So with the proliferation of fully integrated, micro AHRS systems popping up all over the place, and not limited to aviation, it got me thinking about just how reliable these systems are and what are their short-comings, if any ?
They don’t like rapid changes. They are generally not good for aerobatics etc.
We have done a nice project designing a flight data recorder for some specific GA tasks. It was very difficult to find a MEMS gyroscope which did work well in this application. We used different sensors, from some major manufacturers, including their support, of the 10 or so different sensors only one was capable enough for the task. Others where just unusefull at rapid movement (spin for example).
I think they are more usefull than mechnical instruments though, which don’t like these movements either.
The basic MEMS sensors should be in the range of 10^6 to 10^7 h MTBF
The problem mostly seems to be poor algorithms and non-graceful failure modes in aviation, when eg. airspeed and/or GPS fails (Aspen, G1000), or gyro toppling on rolling takeoffs (honeywell, apparently)
Furthermore, certified devices now on the market mostly have been designed 10 or more years ago, while MEMS sensors have made pretty huge advancements since then.
I am really pleased with my Sandel SG102 – mentioned at the end of this report.
It is plug-compatible with the old King KG102A so usable standalone, in nearly all KG102A applications.
Sandel stuff is not “fashionable” nowadays in the light GA segment. They sell a lot to helicopters and the military and the products are built appropriately. Really good stuff.
I don’t think the SG102 is affected by spins, or anything anybody could do in a plane and still have the wings attached.
Peter wrote:
Really good stuff.
How can that be given that their box has a Sub D connector?
Are you serious? 
I think reliability is read differently by all who replied. So the question should be more specific.
The issues I meant can be clearly seen in the other topic on the Pantera spin.
I was actually referring to performance in the OP, but overall reliability is important too.
The Sandel is HSI / directional only, so seems that’s a fairly easy task versus AI and T&B ?
The SG102 does pitch and roll also but it is not a “certified” source. This has been much discussed here here here here here
To get a certified source you need to do some extra stuff, but the SG102 does perfectly work as an AI (with the SN3500 EHSI) and it doesn’t have the serious failure modes discussed in the above threads (e.g. loss of static or pitot = total loss of the attitude instrument) so I wonder what exactly is missing…
Peter wrote:
the SG102 does perfectly work as an AI (with the SN3500 EHSI) and it doesn’t have the serious failure modes discussed in the above threads
So how does its erection mechanism work?
