Getting right to it:
Wes' friend agreed to ship the scope out to me. I've got it and another 465 belly up on my bench at the moment. There are a couple/few issues: the sweep logic chip U870 is being flaky (but works when it works), Q1098 is bad, and this Horiz Amp issue.
I think I've identified one of the key issues at hand: the bias of the input to Q1224 is too negative. The signal is between -320mv and -600mv. On my working scope the input to Q1224 is 0v with a sawtooth wave of +/- 40mv.
I've also verified with an LTspice simulation. Thank you Ozan for sharing the 2235 Horiz Amp schematic. It was the basis for a rewrite to the 465 Horiz Amp. My simulation is mimicking what I'm seeing on the bench: a zero offset 80mv input signal yields nice amplification through the first stages of the amp. A large negative offset yields the disfunction seen on the failing scope. This includes no real signal at the emitters of Q1224 and Q1234, no signal at the collector of Q1234 (a small negative DC value), and a reasonable looking signal on the collector of Q1224.
Normal Q1224/Q1234 collector signals are offset +3v with a 2v signal in both simulation and on the bench. Dysfunctional Q1224 collector is +6v offset in both simulation and bench, but the signal level is somewhat better on the bench than the simulation. Not really a factor, just an observation. Better beta in real transistor than .model?
I can't figure out where this input bias is coming from. This is how I discovered Q1098 is bad. I was trying to see if the B-sweep was contributing, but it turns out no, because the emitter base junction is shorting the B-sweep signal to ground. I've removed the transistor from the scope to eliminate it as an issue. I've verified that neither the emitter nor collector sockets are shorted to gnd.
This leaves, near as I can tell, only Q1038 or sheet 9/Horizontal Amplifier as sources for a negative bias. I'm pretty comfortable saying Q1038/A-sweep are not the source as there is no negative voltage coming from there, and the A-sweep signals inside of Q1038 are fine.
I've pulled up legs of the resistors around Q1224/Q1234 to verify their values. Every resistor has checked good. The bias voltages are according to the values shown in the schematic. One result I found interesting is I tested the input signal with R1221 and Q1224 pulled. The sweep input bias goes to +8v and there's still some signal on it. But this tells me there's nothing coming from the X axis input nor position pots to pull the input so negative. I've pulled P1200 from the position pot anyway. To no avail.
So what the heck!? Everything in the input bias looks spot on. I've even replaced Q1224/Q1234 with brand new 2N3904s. I just got a roll of 100 from Mouser.
It's a puzzle. Any ideas?
Dave
Dave Peterson