S. Dwyer, K. Kawabe, J. Kissel
As we continue to gather evidence for the PUM driver actuator noise impacting DARM (see LHO aLOGs 53329 and 53253), we need an estimate of how much closer we would be to regular DAC saturation if we engaged the analog low-pass filter, transitioning from "State 1" to "State 3" -- and thus requiring a digital compensation filter that would amplify the existing request by a filter with two zeros at 0.5 and 250 Hz and two poles at 6 and 20 Hz (i.e. the ratio between state 1 and state 3).
Attached is plot of the results for ETMX.
I only plot the estimate for the H1 ETMX PUM UL coil, because
- I've confirmed via measurement that UL is exemplary of all coils on the ETMX PUM.
- We assume that ETMX PUM is a "worst case," since it is the only QUAD which receives both LSC and ASC drive at the PUM stage.
The conclusion: by using State 3, we decrease our nominal low noise safety margin by 1 - (12.46 / 9.207) = 35%, but we're still a factor of 9 away.
Note on the current conditions: useism between 90th and 50th percentile region at around 0.5 um/s, in low-ish winds at 10 mph.
It is our plan to indeed code up the engagement of State 3 (ACQ OFF, LP ON) in the LOWNOISE_COIL_DRIVERS state tomorrow during maintenance, such that all subsequent observation ready segments will have state 3 engaged (until otherwise noted).
One hopes that we would switch to State 3 *after* most of the lock acquisition sequence is complete, so the ASD shown here (taken during an observation ready segment) should be representative of the environment and requested drive that we would typically have.
Driver Type Circuit Iso. Stages State Switch State Freq. Resp DC Transconductance [mA / V]
ACQ | LP (z):(p) [Hz] [mA/V]
PUM D070483-v5 QUADL2 1 OFF | OFF (12):(110) 0.27
2 ON | OFF (1.35):(80.5) |
3 OFF | ON (6 12 20):(0.5 110 250) |
4 ON | ON (1.35 6 20):(0.5 80.5 250) V
You can find a plot of the magnitude difference between the states in the plot here, pulled from LLO aLOG 4495
The template for gathering the data lives here:
/ligo/svncommon/SusSVN/sus/trunk/QUAD/H1/ETMX/SAGL2/Data/2019-11-25_H1SUSETMX_Actuator_DAC_Output.xml
The script to process the exported data for the PUM lives here:
/ligo/svncommon/SusSVN/sus/trunk/QUAD/H1/ETMX/SAGL2/Scripts/estimate_coildriverstate_impact_PUM_20191125.m