I have moved the DHARD control to be distributed between AS A and B RF45 Q instead of only on AS A RF45 Q to see if we can reduce the noise in DHARD.

On Monday, I ran an injection in DHARD P and Y (see pitch, see yaw) and determined that both AS A and B RF45 Q have equal SNR for DHARD. The difference is the phase, AS B RF45 is the opposite sign as AS A RF45. Therefore, I figured we could try a new sensing scheme, instead of using just AS A RF45 Q, we would use 0.5*A and -0.5*B signals.

Today, I first tested this by putting in my new input matrix into DC7 pitch and yaw. I watched my new signal compared to the nominal DHARD signal during the engagement of DHARD WFS. There were no offsets in the signal. I ran a quick passive transfer function to determine that the signals have the same magnitude and phase.

Then, Tony was holding us at 2 W to sweep the LVEA before observing, so I used the opportunity to try the new matrix. I put the new matrix in the DHARD B path, aka the DHARD blend filter path. I then ramped halfway between DHARD A, which is just on AS A, and DHARD B which is distributed between AS A and B. This worked well, so I ramped the rest of the way. This worked perfectly for both pitch and yaw. I decided I liked this, so I put the new intrix back in DHARD A and ramped back so I wouldn't have to deal with extra SDF diffs.

I continued to monitor the signals as we powered up and saw that everything was performing well. I updated the guardian to use this new intrix- this matrix is set in the DARM_TO_RF state, and then DHARD is first engaged at DHARD_WFS.

Sensibly, this corresponds to about a sqrt(2) reduction in the DHARD noise above 10 Hz, comparing the signals from today to signals during observing last night.

I accepted the differences in observe SDF as well. Note that I also accepted a CHARD tramp due to some other ASC changes I reported in this alog.