JohnC Veteran Location: East Yorkshire, England
| I'm the guy who's doing the CARVEC system, which I hope will achieve many of the things you are talking about. Whilst developing it, I have come across a lot of problems regarding the things you mention and, like everything else, it's never as straightforward as you think:
The mercury switches are really just gravity-sensors and you would use a solid-state accelerometer to obtain a variable signal proportional to the amount of tilt. However, even when you've done this it would not work!. The problem is that when the heli tilts away from upright, it starts to accelerate sideways. This sideways acceleration cancels out the extra acceleration sensed from the gravity component you are now detecting!. So, as far as your on-board system is concerned, you are at the same angle.
Secondly, two heading hold gyros on the pitch/roll axis would not work either!. The problem here is that they are not sensing pitch and roll, they are sensing helicopter X and Y axis movement. Taken to the extreme to illustrate the point, imagine a heli straight and level: now pitch it up 90 degrees so it is looking upwards. Now rotate it 90 degrees about the Z axis (tail rotor movement). The heli is now at a pitch angle of zero degrees and a roll angle of 90 degrees - whereas the two X and Y heading hold gyros have not turned at all.
A further problem with standard gyros is that the heli vibrates a lot more in the X and Y axis than the Z axis, because they are directly affected by vibrations in the main rotor. The Z axis is at 90 degrees to the plane of vibration and so give a relatively clean signal. I think a normal gyro would cause a lot of servo-movement and possibly shake itself to death if you tried - but I may be wrong as I haven't. I would be interested to hear the results of anyone who has successfully done it ??.
The optical sensor systems like the FMA co-pilot are ingeniously simple - and appear to do an excellent job considering the compromises they have to make. However, they can't realistically form the sensing elements of a proper autopilot - they have too many constraints.
I hope you find this interesting. This is just the tip of the iceberg ...
John C. |