bias and grid resistor power stage

[krash]

Your problem could be excessive screen grid current.

I had EXACTLY the same symptom on an amp I built with a tried and true routine design I use all the time and banged my head against the wall for a long time, until I discovered that different tubes made it go away. I added 2K more screen resistance for each tube and voila, it went away with no ill effects. Maybe these JJ tubes have changed in some way recently, because I never heard this before. I kind of expect that I might be seeing some of my amps come back for this mod as people replace tubes with new ones.

[rfgordon]

I truly appreciate the tip about the negative DC voltage swing on the grid.

I put the meter on the output tubes, and, sure enough, there it was. I checked the grids on the peanuts, too, and there was some on the post-verb mixer stage. Simply out of convenience I had popped in a 2M2/20pF combo that I had taken out of another amp. I swapped it for the 3m3/10pF, and that voltage got steady.

So, on the power tubes, I've only had time to change the feeder caps from .1 mF to .022 mF. Didn't help. So I'm gonna change the caps upstream from there.

Also, I know that blocking distortion can be a result of biasing too cold. I'm running 360 VDC on the plates here, which is much lower than my usual, due to the PT I'm using. So I'm thinking that the 250 ohm Rk per pair may be too big.

I'll also try upping the screen grid resistors from 1k5.

[dobbhill]

Rich:
I had similar sounding problem. It turned out to be a leaky coupling cap.
Just sharing my experience hoping it might help....
D

[rfgordon]

Well, it ain't fixed yet. I've replaced every coupling cap, adjusted the lead dress, tried different power tube Rk values--nothing makes a difference.

My next step is to try higher screen grid resistors on the output tubes.

The step after that is to take the amp to a guy in town who used to work for Marshall...while waving my pathetic white flag of surrender......

[rfgordon]

I scrounged around and found two 1k5 power resistors (1 5 watt and one 3 watt). I figured, what the heck, so I put them in series with the existing 1k5 resistors on one pair of output tube sockets.

Well, waddya know...now there's just the tiniest bit of fiz on the very end of the note decay when the amp is completely maxxed out!

Thanks, krash, you were right on the money!

[krash]

well, glad that tip helped about the screen R's. It really confused me too.

Take a look at the JJ web site, their tube data shows Ig2 (screen grid current) on their EL84 rated at 5.5mA and 3W for grid power. The 6V6 & 6L6 data sheets don't have as much detail but they show 5mA for Ig2 for each of those.

So when I did the math for my EL84 amp it looked like I needed at least 2K to limit screen power to less than 3A and also limit screen current to less than 5.5mA.

Randall Aiken posted some info about this in his comparo of EH and JJ EL84s here:

http://www.aikenamps.com/JJ_EH_tube_tests.html

I would expect that many folks are plugging in circuit values for screen resistors taken from 50s and 60s amps when the tubes du jour had far more robust screen grids compared to today's tubes.

[rfgordon]

This amp is running a pair of EH 6L6s and a pair of NOS 6V6s branded General Motors (man, when was the last time GM put tubes in a radio?).

But, in any event, it's working now. I also ended up putting 220k grid stoppers on v1b, and v3b. The v1b had the 500pF cap across it, a la ODS.

Now the thing sounds really rich. The 6L6/6V6 combination is very nice.