The screws are indeed under the front and back labels
The board is attached to an aluminum back plane heat sink and slides out:
The amplifier is relatively complicated, with a discrete component output stage.
The front end has switchable gain (x1 or x10 using input relay), adjustable offset (+/-12 volt, 12 bit resolution, generated by PWM from the microprocessor) and signal inversion option. All of this is controlled over USB. While the manual mentions a PC control program, the drivers were apparently only available for Windows XP. Drivers were not included with a recent new amplifier and Rigol tech support did not send me a copy when requested. They said the amplifier is only designed to be controlled with DG1000, DG4000 and DG5000 function generators. That does work (with DG4162) and settings can be stored for offline use.
The output stage is configured as a discrete component inverting op-amp with a gain of 5 (1 kOhm and 5 kOhm resistors). To realize gain of 1, the signal is first attenuated by a factor of 10, then amplified by 2 with another op-amp (LM318). The switching power supplies to generate +/-16.5V rails seem reasonably good with spikes of about 20 mVpp.
The amplifier generally meets specs, able to generate signals up to 1 MHz without much attenuation. Full 1 A current operation is limited to about 500 kHz (using 4 Ohm load I tested). The resistor trimmer adjusts the quiescent current, which is about 0.5A from 12V supply. Adjusting this trimmer affects the ringing behavior at high currents. The output voltage offset of the amplifier is about 30 mV.
I was primarily interested in the output noise of the amplifier. A simple modification is to bypass all of the front end and use the discrete component op-amp directly. By cutting one trace and adding a 4 kOhm resistor to the input BNC one can get an inverting gain 1 amplifier. This doesn't affect the output response, but the noise is substantially lower, particularly at low frequency.
Edit: Added noise spectrum using an external linear +/- 16V power supply. The noise spectrum is even better. So I am tempted to cut-out just the power op-amp part of the board and throw the rest away.
Another similar amplifier is Siglent SPA1010, I am curious if anyone has opened one of those. There is also JUNTEK DPA-2698, discussed here
https://www.eevblog.com/forum/testgear/juntek-dpa-2698-10mhz-signal-amplifier-digilent-analog-discovery/ with a questionable SMPS. There is also FPA301-20W,
https://www.ebay.com/itm/Function-Generator-Power-Amplifier-Arbitrary-Waveform-Signal-Amp-FPA301-20W/114018996426 it would be interesting to see the insides. The last two amplifiers are rated up to 10 MHz.
PS. Is something wrong with inline image inclusion?