Menu Sign In Contact FAQ
Banner
Welcome to our forums

Avidyne DFC90 and Garmin GFC700 autopilots, and behaviour with a frozen pitot tube

Can everyone please use the forum names in the @ references.

Most people here don’t know everyone involved personally. Also some may object. And most people won’t write frankly if their identity is blown.

Administrator
Shoreham EGKA, United Kingdom

This is what the DFC90 manual says about “Loss of Airdata”.
I wasn’t fully aware of it but this is a situation where the “Straight & Level” button comes in handy.

What does the autopilot use to maintain altitude in the S&L mode without anything from the ADC? Does it have its own pressure sensor?

Biggin Hill

I think it will use the AHRS for the straight and level function, but without the air data computer. I don’t think it can hold a specific altitude then, only a certain pitch and roll attitude.

Altitude doesn’t come via the pitot tube; it comes via the holes in the side of the plane, so you should always have a usable altitude. The King autopilots have a pressure encoder internally (pressure to 16 bits – a bit like this) which they use for altitude hold. The capture is based on the gray code input; the hold uses the internal baro.

Administrator
Shoreham EGKA, United Kingdom

We are talking about a “loss of airdata” here, and that does come from the ADHRS. If you lose the Air Data you have no altitude indication, only on the backup altimeter you do.

tomjnx wrote:

I still don’t see why velocity input should be mandatory.

I agree it should not be necessary in theory, and there are definitely many Kalman implementations that use gyros+accelerometers only. But my understanding is that the AHRS systems in question here (G1000 in particular) require (in addition to the gyros+accelerometers), either a GPS speed+heading reference or an airspeed+magnetometer heading reference. Failure of either both speed references (GPS and airspeed) or both heading references (GPS and magnetometer) at the same time, will cause a complete attitude reference failure.

I think this could be for one of a number of reasons:

  1. To enforce a higher level of integrity, erring towards showing no attitude reference rather than showing a suspect one. But this seems unnecessary and overly conservative given that a mechanical AI does not provide this integrity checking.
  2. To ensure that the AHRS provides an accurate attitude reference in the continuous-balanced-turn scenario where a continuous orbit causes the apparent gravity vector to draw the attitude indication to incorrectly read level while still banked. Knowledge of speed and rate of change of heading gives an independently calculated bank angle and allows detection of this and allows ignoring of the false gravity vector.
  3. High level of drift in the MEMS gyros used which means that supplementing with accelerometers alone is not enough to provide consistent performance.
  4. Arbitrary certification requirement with no technical basis!

ortac wrote:

2. To ensure that the AHRS provides an accurate attitude reference in the continuous-balanced-turn scenario where a continuous orbit causes the apparent gravity vector to draw the attitude indication to incorrectly read level while still banked. Knowledge of speed and rate of change of heading gives an independently calculated bank angle and allows detection of this and allows ignoring of the false gravity vector.

Just as your item 1, this seems overly conservative to me since a mechanical AI would have the same issues.

ESKC (Uppsala/Sundbro), Sweden

Yes I agree, although if the MEMS gyros have higher drift than mechanical gyros, the gravity vector may need a higher weighting in the Kalman algorithm and this would make the system more susceptible to the balanced turn issue.

Yes – I think MEMS gyros drift massively more than good quality mechanical ones.

But also the goalposts change over time. The “fake gravity vector in a balanced turn” problem is ignored with mechanical gyros because, ahem, there is absolutely no solution to it All you can do is the nonsense with the pendulous vanes, or the electrical equivalent, and these assume everybody will start the engine (the vac pump) with the plane not moving already, and that nobody will do too much turning. But with modern avionics there is an opportunity to fix this properly, so maybe it was decided that is necessary.

Administrator
Shoreham EGKA, United Kingdom
Sign in to add your message

Back to Top