Autopilot Rolls Into Turns Like A Fighter Pilot

[dbwindom]

My first A/P use after upgrading saw ROLL SLIP in light turbulence and an uncommanded “fighter pilot” roll to the right and nose drop, as if rolling in on a ground target! At 170 knots, it was a little disconcerting. Would NOT have wanted to see such action while in the soup on an approach.
It was when I was expecting a left turning intercept of my GPS course.
Also, my fuel flow/use is no longer “spot on” , as it was before upgrade.
I also am getting “Message” audio alerts, but NO new messages.
Lastly, I am not confident I am getting all the Traffic displays I should.

 

[lancair360]

Catching up on some reading for plane, been neck deep in the Boeing 737 for a while.

On my Lancair it would roll smartly into and out of turns, no roll slip, worked great. After 16.2.4 I get roll slip now. Will have to adjust sensitivity. This update the first big step backwards....roll slip, tach error (didn't affect me) and error adding fuel in the EMS. 1 star, would not recommend.

 

[dmanp]

I have the same issue, AP does very sharp turn initiation now after 16.2.4 upgrade. Was always smooth before. Cruise TAS of 110 in a Sport Cruiser. It’s a fairly disconcerting difference from before. Hope this gets worked out soon.

 

[TrevorMills]

I decided to play with the autopilot settings to see if I could correct the 'aggressive' roll during course changes. The first settings I noticed on the 'Roll Axis' screen are labeled "Low Speed Sensitivity' and 'High Speed Sensitivity'. Not sure if they were there before V16.2.4 ( I don't remember them ). Anyway I changed the 'High Speed' setting from 20 to 12 and it seems to have corrected the problem. Rolling into and out of turns is smooth with no 'aggressiveness'.

Not sure it's technically the correct solution but it has certainly helped to make flights on AP much more comfortable.

It would be interesting to see if others find this solves their issue too.

 

[lancair360]

I had to turn mine down from 20 to 12 as well. Fixed the problem rolling into a turn. If I'm rolling into a turn and then reverse the turn I still get roll slip, but that should happen in normal ops so no big deal.

 

[swatson999]

I don't use the A/P that much, so hadn't noticed, but yep...I see the same thing. When it initiates a turn, it's like BOOM, it snaps over right away, then slows to a normal roll-in to a turn. Roll-outs are normal. Adjusting the Heading bug or NAV courseline or what have you, and BOOM, does it every time. I messed with the sensitivity a bit, but only just to futz around and see if it affected it, not as part of any tuning procedure. Lowering it to 8 *seemed* to help, at first, but it was still there.

It's like it's got a step function in the entry to the PID loop, going from 0 to the desired angle as quickly as possible on entry into it.

 

[Rhino]

It's like it's got a step function in the entry to the PID loop, going from 0 to the desired angle as quickly as possible on entry into it.

Don't say loop. You might give it ideas.

 

[jcarne]

Bump this request.

And to paraphrase Dr. Evil: Throw us a frickin' bone here! We're the customers ... need the info!

+1 on the bump.

Dynon I hate to be this guy but if I'm being honest some of these known issues that have not been dealt with make me feel like existing customers are not being valued. I know that supply chain issues have really put you guys in a crunch to find substitutes so I'm really trying to be patient but my patients is really starting to morph into other emotions. This autopilot issue is a big deal at times and probably a major deal for those that fly IMC. Really hoping it gets fixed sometime soon.

 

[swatson999]

I'm beginning to agree with jcarne about the lack of progress on a lot of known issues. And supply chains don't (or shouldn't) affect *software* development.

 

[Rhino]

It could affect it if hardware changes caused by supply chain problems are tying up developers trying to adapt software to the different hardware. I'm not saying that's the reason. I'm saying there are valid reasons that may not be obvious to us who are not involved in the operation. I spent many years in technology, and saw things like that happen all the time, so I'm most definitely giving them the benefit of the doubt. I have a hard time believing developers are just sitting around at Dynon staring at the ceiling. I think they just aren't communicating well with the customer base, which is hardly anything new for Dynon or any other technology company.

