Well, there I was, "fixing up" a 7633.
There are two 7A18 amplifiers to align.
One channel has an appalling square-wave response: all over the place as you select various attenuator settings.
With the attenuator out (5mV/div) the response is way down at HF.
After a bit of head-scratching it becomes apparent that the signal is wrong leaving the source of the input FET. Changing the FET makes no difference. Neither does removing the input clamp diode.
Now there are only half a dozen passives left*: three R's plus 3 C's. All the R's check out OK. In desperation I change the entire assembly, and the signal came good. The culprit was the 1.8pF capacitor to ground (C212). It measures close to 18pF on a simple digital capacitance bridge, and I did wonder if this component has been incorrectly marked and somehow got through the Tek system undetected. However, further investigation revealed that this "capacitor" really has failed in a most unusual manner: here are the results of a few tests:-
DVM ohms: (overload)
Insulation tester: 10Gohm @ 100V DC
Digital capacitance meter: 22pF
ESI LCR-bridge: 12.5pf // 80n Siemens @1kHz (i.e.12Mohm approx)
Two similar styles of NOS tubular ceramic both measured their nominal cap. value and 0nS, so clearly the Tek part has failed high capacitance and high loss, (either series esr or some kind of dielectric fault).
So, next time you're up against some weird fault, don't discount the humble and lowly-stressed ceramic caps!
* PS will post some photos under 7A18.