J. Kissel, F. Clara

Executive Summary
I've measured the frequency response of all the channels of a D0900900-v2 / D0901284-v4 UK 4CH SatAmp chassis.

This is in prep as a "before" vs. "after" for ECR E2400330 -- but also to back up the modeling I've done suggesting that 7.5% of the OSEM calibration error that Edgard finds is from compensating the frequency response of a D0900900-v2 / D0901284-v4 UK 4CH SatAmp with the (z:p) = (10 : 0.4) Hz digital filter that we've been using since the stone ages (see deep-cut comment from LHO:83662).

A simple model based on drawing component values is correct: the measured response of each channel confirms (z:p) = (0.384 : 10.6) Hz, and that we should be compensating with the inverse of that instead, i.e. a digital filter of (z:p) = (10.6 : 0.384) Hz.

Measurement Setup

Fil had already set aside S1100173, a UK satamp for ECR E2400330 upgrade, so I measured the response of that using an SR785.

Fil introduced me to the satamp tester board, D1000878-v1, which is a nice easy interface for driving an equivalent PD current into the board via fixed 242.0e3 [Ohm] resistor connected to the satamp's PD (K, or Cathode) Input (a la LHO:61202). 

The 
    (V_src_SE / I_in) = 242.0e3 [Ohm] (measured)
series resistor *almost* perfectly cancels out the transimpedance resistor and differential driver; this unit's TIA resistors measure to be 
    (CH1,CH2,CH3,CH4) = (120.1, 120.1, 120.1, 120.0) [Ohm] (measured)
making the overall transimpedance of most of the channels
    V_out_DIFF / I_in = - 240.2e3 [Ohm] (modeled) 
where I include the negative sign accrued from sending in a positive current, rather than a negative current that occurs when more light impinges on the PD, a la the negative reverse bias configuration that's used in the field (see G2500980).

That means this measurement should have a DC gain (and sign) of 
    I_in / V_src_SE * V_out_DIFF / I_in = (240.2e3/242.0e3) = - 0.9926 [V_SE/V_DIFF] (modeled).
    
The only down-side is that the board nominally puts the satamp's differential output voltage on the signal pin and shield of one BNC connector, so I had to do a bit of ugly BNC / clip-lead adapter shenanigans to extract the differential voltage (there's a picture in the collection referenced below for future reference).

Attached are detailed measurement setup notes and pictures.

Results

The measurements confirms that the frequency response of the D0901284-v4 UK SatAmp matches the simple of model of (z:p) = (0.384 : 10.6 , 5.2e3) Hz from LHO:83662 to the tune "model / measurement" is within +/-1.5 [%] / +/-1.0 [deg] at all frequencies measured (up to 1 kHz). See four-page collection of plots attached.