Kiwamu, Sheila, Rich, Daniel, Rana
Summary
We are close to saturating REFL_A_RF9 during our hand-off attempts from sqrt(TRX+TRY) to REFL9I.
For the time being, we are switching to REFLAIR_A_RF9 for the handoff. Sheila and Rana have done some rewiring on the ISC rack so that the I&Q demodulated outputs of REFL_A_RF9 are replaced by REFLAIR_A_RF9.
Details
The attachment shows the 9 MHz rms monitors for REFL_A and REFLAIR_A during a lock acquisition attempt. After some discussion, Rich and Daniel have concluded that the RF monitor channels are calibrated into dBm, as measured at the output of each PD with a 50 Ω load. REFLAIR_A is sitting comfortably below −10 dBm. REFL_A, however, peaks as high as +15 dBm. The amplifier in the RFPD is an LMH6624, for which the minimum output swing is reported as ±4.4 V into 100 Ω. Since the amplifier has a 50 Ω series resistor on its output, that's 20 dBm of power delivered into a 50 Ω load. If we want to keep distortion to a minimum, we should stay well below this limit.
we swapped the the I&Q cables at the demod board side, so the REFL_VAC I&Q now go into the AA inputs for REFLAIR and vice versa. This seemed more sensible than doing the RF, since that would introduce weird phase shifts. We then swapped the digital demod phase numbers between the REFLAIR and REFL_A screens so that the demod phase follows the cables.
I'm confused as to why you needed to move cables. Wasn't the REFLAIR signal already going to the CM Summing Module, so you could have used that path? How about a diagram of the setup?
Rightly so. We forgot to mention why we did this...
The LSC model wasn't wired to permit normalization of REFLAIR with sqrt(TR), so we swapped cables rather than rebuild the model and risk that kind of fallout. We should change the model to allow this and swap the cables back eventually.