This calibration fixture standardizes the horizontal and vertical signal paths for 7000 series mainframes. It also allows to measure the bandwidth of a mainframe alone by using the CW IN connector. CW IN signal path supports up to 500 MHz sinewave inputs.
The user manual and more information can be found on tekwiki.
Part of the pulse generator circuit. Differential structure, working a current switches. Note the board cut-outs under the diodes to prevent leakage (?). These diodes are Schottky types, selected and matched at the factory. Does anyone know if there is a modern replacement?
Gold device CR280, the dreaded tunnel diode, GESMTD708 in the pulse generator. Unobtanium part.
Note the cut-outs in the PCB for the transistors.
U30 amplifies the CW input signal. This Tek-made part 155-0078-00 is a 16-lead high-speed variable gain amplifier. Nominal bandwidth 1.05 GHz, max gain 2.82 times (this were the sixties!). Internally it is basically a bipolar differential pair with each collector loaded by a current-steering diffpair. External load resistors convert the output current back to a voltage.
Nice dead-bug assembly. These are intentional, no bodges. Note the 'silk-screen' of the diodes on the circuit board.
These four diodes constitute a balanced peak detector circuit that detect the amplitude of the amplified (by U30) CW signal. The output is converted to a single-ended voltage (circuit around U90, a 741 opamp; also compensates the loop) and applied to the error amplifier U94 (another 741). The feedback circuit steers the gain of U30 such that the output amplitude remains constant (between 0.3 V and 0.7). I cannot help but wonder how much the designers knew about Bode's feedback theory. Probably quite a lot, but this had to be coupled with a huge amount of experimentation.