[GreenKeys] Need help with 33 parity keyboard (and rant)

[Charles]

I have reached the limits of my patience, not to mention capabilities 
(unless I grow a third hand) trying to fix my 33ASR's keyboard.
It was working perfectly until the worn motor drive belt popped off the 
pulley while I was typing earlier this week.

Once I put the belt back on temporarily, the keyboard now always has Bit 5 
and Bit 7 set!
It's a parity keyboard so the status of bit 8 will change to make an even 
number of bits. But, for example, Ctrl-Shift-@ should transmit all 0's, and 
it has those two bits set also.
Typing a CR echoes as "M" (not Ctrl-M). Always the same two bits.

The next time I plugged in the machine, the selector magnet was making a 
loud buzzing sound. I tracked it down to an open filter capacitor in the 
power supply. Replaced it with another from my junk box - which also turned 
out to be bad! Got that fixed, back in local loop mode with the wrong 
characters still coming out.

Today I took the keyboard apart (which I've done before) and couldn't 
identify anything jammed, disconnected or otherwise injured. I was looking 
for a disconnected t-lever, bent codebar, etc. All the contact wires look 
fine (and would not have changed just because the motor stopped during 
typing).
Putting it back together was the usual three-handed struggle.

Then I compounded my problem by attempting a shortcut changing the drive 
belt. I accidentally "spread" the distributor clutch apart, and to get THAT 
back together I had to take the distributor shaft out anyway, which is what 
I'd been trying to avoid in the first place! Put it back in- and realized 
I'd forgotten to put the new belt over the gear! Did it all over again. Next 
I found that the clutch would never disengage - after RTFM, I found the 
little spring lying under the typing unit that is supposed to keep the trip 
lever in position. Sigh.

Now that it's all back together, the typing unit works when sending it 
characters from my PDP-8/A. But my keyboard now won't toggle the H-plate 
enough to trip the lever, regardless of the H-plate position setting (too 
loose makes it run all the time, too tight won't trip).

I can only assume that I not only didn't find the keyboard problem but 
multiplied it by reassembling something incorrectly.
Can someone please fix it for me... or just sell me a good (non-parity is 
fine) 33 keyboard? Once I have a known good keyboard I can make the other 
adjustments to get it to work in my 33ASR.
I wasted a whole day and I'm about to pitch it out the window!

thanks
Charles