J. Kissel

Continuing my low-level, background, fact building exercises on satellite amplifiers (the only previous installment thus far being LHO:85348) here I compare the three types of satellite amplifiers
    (a) D0900900 / D0901284 :: UK 4CH SatAmp 
    (b) D1002818 / D080276 :: US 8CH SatAmp 
    (c) D1900089 / D1900217 :: US 4CH SatAmp
again, but instead now comparing their transimpedance sign and frequency response. 

Namely, if the systems design intent was
    - the US 8CH SatAmp has been wired up to the OSEM in a POSITIVE reverse bias configuration, with ANODE (A) connected to the NEGATIVE terminal of the transimpedance opamp, and 
    - the US and UK 4CH SatAmps have been wired up to the OSEM in a NEGATIVE reverse bias configuration, with CATHODE (K) connected to the NEGATIVE terminal of the transimpedance opamp,
then these better have differently signed overall transimpendance in order to achieve the same sign at the ADC.

As such, I performed a similar frequency response measurement on an example instantiation of a D1002818 and D1900089 chasses as I've been doing for the UK 4 channel satamps -- see T080062 for the full procedure, and the "only" difference is the physical connection of the SR785 to the device under test. To describe that, I attach here a collection of diagrams of the physical setup for each chassis.

The results are clear: driving a POSITIVE voltage through a series resistor into the NEGATIVE input of the transimpedance op amp (i.e. mimicking the flow of PD current [conventional current, not electron flow] into the NEGATIVE input) results in a
    (a) NEGATIVE differential output voltage for a D0900900 UK 4CH chassis,
    (b) POSITIVE differential output voltage for a D1002818 US 8CH chassis,
    (c) NEGATIVE differential output voltage for a D1900089 US 4CH chassis.

Note, the NEGATIVE input to each chassis transimpedance amp is connected to across pages of the circuit diagram to connector pins labeled as 
    (a) K (CATHODE) for a D0900900 UK 4CH chassis,
    (b) A (ANODE) for a D1002818 US 8CH chassis,
    (c) K (CATHODE) for a D1900089 US 4CH chassis.
implying the correct outer-layer usage.

The second attachment shows the frequency response of the three chassis. Note all of the tested instantiations are pre-ECR E2400330 whitening filter change, so they all have the old z:p = ~0.4:10 Hz frequency response. The legend in the caption of the upper right panel shows the predicted z:p values from the schematic, and the data in the upper and lower right panels show that particular z:p divided out of the data to show how close these instantiations are to the ideal. Yes, the 8CH satamp frequency response is slightly different than the two 4CH responses; this will be rectified with ECR E2400330. The deviation of all three responses from ideal is consistent with the diversity of the instantiations of the UK 4CH sat amps, namely the uncertainty in whitening filter stage capacitance value (which should be a total of 20 [uF], but can vary up to 5%; see LHO:85396).

As such, when using these chassis, it is critical that the chassis-to-flange and in-vacuum wiring is connected in such a way that connects this NEGATIVE input to the right terminal of the OSEM PD.

For detailed breakdown of each stage of the sat amp frequency response and sign, to understand *why* and *how* they're different, check out G2500980.