 

[AlanR]

I wonder if this problem of increased aggressive roll is just happening on Skyview HDX versions, because updating to 16.2.4 on Classic D1000's in my Bristell NG5 appears to have made no difference to roll at all!
The autopilot works perfectly as before without displaying any extreme or aggressive fighter roll type symptoms. I am using exactly the same settings as I did before updating the software.

 

[swatson999]

I have a SV Classic Touch, and it exhibited this behavior.

 

[Seagull]

I made a small course with left and right turns, the legs were 1-3 miles long. I flew the course with 16.0.1 then landed and upgraded to 16.3.2, and re-flew the same course. There was no difference in the way the plane navigated the course or the turns. I flew both at about 4400' and 100kts.
The photos were taken in the same turn of the course, you can zoom on the Dynon to see position.

What I did notice is the alert audio has increased with 16.3.2. I had it at 52% in 16.0.1 and had to lower it to 15% in the latest 16.3.2.

the course

firmware 16.0.1

firmware 16.3.2

 

[Dynon]

A short update here: We're still driving to root cause on this, and don't yet have it identified yet. We plan to address the behavior in our next software update, which currently doesn't have a scheduled release date. Nothing in the 16.2.4 ostensibly should have changed autopilot behavior at all, though, so it's a bit perplexing.

 

[Seagull]

Something new I just found out is when I bank hard I get an audio alert “bank angle”. How do I get rid of that, I can’t find any reference to it in the menu?

 

[Seagull]

I

It's a new alert with v16.3, as of May 2022.

"Bank Angle - The autopilot is engaged, and the bank angle of the aircraft is greater or equal to set Bank Angle Limit plus 5 degrees. The alert is cleared when the autopilot is disconnected, or the bank angle is less than the set Bank Angle Limit. The audio alert repeats every 5 seconds until cleared."

Next time, try pressing the AP disconnect button twice within one second, to turn off the Flight Director and the alerts that go with it.

See this thread for more info:

Nuisance "Bank Angle" Warning after Software Upgrade

Anyone else having this? Upgraded to the latest software version yesterday and flew today....the flight involved assertive autopilot off maneuvers (generally non-aerobatic) as part of formation flying (rejoins, etc.) 60-70 degrees of bank, pitch +/- 15 degrees. Any bank angle over about 30...

forum.flydynon.com

Thanks, that solved it. I have the AP button on my grip so it is real easy.

 

[jcarne]

Gents, Don reached out last week with news of engineering identifying and fixing the issue with the autopilot. I had a chance to test it today and everything seems to be back to working. I could not replicate the previous issue with either the heading/track bug or the nav feature. I only have about 1 hour on the update but I haven’t noticed any issues as of yet with anything else. So if you want a fix for the autopilot you might reach out to Don as I don’t know when they will issue a full update. I will do some approach testing tomorrow and if anything weird is found I’ll report back. The update he sent me was 16.2.7.9443. If you are newer than that you will require two other files from him as well as a force network load on the modules (a lot easier to do than it sounds).

 

[Seagull]

Thanks for the update.

 

[jcarne]

I'm glad they found the issue (and a fix). Did he say what caused the problem, since Engineering originally indicated that they made no changes that would affect the AP performance, and this AP bug seemed to affect some folks but not others. Curious minds want to know.

I would be hesitant to try this version of the SW, as it may not contain the Tach Time fix, which also affects my airplane.

Thanks for the info.

He did not say, I am curious as well. I am also curious on the tach time as I had that issue too before 16.3. I will look at the data tomorrow and see if that is fixed too. I’m hoping so as this is 16.2.7 and not 16.2.4 where I really had issues.

Since I am doing instrument training right now an autopilot fix is higher on my priority list than tach time but I’m really hoping that fix is packed in there as well.

 

[jcarne]

Okay guys, more testing done today. The autopilot seems to be working normally again. As mentioned though by RV8JD the tach time issue could be back. I can confirm that it is back. I have reached out to Don to see about getting a fix. So, if you have different ignition types and don't want the tach issue to deal with I would hold off until an update is available with both fixes in it.

As far as the autopilot goes it seems to fly normally again including approaches. You will however have to tune the pitch axis again. No biggy as this is pretty much standard ops with any update it seems. Hope this helps.