[DavGoodlin]

Usually when I'm stumped on a tube set, something of a high-impedance nature eventually becomes obvious. But this case its phantom feedback even though the tuner works or is disconnected.
It sounds almost right until the treble is increased past 50%, then it howls with regeneration and volume, bass and bass boost switch have no effect on it.

I fully-recapped this amp intending perfection, even the wax-dipped zenith-oem maroon ceramic disc caps looked suspect and had cracks in them. I did them all using exact replacements that are brown or black dips from JustRadios. Only the grid-leak bias resistors 680K were out of tolerance.

I have narrowed it down to feedback from the speaker winding via a series RC 39Kohm x.01mf to provide negative feedback to the 12AX7 driver's cathode. The bass-boost switch extends de-emphasis further, back to the volume control. I *may* have reversed the feedback loops coming from the speaker windings and will check that next. I did not notice an issue the other times I traced out all my work. Even when I disconnect one of the speakers, it goes in to very loud self-oscillation at about 700 hz.

On many Zenith console stereos with tube amps, single-ended 6BQ5/EL84 beam pentodes are employed when most other 6BQ5s (RCA-Motorola-Fisher-Magnavox) are push-pull pairs. I heard that Zenith used Amperex-sourced tubes for some reason. I noted this difference when I checked cathode voltages on the outputs with weaker RCA tubes and the Zenith originals "made in Holland".

With the tired RCA EL84s I use for initial power ups, cathode voltage was only 5.2 volts, BUT when the Zenith-branded tubes were used, rose to the proper 10.8 volts. It also got a great deal louder with more bass, with the treble at 30% of course.

All this makes me wonder if I did not inadvertently rewire this amp into a multi-vibrator.