older APs did a lot with very low computer speeds…

Today, the computer speeds must be incredibly high in comparison?

PP thoughts only, not a computer guru…

Best regards,

-a-

[Davidv]

I’ve said this in some other threads, but I spent 6 months chasing a similar only to found out it was bad servos.  Once the Servos were replaced it was 95% better.

[LANCECASPER]

I have a GFC700 which I really like, but had a KFC150 in a previous airplane that worked very well and wasn't ready to be replaced.

If I had a Bravo or Ovation with a KFC225 that was working well there is no way I would rip it out and do the GFC500. The KFC225 was a top of the line $40,000+ autopilot 20 years ago. Although the GFC500 is a newer design and does interface nicely with other Garmin products, just looking at the servos from both, the KFC225 is a much more robust design. Where people get at odds with their King autopilot is if they or the previous owner have taken it somewhere that doesn't have intimate knowledge and they do more harm than good. If replacing a Century or King at the end of service life I can see going to a GFC500, it makes economic sense, but IMHO there is a lot of life in the KFC225 if serviced by someone who really knows the autopilot. 

[FlyingDude]

On 5/1/2022 at 6:59 AM, Davidv said:

I’ve said this in some other threads, but I spent 6 months chasing a similar only to found out it was bad servos.  Once the Servos were replaced it was 95% better.

What was actually "bad" with them?  Did they tell you?

[carusoam]

7 hours ago, FlyingDude said:

What was actually "bad" with them?  Did they tell you?

There were quite a few of the early installs that had bad servos…

Fortunately, it didn't take too long for Big G to figure out serial numbers for the batch… or whatever group was affected…

There is plenty of reports about them around here somewhere…

Best regards,

-a-

[Davidv]

On 5/2/2022 at 4:06 PM, FlyingDude said:

What was actually "bad" with them?  Did they tell you?

No, they didn’t know, only that there was something internal with the clutches that showed there were some defective pieces.

[dzeleski]

I do not have issues with my GFC500 when level but using IAS climb specifically it hunts even in clean air. The situation is made much worse if you set your targeted speed instantly. Its so bad that in IMC it makes you feel like you are going to flip backwards, not very pleasant. My installer is very aware of this situation and said it is "fixable" with gain settings but its not legal to do so we cant do it. Why Garmin doesnt give us a valid gain range instead of a fixed value is beyond me. These airplanes are all slightly different and small tweaks should be allowed within reason.

I work around this problem by slowly decreasing the IAS bug speed by 10, pausing, let it react, adjust, pause, adjust, etc. Annoying yes, but less annoying then the massive control inputs that otherwise would be added. Another work around is to set VS mode first at say 750 fpm, wait for it to stabilize and then select IAS. VS mode for me is always locked nearly perfect, I just dont like to stay in VS mode during climbs for obvious reasons.

Finally I did have one install issue and that had to do with the envelope protection. My airplane is in MPH and the protection speeds and v speeds are stored in knots. I went up and tested all of these and noticed they were all quite wrong especially during decent I was getting an airspeed warning way before Yellow or Red line. My installer reached out to Garmin on the proper solution (which is to just convert the speeds to mph) and they flat out asked if pilots are trying to kill them selves by testing these limits. Thats when I realized there is a serious disconnect between what is going on within that company and what is going on in real life.

[StevenL757]

@TrekLawler for comment.

[Larry]

When I change the IAS in climb mode, I try to not change more than 2 knots.  Much more than 2-3 knots and it will pitch up or down to quickly achieve the speed then reverse the pitch to hold it.  I'm not sure if that is a feature or a problem, I wish it were more subtle.  I also get the "hunting", but its only about +/- 2-3 knots.

[Larry]

OP here.  I followed some of Skip's suggestions.  In pitch and IAS mode it seemed better.  I tried alternate static, but it was too bumpy to say if there was a difference.

 

[PT20J]

3 hours ago, dzeleski said:

My installer is very aware of this situation and said it is "fixable" with gain settings but its not legal to do so we cant do it.

Not to knock your installer, but how would they know it was fixable if they hadn't successfully fixed it by tweaking the settings which they say they cannot do?

I have tried IAS mode in a C-172 as well as my M20J and they both behaved similarly. 

Skip

[dzeleski]

2 hours ago, PT20J said:

Not to knock your installer, but how would they know it was fixable if they hadn't successfully fixed it by tweaking the settings which they say they cannot do?

I have tried IAS mode in a C-172 as well as my M20J and they both behaved similarly. 

Skip

Because its allowed in the experimental world and they can tune it when it happens. Its specifically not allowed for certified installs.