Menu Sign In Contact FAQ
Banner
Welcome to our forums

Just how reliable are MEMS based AHRS ?

Peter wrote:

See for example here

That discussion doesn’t answer my question. Basically the postings said that you need GPS, airspeed etc. to determine when the aircraft is in level flight so that the gyros can be aligned. Fair enough, but they didn’t say why this can’t be done in the same way as with a mechanical gyro.

ESKC (Uppsala/Sundbro), Sweden

alioth wrote:

That surprises me. These MEMS sensors are used in flybarless systems for RC helicopters. They cope with some pretty extreme aerobatics from the 3D guys.

We also tried these, without succes. Tried several portable solutions and tried several certified products, none of them was useable, apart from a 10K commerical products, which works fine.

tomjnx wrote:

Huh? Why then do those certified backup attitude indicators not need GPS or airspeed input?

It is clear that these aerobatics manouvers are not part of the certification process. I can see why, a mechical gyro doesn’t like aerobatics either. It is certified to be replacement of such gyroscopes, so why should it’s performance exceed that of a mechanical one.

Airborne_Again wrote:

So what are the position, heading, airspeed etc. used for?

This is used for cross checking. When this data does not agree one of the data sources is invalid.

JP-Avionics
EHMZ

Jesse wrote:

This is used for cross checking. When this data does not agree one of the data sources is invalid.

That doesn’t make the algorithm any less stupid. It’s relatively easy to loose airspeed (eg. by icing up the pitot), or to loose GPS (eg. by RF interference). It is however rather difficult to interfere with basic physics, and to my knowledge nobody managed to modulate the gravity field with anything less than planet size. So by “gating” your high reliability angular rotation and linear acceleration based attitude with the comparatively lower reliability airspeed or GPS velocity, you’ve successfully reduced the reliability of the overall system. It’s IMO rather stupid to draw X over the screen in IMC even tough you’d have a most likely perfectly valid attitude, you’re just not fully sure about it.

LSZK, Switzerland

I agree.

and to my knowledge nobody managed to modulate the gravity field with anything less than planet size

I reckon that (interacting with gravity without manipulating huge masses) will be the biggest scientific discovery in the next 100 or 1000 years – IF they ever make it.

Administrator
Shoreham EGKA, United Kingdom

tomjnx wrote:

It is however rather difficult to interfere with basic physics, and to my knowledge nobody managed to modulate the gravity field with anything less than planet size. So by “gating” your high reliability angular rotation and linear acceleration based attitude with the comparatively lower reliability airspeed or GPS velocity, you’ve successfully reduced the reliability of the overall system. It’s IMO rather stupid to draw X over the screen in IMC even tough you’d have a most likely perfectly valid attitude, you’re just not fully sure about it.

I completely agree with that. However it won’t give you a red X directly. It will continue to indicate, stating that you must cross check it’s indication and (with Aspen) to check pitot heat. When the situation gets worse, it will in the end show a red X.

JP-Avionics
EHMZ

My understanding is that (lower end) MEMS drift more than acceptable for an AI. So being able to detect wing level by other sensor is used to remove the drift.

Nympsfield, United Kingdom

As aside, I keep reading more and more on forums about Aspens and G500/600’s that fail in flight. Many of them with weird intermittent failures and red crosses all over in IMC. Some rectify themselves with a restart, some don’t. But in any case – if its failed once in IMC, do you feel comfortable it will “repair itself” should it happen again? As the technology gets older and older (G600 is well over 10 years now), expect this to happen more and more. Just remember how those old 80’s cars with the then fancy new LCD or digital dashboards look or work today – that’s how these systems will be in 20 years time. Try to get an old Casio watch repaired today, or a Camaro z28 speeder, or an Aston Marion Lagonda panel. Unreliable and unsupported.

Complete reliance on glass, which seems to be every panel remodelers goal these days, might come with a price. I understand the desire to get rid of old vacuum systems, but there’s something to be said for complete redundancy, not just another unsupported electric unit redundancy.

An equally unsupported cessna autopilot recently tried to roll me over in IMC. I had a vac pump failure in flight (vmc, fortunately) and a failed attitude indicator which toppled at night over the sea.

The worst hardware failure i had in glass was the brightness button coming off… But of course they are much newer.

I agree with the need for proper redundancy, I would want a unit that is completely isolated from the main avionics (I had more failures of the G1000 because the service centre botched a software update, one killing the entire system, one breaking the Engine instrumentation after a few minutes), but electrical with integrated backup battry is fine; glass or not – don’t really care.

Last Edited by Cobalt at 26 Nov 02:08
Biggin Hill

Jesse wrote:

This is used for cross checking. When this data does not agree one of the data sources is invalid.

So a solid state AI will work perfectly fine without any air or navigation data, but the risk of incorrect indications is greater than with a mechanical AI and that’s why you need the additional data for cross-checking. Have I understood you correctly?

ESKC (Uppsala/Sundbro), Sweden

The r/c helicopter sensors will keep them upright for a few seconds, but really they’re fly-by-wire devices that help damp down instability in roll and pitch, and a coupling between cyclic, collective and yaw. They do drift quite a bit over a minute or two. Also very sensitive to vibration.

Self-levelling devices for radio-control typically use infra-red sensors to look for the horizon.

Sign in to add your message

Back to Top