Rana and I went out to look at the breakout board for setting the gains/whitening for PR3. We found three out of the four segments were set to have one stage of whitening; the fourth had no whitening. Regardless, all four segments had an antiwhitening filter engaged in their filter banks.
We flipped bits on the breakout board as follows: on each segment, bits 2, 3, and 5 are high; all others are low. This gives one stage of whitening, and somewhere between 5000–10000 dc counts on each segment.