Where Does the Signal Go?

[clyde]

EF86 into long tail pair, 2 EL84's.  I've done this many times as 2nd channel in VOX builds.  This was started awhile back, low distorted output, left and revisited.  I subbed in different OPT, problem apparently solved, beautiful tones as expected, full volume on tap.  I removed the old OPT, mounted the new one and back to original symptoms.  Putting 1V into the jack with signal generator, 1V at the grid, which goes down to 50mv (grid) when B+ comes up, both plate and screen measure 30V!  WTH?  290V B+ to 220k plate R, 1M to screen, 2.2k cathode R, app. 85V plate, 100V screen, 2.3v cathode when operating normally. All other voltages in the amp stay stable when the EF86 acts up.  Plugging in the signal generator without turning it on causes the screen and plate voltages to approach B+ values.  ????   

[pdf64]

Have you verified it isn’t oscillating?

[clyde]

Non-oscillation verified.  It was the output jack partially shorted due to corrosion.  So much for using recycled parts w/o verifying their state.  The effect on the input stage boggles my mind, totally insane.  Tube amps are very weird creatures. 

[Lectroid]

@pdf64,

How can you verify oscillation?  I have a very ringy, buzzy circuit I suspect is oscillating somewhere, but I haven't a clue as to how to track it down?

[pdf64]

Ideally scope the output across a dummy load, preferably a reactive one.
If no scope available, consider that an amp that’s in free running oscillation will have somewhat differing DC conditions to when it’s idling.
So measure the HT and cathode VDC in the suspect mode, and then at idle. Much of a difference between the 2 sets of readings may be attributable to oscillation.
To ensure idle, it may be necessary to mute the signal path at several points.

The problem is that sometimes oscillation won’t be free running, the amp might be stable at idle, it can require an existing signal to parasitise, ie parasitic oscillation.
A scope is kinda essential to identify that.

[Lectroid]

Okay, so connect amp output to dummy load (eg: 8 ohm 25W resistor) and clip scope leads to the two speaker leads.  Right?

What frequencies do you inspect?

[pdf64]

From the symptom description, I suggest 1k-100kHz