Evan G, Jeff K

Summary:
We finally processed the data Jeff collected (aLOG 21325) to investigate the mysterious zero near 100 Hz. Unfortunately, while we began to further understand the calibration measurements, we also uncovered new issues needing investigation.

Solved:

1. We can verify the DC transconductance of the UIM BOSEMS is ~590 uA/V, and not off by a factor of 2 (see 21127 and 21142), and is an issue with the fast I MON
2. There is some "zero-like device" (see details below) that is at ~100 Hz

New mysteries:

1. The frequency dependence of the "zero-like device" appears to be f^{3/4}... WHUT?!
2. The state 3 and state 4 zero-pole might not be fit that well but we are not sure if it is real, but it looks smallish
3. The CAL model uses incorrect zero-pole pair in the UIM stage, and not using any of the fitted low-pass filter values

Details:

Taking Jeff's measurements of the driver electronics (plotted magnitude only called rawXX.pdf attached) and dividing by the fitted values for the low-pass filters from aLOG 21283, we can see the remaining effect of the BOSEM in the circuit (see XX.pdf attached). We plot the magnitude and phase of each state. Left hand plots are over the full range of 1 Hz to 10 kHz, while right hand plots are zooms from 1 Hz to 200 Hz.

Fortunately, from these measurements, we can confirm that the DC transconductance is: UL = 578 uA/V, LL = 589 uA/V, UR = 582 uA/V, LR = 600 uA/V. Earlier measurements (see 21127 and 21142) had found that this was off by a factor of 2, but we do find the expected value. However, when trying to understand the "zero-like device" we found that the frequency dependence at higher frequency goes like f^{3/4}! Now we are confused.

Looking closely at the low frequency region, and in the zoom (right hand plots) we see that the fitted zero-pole pair doesn't quite take out the low-pass filters for state 3 and state 4 (last two low-pass stages). This could be because of the buffers in the driver electronics, but we are not sure if this is really the cause.

We also started looking at the parameters file for the CAL model and found that the UIM parameters have the wrong zero-pole pair and, in addition, are not using the fitted low-pass filter values. This needs to be rectified.

The script that plots these results is in:

/ligo/svncommon/CalSVN/aligocalibration/trunk/Runs/ER8/H1/Scripts/Electronics/analyze_uimdriver_wBOSEM_20150908.m