I assume that J3 is for programming U9, which I think is a Silabs 8051 micro (does DC offset adjustment and controls the gain switching relay). The micro also does DC offset cal when you start it up while pressing both buttons (see manual).
The other side of the PCB is very boring - mostly ground plane, a bit of decoupling and a small circuit that doesn't seem to be part of the signal path. I've attached a (cellphone) pic.
Note that the power input seems to be isolated - there is no DC path from any pin of the USB input connector to the rest of the PCB. The output USB connector has one of it's data pins grounded, the other reads high-z (about 2 Meg) - no idea why they have done this!
IC part numbers I've found so far:
http://www.analog.com/media/en/technical-documentation/data-sheets/ADA4817-1_4817-2.pdf (main input buffer)
http://www.ti.com/lit/ds/symlink/opa2171.pdf (low freq op-amp, involved in offset circuit, marking OPMI)
http://www.analog.com/media/en/technical-documentation/data-sheets/AD8001.pdf (output diff-SE converter opamp)
DC-DC part number is AH4V55V5S-1W, made by "SUCCEED"
I haven't gotten any further working out how CMRR is maintained when switching gains - still looks to be done by switching the ouput resistors between the ADA4817 and the AD8001 from 100 to 1000 ohms. This continues to not make sense to me as i would have thought that CMRR matching would be a nightmare due to needing to match 2 resistors + relay contact resistance between 2 separate ranges. I'll look into it further to try and make some sense out of it.
As for CMRR measurements, I measured one probe at 1MHz as ~54/52 dB (50/500x) and the other as ~47/50 dB (i.e. not quite meeting spec). At 10MHz I got ~35dB for both ranges on the better of the two probes, but I'm less certain of those numbers.
Edit: forgot to mention that the screws are under the label (ugh) but it thankfully came off without visible damage. There are 6 (!!) of them going into brass threaded inserts, and the case is nice and solid with a decent chance of passing a Dave-test. Plastic fingers from both case halves go into the slot between the HV inputs - I'd say that the resistor/capacitor network would arc over before anything else in this thing. Ouput bnc/cable is not a strong point of this thing (feels cheap, especially compared to the nice case) but i guess it does the job and is easily replaced (at which point the 75R output could be made into 50R).