Cure for a thumping trem

[stokes]

Tremolo in a Brown Princeton clone is thumping like mad,gets worse as intensity is raised.Last year I built one of these for my son and it doesnt thump.I built the second one for myself,identical except I used EL84's.The tremolo works but it has a loud low freq thumping when not playing,not as noticeable while playing.I've tried changing out the caps,different tubes,I lowered the current draw on the power tubes which lessened it some but not enough.The plate of the osc tube is fed by the same B+ node as the OT.I am thinking about either moving the feed to the screen node or adding a resistor/cap before the plate resistor to de-couple the stage.What do you guys think?Anybody got any other thoughts on quieting this?The amp sounds great when the trem is off.

[tubeswell]

Yes I'd try moving the HT supply to a different HT node (decoupled) with a R/C power supply filter, and change the LFO ground return path(s) to ground at that filter cap's ground return

[stokes]

Ok,problem solved.I moved the plate feed to the screen tap in the B+ rail,not much,if any improvement.The problem got extremely severe if the intensity knob was turned past the halfway point.I remembered a "fix" for the Blackface Princeton tremolo being too weak.In the Blackface there is a 1m resistor in series with the .01 cap fedding the intensity control,to strengthen the trem you reduce this resistor.In the Brown Princeton the resistor is 220k,I increased it to 470k and it cleared it up substantially.If the intensity is turned up all the way the oscillation is back some,but not too bad.I'm going to try the 1m resistor but the 470k is good.I built this amp because the one I built for my son had such a good trem,I guess the EL84's have a different effect on the trem,its the only thing different in the 2 circuits.

[tubeswell]

Yes the bias voltage on EL84s is about 1/2 of what it typically is on 6V6s, so bias-wiggle trem that is designed for 6V6s will slam the EL84 grids hard, unless it is mellowed. Another 'solution' might be to run the EL84 bias a wee bit hotter - the increased tube current 'eats up' some of the grid bias wiggle. Just be careful how hot you go, bias wiggle can cause the output tubes to exhibit hotter plate dissipation when engaged.

[stokes]

I originally had them biased at 70% which had the bias voltage at -17vdc.I raised it to -21vdc which lowered the current draw and lessened the effect some.My thinking was to bring the bias voltage closer to the -35 needed for the 6V6's.Anyways,all is now good,the amp sounds great and the trem is good.