J. Kissel, D. Bhattacharjee

Chassis Drawing: D2100631
Circuit Drawing: D2100630
Chassis Serial Number: S2101608
Original Upgrade ECR to Cover the whole electronics update of the GW DCPDs: E2100138
Measured with implemented ECR implemented on the whitening filter not covered in Traveler: E2200045
Design Doc: E2100082
Design Critique based on experiences from this measurement: E2200080

Apologies for the way-overdue aLOG!

Dripta and I measured the new OMC DCPD's whitening chassis chain on Feb 15 2022. I attach several hunks of stuff debriefing the measurement setup here, and Dripta will attach fit results in a comment below. Recall that we expect the design of the chassis to have a DC gain of 1 [V_diff / V_diff], with an optional gain of 2x controllable via switch (hence "hi" or "lo" gain, or as is its written on the front of the chassis "run" or "acquire," respectively), with a frequency response of 2 zeros at ~20 Hz, 2 poles at ~100 Hz, and 2 poles at ~10 kHz, and a single extra pole from the differential receiver, also at 10 kHz, such that in the gravitational wave band between ~200 Hz and several kHz, the circuit provides a a whitening gain of either 25 or 50.

First attachment: 20220311_meas20220215_H1OMC_WhiteningChassisTFs_OMCA_DCPDA.pdf
    (A) Page 1: Here's the final answer first, showing the excellent quality measurements of the (OMCA) DCPDA and DCPDB channels. Note for later, these are driven with a SRC 785 source voltage of 0.5 V_pk single-ended. 

    The first interesting points appear here: 
        (i) The balance of the two channels is quite excellent, with the ratio of the A channel to B channel deviating by at most 2.4% at ~50 Hz, where the mis-match is likely from the component differences of, C20, R90, R41, and C21 R91 R42 which determine the ~100 Hz pole frequencies. At DC and high-frequency the match is already around 0.1%. These statements are true for the high and lo gain settings.
        (ii) Of subtle note that only the calibration group cares about, but each channel's ratio of hi to lo gain is not exactly 2.0, nor is it perfectly frequency-independent.

    (B) Page 2: Here, because we've seen whitening stages behave non-linearly in the past (see 47254 and resulting ECR E1900064), I took quick measurements from 10 Hz to 102.4 kHz with different excitation voltages. With the SR785 source voltage set to 0.05, 0.1, and 0.2, V_pk single-ended, which meant the chassis gets 0.05, 0.1, and 0.2 V_pk differential (or each leg sees 0.025, 0.05, or 0.1 V_pk, single-ended), and the output to the AA chassis at peak amplitude around 500 Hz in hi-gain mode would be 2.5, 5.0, or 10 V_pk differential, which is 1/8, 1/4, or 1/2 the range of the ADC which can accept 40 V_pkpk differential, or +/- 20 V_pk differential.
        (iii) This measurement does show there is some small level of statistically significant non-linearity in the system under reasonably expected signal amplitudes, at the level of 0.25% peaking at 50 Hz. BAD BAD BAD Again, we're talking calibration-group level of caring, but it just gives me the willies thinking that we have such unexpected and unwanted behavior in the system. This comparison was only done with the A channels in hi-gain mode, because I had already spent hours getting the data from (A), so I can't conclude a pattern, i.e. if this was only and issue with one setting of one channel, or if it was systemic to the design, but we should make sure it's not when we redesign the chassis.

    (C) Page 3: This page is just another warning to anyone and future-us setting up an SR785 to take long > 5 minute measurements, and expecting exquisite quality of data that will no confuse a fitting algorithm. As one can see in this plot, I forgot to turn off the Auto-Offset feature within the channel input configuration menu during one of these quick measurements of the hi gain channel A. The sweep direction was up in frequency, so it wasn't until the 5 minutes into the measurement, when frequency vector was up at 12.69 kHz, but *bang* I got hit by the mistake, and you can clearly see the measurement goes bogus above this frequency.

Second Attachment: D1900027-v3_aPLUSO4_D2100630-v2_OMCDCPDWhiteningChassisSetup_WhiteningMeasurement.pdf
    - The usual kissel-o-gram of how to connect clip leads to chassis.

Third attachment: 2022-02-15_OMCWhiteningChassis_MeasSetup.pdf
    - Pictures of the front and back of the OMC whitening chassis and connected in measurement setup during the characterization of the whitening filter within the chassis.

Fourth attachment: 2022-02-15_OMCDCPD_WhiteningChassis_MeasurementNotes.txt
    - Notes detailing the exact SR785 configuration setup during the (final, successful) measurements.

The stored TF data for the two channels worth of two gain settings (as well as the TF of the measurement characterization system) live in 
    ${CalSVN}/trunk/Common/Electronics/H1/Data/OMCWhiteningChassis/S2101608/2022-02-15/
        2022-02-15_OMCDCPDWhitening_OMCA_DCPDA_HiGain_mag.TXT
        2022-02-15_OMCDCPDWhitening_OMCA_DCPDA_HiGain_pha.TXT
        2022-02-15_OMCDCPDWhitening_OMCA_DCPDA_LoGain_mag.TXT
        2022-02-15_OMCDCPDWhitening_OMCA_DCPDA_LoGain_pha.TXT
        2022-02-15_OMCDCPDWhitening_OMCA_DCPDB_HiGain_mag.TXT
        2022-02-15_OMCDCPDWhitening_OMCA_DCPDB_HiGain_pha.TXT
        2022-02-15_OMCDCPDWhitening_OMCA_DCPDB_LoGain_mag.TXT
        2022-02-15_OMCDCPDWhitening_OMCA_DCPDB_LoGain_pha.TXT
        2022-02-15_OMCDCPDWhitening_MeasSetup_mag.TXT
        2022-02-15_OMCDCPDWhitening_MeasSetup_pha.TXT
    

The script to quickly plot the data in this aLOG also lives in that same directory, 
    ${CalSVN}/trunk/Common/Electronics/H1/Data/OMCWhiteningChassis/S2101608/2022-02-15/
        quick_plot_data_20220215.m

Again, stay tuned for Dripta's fits of these measurements in the comments below.