Rather than beating the NF1811 dead horse any more I've switched it out for the new KA 26 MHz detector.
I found that after removing the focusing lens that was previously the last element before the BN detector, the beam size was roughly 300 µm (radius) about where the PD needed to go. Here the photodiode size is 2 mm diameter. At this size the beam fits in about 1/3 of the diameter of the new detector area, so this looks like a perfect fit. As a bonus we have one less optic in this critical part of the optical path.
The new detector was mounted at 3" on two 1" diameter posts. There are bolt holes at 1" and 2" spacing on the base of the detector but I didn't have a more solid base in stock. This mounting should be good enough for now, but can be improved on.
The the incident beam was angled at 30 deg and when centered gave a DC output voltage of about 115 mV from a DC power of 170 µW. This is about right for the DC path where from Koji's schematic (see PSL:2162) the DC path has transresistance of 10 Ω followed by gain of 101 (total G=1010): we should expect 129 mV, the small discrepancy could be my sloppy power measurement. This power is down from the usual value of about 800 µW mostly because I refloated the table and didn't realign the refcav input beams. Hopefully I'll get to fixing that issue tomorrow.
The activity on the table unlocked the cavities a few times which caused the beat note PID to kick the frequency around quite a bit. I was unable to relock the PLL tonight but the peak power as the beat note slewed across 26 MHz was about 55.2 mVrms (-12.2 dBm). We should expect that with about 85 µW in from each path, 100% overlap, 0.75 A/W responsivity, and 1.27 kΩ gain that the beat note would be on order -2.8dBm (0.162 mVrms) in this case. I'd say there is some optimizing to do with alignment. Could have also missed the exact point where the BN crossed over 26 MHz.
Its going to take all night for the BN frequency PID to settle (bring on the intelligent NL controls). This will have to wait till tomorrow.
Not sure what the largest permissible beat note power can be here. Koji went with a MAX4107 which has a slew rate of 500 V/µs. There were comments from in previous posts about slew rate (see PSL:2161). What is a good rule of thumb for slew rate of signal vs the rated response of an op amp in situations like this?
Edit Mon Apr 23 13:47:34 2018 (awade): added detector diameter.
This beam is WAY too big for the PD. If the beam radius (wz) is 100 microns and the PD active area diameter is 300 microns, than you're always scattering a lot of beam off of the metal of the can. For new focus 1811, the beam radius should be ~30-50 microns.