Originally Posted by
alf5071h
JPcont,
Your alternative perspective (#370) is interesting. Whilst the basic information (AoA) is the same for both ‘systems’, each system has alternative views and thus use of the information. MCAS receives a valid, but inaccurate value of AoA, the computation and activation works exactly as designed - except that the output is not what the aircraft situation, nor crew requires.
I try to explain why I feel that the MCAS is based on the rotten foundation rather than badly implemented software:
Control law works against some constraints. The simplest one is (set point - measurement)*coefficient. The control law might include integrator or there might be integrator in the system e.g. hydraulic cylinder, jackscrew etc., basically everything that acts the same manner.
When there is two control loops, there is also two active constraints.
Lets think that we have a reservoir (pitch angle) and the level of it is controlled by two independent double acting pumps (elevator, stabilizer) with distinct integrating control laws (PI controller) (jackscrew with MCAS, pilot) with slightly different level set points. The outcome is that integrator values decreases/increases until one pump is at full power (physical constraint). Then other controls the level accurately but the system acts stupidly.
I know that this is artificial and somewhat false example. However, the point is that this type of distinct control laws can have, not so nice, side effects. If the MCAS kicks only ones, pilot can easily out trim it and it has no effect. It seems to be either potential dangerous or ineffective. So, if I were regulator, I would have couple of words about this invention...