Mystery Oscillation Solved

[k4kyv]

I once had a most mysterious problem with a transmitter - it would go into (very stable) self-oscillation on 160m with nothing but the filaments on, in standby mode with the HV power supply completely turned off by cutting off the a.c. voltage to the HV power transformer.  At first I noticed it as an unidentified carrier that kept popping up near my operating  frequency.  It was very stable, so I assumed it to be someone tuning up near-by, but they never ID'ed.  This drove me up a wall for a couple of weeks, as I thought it must be someone local, since the s-meter did not move to indicate any QSB.  One time I just happened to bump one of the tuning knobs on the rf exciter, and the carrier went away.  Upon further investigation, I discovered that rotating the tuning knob on the first buffer stage changed the frequency of the carrier!  It would have made a good VFO, it was so stable.  But how could an rf stage go into oscillation with no DC voltage on the tube?

The offending stage used a 6V6, as a buffer stage to the external VFO.  The power supply used a 5R4 rectifier tube.  I started to measure voltages with my DVM, and discovered that the plate of the 6V6 stage showed about +15 volts.  I could short out the B+ line to that stage, and the oscillation would stop.  I pulled out the 5R4 rectifier, and the oscillation would stop, but shorting out the transformer winding  with a clip lead did not kill the DC voltage or the oscillation. Therefore, this was not due to the plate transformer picking up stray 60~ a.c. by magnetic coupling to another transformer.  After some head-scratching, I finally figured out what was happening.  I turned off the DC voltage from the power supply by opening a relay in the a.c. line to the primary of the plate transformer, but the filament of the 5R4 stayed on all the time, since it was fed with a separate filament transformer.  With no transformer voltage supplying the rectifier, there were still enough stray electrons emitted from the heated filament, with high enough velocity to randomly hit the rectifier tube plates, to cause a charge to accumulate and generate about +15 volts DC output from the power supply.

My solution to the problem was to change the 6V6 buffer stage to a 6AG7, which is a much better shielded tube at rf.  That took care of the oscillation.  That 6V6 stage must have been extremely prone to self-oscillation.  I still use that rf exciter unit to drive my lower power 8005/805 homebrew transmitter, usually tuned to 40m.