If your AHRS is operating normally then it is not the cause of any issue.
The AHRS is used to detect CHANGES in attitude. Anything else is secondary to the AP function. The AP does not care if if you are flying left wing low or pitch up - it's only interested where the aircraft is going.
Practically all issues are related to coupling of the servos to the control surfaces in our experience.
The first thing to check is the shear screw on the servo arm - that needs to be tight. Once installation is near completion the screw should be fitted with a drop of locktight. If this is not done it WILL come loose and the AP has no chance of controlling the aircraft properly.
Similar should be done to the large grub screw in the servo boss - this needs to be tight and a drop of locktight needs to be applied so it stays that way. The forces acting on the servo shaft can be pretty big.
The AP relies on small corrections on the servo arm making it though to the control surfaces. If this does not happen you will get oscillation as the AP will forever be behind the aircraft - no possible setting you can do on the EFIS will fix that.
I would also like to mention another example: A sinus came to my attention where the pitch control was really bad. The installation was done well. When I got to see the aircraft and tried the servo test the entire floor of the aircraft would flex - the servo was mounted on a structure that was connected to the fuselage floor - and this was pretty thin to save weight. It was easier for the servo to flex the floor than move the elevator !
Look at your entire control linkage scheme - make sure a small movement of the arm WILL feed to the control surfaces and don't let gravity on a surface fool you - a small amount of play somewhere can have big consequences once it gets to the control surface. Make sure your mounting brackets etc do not flex under load.
In most cases it is pitch control that needs attention as roll is much more stable than pitch on most aircraft.
The AP is programmed to hold the aircraft within 50ft using a "herd mentality" - it will not force the altitude to 0ft but rather try gently. In calm conditions you can get there but in bumpy stuff it will hold off and correct only for the big stuff with some determination. This is to avoid unnecessary control inputs - much like you would do as a pilot.
Why Doesn't The Autopilot Do Its Job Properly?
Forum rules
Please keep your posts friendly and on topic. No politics or discussions of a controversial nature not related to our favorite subject of flying and avionics. Offending posts may be removed or moderated.
Please keep your posts friendly and on topic. No politics or discussions of a controversial nature not related to our favorite subject of flying and avionics. Offending posts may be removed or moderated.
Re: Why Doesn't The Autopilot Do Its Job Properly?
Thanks for the reply, Rainier, but I genuinely don't think it is anything to do with the mounting of the autopilot servos or the connections. Everything of mine has Loctite and is marked to show any movement. Hopefully the attachment will show the 'cage' in which the servos are mounted and secured to the airframe. The Garmin & Dynon servos are mounted using the same hardware.
The altitude keeping definitely strays outside of 50'. I would actually be quite happy with 50' but very often I will take the autopilot out when I'm flying below controlled airspace, as I don't trust it enough to stop me straying into it. When my wife is with me I'm under orders to hand-fly because she can't stand the way it wanders around in roll.
I would genuinely love to know what the others are doing so they don't get the same results in the same aircraft.
The altitude keeping definitely strays outside of 50'. I would actually be quite happy with 50' but very often I will take the autopilot out when I'm flying below controlled airspace, as I don't trust it enough to stop me straying into it. When my wife is with me I'm under orders to hand-fly because she can't stand the way it wanders around in roll.
I would genuinely love to know what the others are doing so they don't get the same results in the same aircraft.
- IMG_0583.jpeg (227.57 KiB) Viewed 8882 times
Re: Why Doesn't The Autopilot Do Its Job Properly?
Today was so calm and smooth I could have fallen asleep. Somehow my '50ft herd' turned into +700' (from the 3000' Alt Hold) and then I got bored and disconnected the autopilot at 2100' with it still descending. +700/-900......yeah, that's closeThe AP is programmed to hold the aircraft within 50ft using a "herd mentality" - it will not force the altitude to 0ft but rather try gently. In calm conditions you can get there
- tempImageFdKU5u.gif (234.34 KiB) Viewed 8871 times
Re: Why Doesn't The Autopilot Do Its Job Properly?
- tempImageg6LmQl.gif (152.97 KiB) Viewed 8871 times
Re: Why Doesn't The Autopilot Do Its Job Properly?
Is it possible that the servo is 'crossing zero' and causing the issue.
Re: Why Doesn't The Autopilot Do Its Job Properly?
No, that causes all sorts of chattering with the test and, normally, results in the stick going to one position and 'knocking' against the stops. I've seen that before and changed a guy's set up so it didn't (I got lucky with mine and it didn't go through zero from day one).
The autopilot works; just really not very well and yesterday's dead smooth weather gave it ample opportunity to behave itself (literally smooth, smooth and zero wind). To climb 700', descend 900' and still keep going is not acceptable.
The autopilot works; just really not very well and yesterday's dead smooth weather gave it ample opportunity to behave itself (literally smooth, smooth and zero wind). To climb 700', descend 900' and still keep going is not acceptable.
Re: Why Doesn't The Autopilot Do Its Job Properly?
No that is not acceptable.
That is not what the AP is doing on its own. It's got a problem controlling your pitch. You need to sort out the mechanics as I tried to explain. If you do that it will hold rock solid with only the tiniest of control input.
