Thx Peter. Comment about floating voltage reference well noted.
I have read some rather detailed posts by someone called flyer101flyer on the pprune forum. My present assessment is:
1. Mechanical TCs have huge issues that will vanish in the solid-state version.
2. The current state of the available technology is such that building a prototype will be very easy and inexpensive
3. Disregarding certification issues it seems that the idea makes a lot of sense from a manufacturing perspective.
I believe the best way forward is to
1. procure/install both device types on a turntable, compare output to roughly calibrate the output of the piezo version.
2. Install both devices on a board, connected to a laptop for data acquisition take it flying.
My Extra 400 is very useful for that: one can easily affix a test board on a rear facing seat in front of the “engineer”, a 24 V socket is installed in the cabin and there’s even a spirit level integrated in the cabin door so the straight and level pitch angle can be correctly set.
I wonder if it could be legally installed with a switch so that the mechanical TC remains in place
In general this is not strictly legal because any change in what feeds an autopilot (and the pilot needs to know about) needs an amendment to the autopilot AFMS (the pilot’s handbook). The other thing which gets you is another strict interpretation of the regs: any connections not shown in the IM are illegal.
A lot of people have been up this road before, on stuff like a KI256 replacement (that one is worth way more $$$ than any other GA gyro replacement). To legalise it you would need at least a field approval (if N-reg) and a new AFMS which would need an approval by an ACO. I did that here but these things are very hard to do from Europe.
the vertical axis of the TC is mounted with a 35 angle forward in order to generate a blended signal from yaw and bank inputs.
Yes; that’s how the TC works. It’s fairly crude. This is why I said you might need to do some software, but if you can get a rate of turn directly, that’s even better. BTW I heard the STEC TC output voltage is relative to the reference voltage, not to aircraft ground.
I think the dedicated device Peter and I were discussing is a much simpler solution than any software based approach that does a lot more than rate of turn.
We are really trying to remove a complex and failure prone mechanical device and replace it wth a “plug compatible” alternative, no more.
I was not expecting such an elegant option to arise – these sensors look very very promising.
Also found someone on the pprune forum who apparently has tested solid state TCs (but not in relationship with an autopilot).
I am not yet in touch with him personally but his posts seem to confirm that the solid state solution will be far better than a mechanical gyro.
Apparently much less sensitive to turbulence, more symmetrical behavior etc…
He says that the vertical axis of the TC is mounted with a 35 angle forward in order to generate a blended signal from yaw and bank inputs.
In an ideal world we’d find a supplier willing to produce a 1:1 replacement with adequate resolution and full swing voltage
I wonder if it could be legally installed with a switch so that the mechanical TC remains in place.
The stratux folks are working on a AHRS using an IMU like that.
I know Sparkfun has a 6dof IMU module:
https://www.sparkfun.com/products/10121
More than you need for a TC but they have a whole range of breakout boards for things like this.
Peter I think 3 degrees per second full scale is not enough. More like 30? We’ll need testing.
I have asked them for a quote and also found another type which IS temperature stabilized. It uses a panasonic chip
(It seems that these sensors are very common in the automotive industry so they should be cheap)
Output voltage has a midpoint at +2.5 V and very low current which is a small nuisance (need to implement the offset to achieve a +/- 1 V full scale swing and increase current capability probably.)
Version R050 does 50 degree per second full swing and gives 25 mV per degree/sec, so that would be 1 V for 40 degree per sec. Seems like a good starting point..
Summary of project:
- make/find a test bed (rotating table, 24 V power supply, 2 voltmeters)
- Buy a used TC
- Buy a solid state sensor which will also be used as the rate of turn measurement instrument on the test bed
- Install, test and map the characteristic of the mechanical TC
- Select the most appropriate off the shelf sensor type
- More testing
- Implement the additional electronics to emulate the +/- 1 volt. This will need some careful design as it will become the weak link…
I wonder how much better the AP will work with this…
That looks interesting. I wonder if the specs are good enough, with sufficient stability, for the 3 degrees per second (full scale) which you are looking for?
Also it doesn’t look like it is temperature stabilised so you would need to implement that
“Obviously if you could get an AHRS module which outputs the rate of turn directly, in analog, you are done”.
http://watson-gyro.com/product/rate-gyros/solid-state-angular-rate-gyro-ars/
Or that?
http://www.strainsense.co.uk/products/angular-rate-gyros/12206ac-uniaxial-angular-rate-gyro
I have never designed anything with an AHRS, but if I was, I would buy one of the ready-made AHRS modules. These are temperature compensated and trimmed, so the really crappy drift of a cheap (anything below a fibre optic gyro) AHRS becomes, ahem, almost not very crappy at all 
But I think these give you the pitch and roll signals, not a turn rate. One could mathematically derive a turn rate from the others, I think, but I have no idea how. You may need a roll angle together with a [rate of change of] heading… One guy who was really clever with this stuff was @DavidS and @Lionel may know too.
Obviously if you could get an AHRS module which outputs the rate of turn directly, in analog, you are done. You just need to fake the “tacho” signal.
Peter you are confirming my gut feeling. As to keeping the TC on board well look at my instrument panel.
(insert pic here)
Can you see it? Rrrrright behind the iPad.
I would of course never ever consider the idea of having both the TC and the TC emulator on board with a hidden switch that is solely used for ground testing of the emulator…
Jokes aside – if it is a relatively linear depiction of turn rate (which really is degrees of heading change per second) then maybe a simple OpAmp would do the trick?
Who needs a CPU for this?
