J. Kissel, A. Pele, S. Ballmer, K. Izumi, S. Dwyer

We're slowly but surely chugging through implementation of a well-understood ALS DIFF control scheme. On Monday, I designed a control system that used the real ALS DIFF noise, the increased UIM driver strength, and 2 QUADs worth of actuation (see first attachment for design plots). Over the past few days, Arnaud, Sheila, Stefan, and Kiwamu have been measuring all the necessary transfer functions to confirm my model of the plant. I've compiled the results and attached the comparison of what we have thus far (see second attachment for comparison of model against measurement). There is excellent* agreement between model and measurement. We have installed the complementary distribution filters, and we're now fitting the measurement results to obtain plant inversion filters (as mine were just low-Q guesses, because I knew optical lever damping, different top-mass local damping, etc. would cause subtly different frequency dependence).

Of note: 
- *There is still a factor-of-4 discrepancy between the model and measurement at the test mass. Kiwamu and I have modeled the ESD in two different conceptual ways, and come up with the same factor of discrepancy. For now, we'll just accept the measurement as canon and move on, but it would be really nice to understand this mismatch.
- Unlike what I've modeled, Stefan has convinced me to go forward with inverting the frequency dependent plant to look like a 1 [Hz], Q=1 single pendulum transfer function, with a DC actuation strength of 1e-12 [m / ct]. 

The DC gains between the stages are as follows:
TOP: 4.769e-10  [m / ct_{TOP}]
UIM: 5.328e-11  [m / ct_{UIM}]
TST: 1.760e-13  [m / ct_{TST}]

For gain ratios of:
TOP / TST:  3.69e-4 [ct_{TOP} / ct_{TST}]
UIM / TST:  3.30e-3 [ct_{UIM} / ct_{TST}]