In a nutshell - if the AP changes the pitch servo by a small amount and nothing happens, it will continue increasing the pitch deflection (on the servo) until something does - then it overshoots as it's too much and then it starts to correct - with the same result.
Small changes are not feeding through to the control surface.
Just as example: I had a case with a pipistrel sinus with similar behavior. Very nice installation, could not fault it and servo perfect. But what was happening - as the servo moved - the first thing that happened was that the entire floor of the aircraft where the servo bracket was mounted to flexed - without anything happening at the elevator. Only once the floor could flex no more did the elevator surface start to move. Needless to say, once that little issue was fixed - it held pitch perfectly.
Not sure how your aircraft's pitch servo is coupled to the elevator - but I am sure we have an issue here. It's always like this. In the many years I have done this it always comes down to the same thing.
That is not what the AP is doing on its own. It's got a problem controlling your pitch. You need to sort out the mechanics as I tried to explain. If you do that it will hold rock solid with only the tiniest of control input.
In a nutshell - if the AP changes the pitch servo by a small amount and nothing happens, it will continue increasing the pitch deflection (on the servo) until something does - then it overshoots as it's too much and then it starts to correct - with the same result.
Small changes are not feeding through to the control surface.
Just as example: I had a case with a pipistrel sinus with similar behavior. Very nice installation, could not fault it and servo perfect. But what was happening - as the servo moved - the first thing that happened was that the entire floor of the aircraft where the servo bracket was mounted to flexed - without anything happening at the elevator. Only once the floor could flex no more did the elevator surface start to move. Needless to say, once that little issue was fixed - it held pitch perfectly.
Not sure how your aircraft's pitch servo is coupled to the elevator - but I am sure we have an issue here. It's always like this. In the many years I have done this it always comes down to the same thing.
Re: Why Doesn't The Autopilot Do Its Job Properly?
Today I flew two of the same aircraft. Same weather, same speeds, same everything apart from the colours. 1 equipped with Dynon HDX and autopilot. 2 (mine) equipped with iEFIS Lite and autopilot.
Flight in 1 resulted in a seamless WIFI route transfer from my iPad to the HDX, followed by 'NAV' doing it everything as it should; where the autopilot intercepted the route, maintained the altitude and did what an autopilot is meant to (similar to my professional life).
Flight in 2 resulted in aborted multiple attempts to load a GPX file into the routes, using a mini SD card and, therefore, no 'NAV' facility at all. Worse than that was a constant dive for the deck every time the autopilot was engaged in straight & level and more wandering than the RN fleets did during WWII to avoid being torpedoed. Absolutely unusable!
Combine all of that with the joy of beholding Zenith & Nadir, whilst cruising straight and level (hand flown) with nothing else going on and I realise, now, how stupid I was to have fitted any of this. I really should have saved my time and money and gone for Dynon in the first place
Flight in 1 resulted in a seamless WIFI route transfer from my iPad to the HDX, followed by 'NAV' doing it everything as it should; where the autopilot intercepted the route, maintained the altitude and did what an autopilot is meant to (similar to my professional life).
Flight in 2 resulted in aborted multiple attempts to load a GPX file into the routes, using a mini SD card and, therefore, no 'NAV' facility at all. Worse than that was a constant dive for the deck every time the autopilot was engaged in straight & level and more wandering than the RN fleets did during WWII to avoid being torpedoed. Absolutely unusable!
Combine all of that with the joy of beholding Zenith & Nadir, whilst cruising straight and level (hand flown) with nothing else going on and I realise, now, how stupid I was to have fitted any of this. I really should have saved my time and money and gone for Dynon in the first place
Re: Why Doesn't The Autopilot Do Its Job Properly?
Would you be so kind as to show:
A screen shot of your pitch control setup menus contents.
A screen shot of the autopilot engagement screen.
A screen shot of your pitch control setup menus contents.
A screen shot of the autopilot engagement screen.
Re: Why Doesn't The Autopilot Do Its Job Properly?
also check the minimum airspeed setting. I have an odyssey so not necessarily the same, but I assume there is a similar setting. I was experiencing similar behavior where the autopilot always wanted to dive, after years of working acceptably. somehow the minimum airspeed setting was corrupted, along with a handful of other settings in the memory; once I corrected that it was back to behavior as usual.
If it did change the minimum airspeed on it's own, along with other settings randomly: the root cause of my issue was a faulty internal SSD card. once that was replace the settings stopped randomly changing, and it's been functioning well again for a couple years.
ETA: PaulSS, I just read some of your follow up posts and see it's wondering high also not always a dive so this likely wont help you, but maybe it will some others.
If it did change the minimum airspeed on it's own, along with other settings randomly: the root cause of my issue was a faulty internal SSD card. once that was replace the settings stopped randomly changing, and it's been functioning well again for a couple years.
ETA: PaulSS, I just read some of your follow up posts and see it's wondering high also not always a dive so this likely wont help you, but maybe it will some others.
Last edited by odens_14 on Tue Jun 14, 2022 3:44 pm, edited 1 time in total.