Menu Sign In Contact FAQ
Banner
Welcome to our forums

DFC90 - Sudden "Autopilot Disconnect" in Cruise

No, it could just be that you servo was on the way out. You have to be able to trust it (to an extent) as that is the whole point.

In the PA-46 it used both pitch and pitch trim servos. If it couldn't work the pitch servo it would use pitch trim which would eventually trigger the TRIM IN MOTION message if it ran for too long.

If it worked perfectly for so long then it is more likely the servos of course.

EGTK Oxford

alexisvc,

One idea might be to call the guys at Avionik Straubing and describe your problem to them (also, it is just a short hop from our home field). They might have encountered this problem before and can certainly be more outspoken about a specific product/problem than the manufacturer of said product.

Their specialist for APs is Martin Scheifl.

Good luck with the trouble shooting.

RXH
EDML - Landshut, Munich / Bavaria

If you put a DFC90 into the SR20 it will use the pitch trim.

It will do the very same thing in a SR22 and a C182 and a PA46. This autopilot is generally programmed to do everything with the pitch trim servo and only use the pitch servo if it needs a faster reaction (flaps, gear, etc.).

This took me a while to understand. In the Cirrus you don't see it because there is no manual trim but in the C182 the trim wheel is constantly moving, it basically never stops and at first, both my avionics installer and I thought this was a malfunction.

You should be able to trust the autopilot.

The way the DFC90 implements the pitch axis is simply wrong. It should drive the pitch servo and then trim out the torque on that if it goes outside the limits.

An early failure of a roll servo is nothing to do with any kind of greater precision. It is simply poor software which drives the servo backwards and forwards all the time. In normal flight there is no need for such constant control surface motion and in any case the control linkage play is substantial.

Administrator
Shoreham EGKA, United Kingdom

The way the DFC90 implements the pitch axis is simply wrong. It should drive the pitch servo and then trim out the torque on that if it goes outside the limits.

That is a very bold statement and I think it is unfounded. It is different from how the others work but it works extremely well which is all that matters.

The DFC90 basically locks the pitch servo in position and almost never drives it. Same for the roll servo, it only moves when necessary, I specifically checked for that when I tried to better understand the design of this AP.

This is what I got from Alec Vincent of RGV in Gloucester (who are the CAMO for my plane):

The roll trim motor start up voltage should not be more than 2.0 V DC, if this is slightly high and you had a reasonable X wind that may be an issue. Might be worth carrying out an in-flight roll trim tab test, see below from SR22 MM. If the Tab is out it could enhance any issue caused by the above factors

He also sent me the procedure from the Maintenance Manual how to test the position of the roll trim tab and how to adjust it ... I will try that before I do anything else.

RXH: Yes, that was my plan anyway, I know those guys well! I also told Avidyne to contact AS and to set up a strategy to find the failure. Why pay thousands of "Flex Care" Warranty and not use it :-)

The roll trim motor start up voltage should not be more than 2.0 V DC, if this is slightly high and you had a reasonable X wind that may be an issue

I'd like to know how the DFC90 manages to establish the relative motion between the frames of reference of the earth's atmosphere and the earth's surface (commonly known as "wind")

That's a joke, surely.

What is the P/N of the roll servo? I will try to get the MM for it, or at least a data sheet.

To suggest a minimum control voltage for the servo to start turning is reasonable. I don't know what these are but the King servos use -9V to +9V but they don't have a lower threshold as such. There is an internal tachometer so even if you applied say 0.5V they will still run, due to the feedback loop in there. The STEC servos may be open-loop, like all the earlier King ones.

But all that checks for is whether the mechanics are seized up, or the commutator is shagged.

That is a very bold statement and I think it is unfounded. It is different from how the others work but it works extremely well which is all that matters.

When you fly a plane, you first control the attitude (pitch and roll) and second control the trajectory (altitude and heading) and third control the navigation (LNAV/VNAV etc). You don't sit there twiddling the yoke left/right for the roll attitude and twiddling the pitch trim wheel for the pitch attitude. It works, but you get a poor transient response. Also the pitch trim jackscrew is going to be going up and down like somebody's knickers.

Administrator
Shoreham EGKA, United Kingdom

Peter, I'll get you the number, have to check the the Maint. Manual!

You don't sit there twiddling the yoke left/right for the roll attitude and twiddling the pitch trim wheel for the pitch attitude. It works, but you get a poor transient response. Also the pitch trim jackscrew is going to be going up and down like somebody's knickers.

Fair enough. However, this method works extremely well. Have a look at all the DFC90 test reports, every single of them says how precise and smooth the autopilot operates, only comparable to the GFC700. The AP has zero tendency to oscillate or over-react. It is an extremely well working system so it can't be that obviously wrong...

The SR20 does not have a pitch servo although all of them have a 2-axis AP (S-TEC 55x, now Garmin GFC700 and DFC90 retrofits).

To suggest a minimum control voltage for the servo to start turning is reasonable.

The DFC90 installation manual has very strict requirements for the voltage thresholds and part of the install is to verify that they are met. Avidyne state that the AP will not function correctly if the servos are not within the allowed range. The other important thing is cable tension, the DFC90 is very sensitive there.

I have the installation manual available in case you need to know something.

Now what's REALLY interesting:

  • Cirrus Service CEnter says that the +/- 3.5 Volts are "normal"
  • Avidyne says it should not be over 2 Volts and that it will not work ...

(no comment)

But I am pretty sure that the trim tab has to be adjusted as well, and (other than Peter :-) I am sure that a constant strong crosswind DOES create a greater load on the load servo. After all in GPSS mode it works a lot to keep the A/C precisely on course

Sign in to add your message

Back to Top