Alain,

Yes the Twinsync will compensate for non-linear servo movement. All servos have some amount of dead band. If the deadband is greater than .01ms (10us) then you will loose some accuracy. Most expensive coreless digital servos have about a 1-2us deadband. Most of times the RF link between the TX and RX introduces 2-10+ us of jitter frame to frame. So you want a servo with a deadband of 10us or less. The twinsync will correct for servo positioning error but not servos that jitter due to a worn or dirty pot or overshoot and correct.

Also if you do the test you talk about, do it without the TwinSync installed. The twinsync makes the servo output a little notchy due to a high deadband and filtering on the throttle input. It has to do this so frame-frame jitter from the receiver is not sensed as a throttle stick change. A throttle stick change disengages the syncronizer, moves the servos, then rengages the sync unit.

However, the resoloution of the TwinSync is better than 1% driving the servos but you can not see it until both engines are running and it is controlling the servos. Try running it with the best servos you have and the worst and report on the difference. I am willing to bet what you will find is you will have to slow down the timing with fast servos and speed it up for really slow ones to get the best time to sync. In the end I think you will find the entire system performs the same because throttle response time and your endpoints (higher is better) are the limiting factors.

(I have about 250 assorted servos. I did bench testing during development with everything from Futaba 9452s (0.07sec/60 deg) to some really old 0.4xsec/60deg Kraft servos). The 9452s are amazing servos. They are so fast you can visually see stick bounce if you turn loose of the spring loaded transmitter stick.

OK... I've given out enough trade secrets...