Q-factor of spring on a Rotary Actuator

[logbook]

I have just designed and built a rotary actuator which uses flexures rather than bearings. These things are essentially springs which have the ability to support an armature and allow several degrees of rotary action against a light spring force.

I have just done an open loop test on the device and the classic second order response is highly undamped. I have no explicit damping. There seems to be no way to calculate the resulting Q because the flexure has no energy loss data.

I have tested a similar design of actuator and their open-loop response is much less peaky. To be very clear I have found a random plot of a second order system to help explain.

The peak value shown is 20dB (zeta= 0.05). In linear terms the similar actuator has a peak at x2 the LF value. Mine has a peak at x110 the LF value! I know the manufacturer of the flexures on the similar actuator and they are made by a different (rival) company.

The peaks of both actuators occur at similar frequencies, in the region of 2Hz to 3Hz, where the windage is evidently pretty low. My feeling is that the reason my actuator has such a high Q is simply that the flexures are extremely low loss. I am thinking that low loss on a spring probably means a better component in terms of lifetime/reliability. Any thoughts/experience on this?

Having said that, this high Q is a real liability in terms of measurement and it worries me that wrapping the closed-loop control around it is going to be more problematic with the higher Q.

I don’t normally do electro-mechanical design. so this stuff is outside my comfort zone.

 

[israelkk]

Without a detail design of your flexures vs the other flexures design I.e. Exact dimensions, material, heat treatment, design requirements and assembly there is no way to compare them. Referring to your last sentence I suggest you ask for real expert assistance.