fast piezo driver

[gkasprow]

We get more and more questions about fast piezo drivers that can work with Pounder. Two prototypes were built so far - one as a Stabilizer mezzanine and the second as an EEM 8-channel unit with Zotino-type DACs integrated.
But they were very slow, primarily used for alignment
What are the parameters of such a driver that would be a part of the feedback loop?
We would probably need to go for PA98.
Here is the list of APEX amps.

The idea is to make it as an EEM/DIOT card with SMA in/ SMC out
1 or 2 channels
We have the funding to make prototypes quickly and ship them to beta testers.

[jordens]

Pounder doesn't have any DC DACs.
What's the actual requirement?
Isn't the hybrid approach with a slow, low-IQ, cheap PA97 plus an AC-coupled, fast, ~10V-range, medium-current amp generally favored for the relevant use case?

[gkasprow]

I'm chatting with Michal Hejduk. He gladly agreed to post our conversation here.

Thank you for the interesting suggestion. I will try to explain our envisioned adaptation. We are designing an optical cavity, the mirror of which is situated on a piezo-transducer PZT as indicated in the attached figure. We will need electronics that drive the EOM and use this driving signal as a reference, which is going to be mixed with the signal reflected from the cavity mirror. Up to here, the mechanism is same as in the standard PDH stabilization application (as described in the attached document from Moglabs). However, instead of driving a piezo-actuator or whatever in a laser, we want to move the mirror mounted on the PZT. The reason for that is to stabilize the cavity against vibrations. The laser is stabilized by a separate wavementer well enough.
As described in the attached thesis, the vibrations are going to have hundreds of Hz at most, so typically we would need the PDH error signal sampling rate two orders of magnitude higher to be on the safe side. A constant DC signal must be supplied to the PZT to keep the cavity resonant with the main frequency component of the laser beam, i.e. at the more-or-less constant length that would be just little bit adjusted by the error signal. The typical DC signal values are in the piezo datasheet from Jena, also attached.

It should be possible to detune the PZT programmatically, probably moving from the carrier resonance to a sideband resonance. I still have to think this through.

I would appreciate it very much if you could suggest anything from your portfolio or if we could work together on the solution.

[sbourdeauducq]

Have you explored the vibration-compensated fiber spools you mentioned a few months ago? Setups as depicted tend to be expensive and bulky. If the fiber stuff works then the price is slashed by at least 10x and the entire setup fits in a small suitcase.