EL84 pp blocking distortion

[octal]

Do you have screen resistors as well, or are the screens directly connected to the ultralinear taps?

I once tried EL84s connected in UL w/o screen resistors and the screen grids lit up like a Christmas tree when I brought it up to clipping. I know that screen grid resistors are often left off HI-FI UL amps, but a guitar amp is a whole different kettle of fish, and load testing is particularly brutal on tubes.

For that matter, why not just configure them as non UL to eliminate that as a potential trouble source? You can go back to UL once you have the problem sorted out.

I'm wondering if you somehow managed to damage both transformers or multiple sets of tubes...

Nathan

[ValveMan]

I'm using 100r resistors on the screens - thought that would be enough - you reckon they should be higher. I guess it's possible there's now something wrong with the development board/chassis that's fried some stuff.

I'll bring up the new build in pentode mode first & see what things look like.

Reason for using UL is to make it a little friendlier to acoustic guitars. The first several test units sounded sweet. It's only since I started tweaking the dev unit to add gain & increase volume for a particular pickup setup (still keeping it clean) that I ran into this issue.

Another possibility is that the PT isn't big enough (272BX rated to 115mA) since increasing the gain in the pre-amp. Although I did try a 272DX which puts out 125mA . .
It's connected to a 5Y3GT running 2 12AU7's, drawing ~11mA between them and the EL84 PP.
Altogether there's 60uF & an 8H choke for the B+, then 3 20uF caps for the 2 12AU7's (one for the PI and one for each half of the other 12au7)

Any other words of wisdom are very very much appreciated. I'll clean the schematic up after this build in case it's needed for deeper discussion.
Cheers!

[Wombaticus]

I haven't had an opportunity to do much more than skim over the thread, but I'll point out that the 470uF cathode bypass cap is unusually large.

In a cathode biased power stage, excessively large bypass caps will dramatically increase the recovery time from bias excursions and can lead to blocking distortion.

[ValveMan]

Thanks W,
It had 50uF/50v when the problem first started (now too) - I'd experimented with various sizes without much of a difference.
Keep the questions coming! :-)
Cheers
Gerry

[Bruce / Mission Amps]

With respect to the 12AU7 PI, I was referring to trying low resistance, 12K to 47K plate load resistors, not the grid load resistors, although I would expect 100K to be fine in that spot.
It really just looks like there is too much stage gain in the amp and the EL84s are over driven.
Once those EL84 grids fold over like that, your stuck there for a while! ha ha
The last real EL84 amp I built ended up using two average level gain stages with a std cathodyne driver and I found that I liked the sound of the amp better with 2200 ohm 2K2 screen resistors, not 100 ohm.
I think the main thing is what you already know, EL84s require so little drive to turn on full blast that it very easy to over build the preamp stages and end up with a diode in the power section!

[ValveMan]

Hi Bruce
Yup - I had 22K plate loads on the cathodyne. I might actually go there again once it's up & running - the signals looked really nice.

The overall gain can a wide range, but I'm not getting past +5V (10V pk-pk) into each EL84 before the clipping kicks in.
I'll try larger screen resistors & see what gives.
Were you configuring your EL84s as pentode or UL?
Cheers
Gerry

[pdf64]

Have you tried it configured as pentodes, rather than UL, yet?
Just to eliminate it, I'd try a regular OT in there - you might be doing something silly with the taps, which just isn't registering with you.

[Bruce / Mission Amps]

Whoa!! You mean 5vac per grid or 10vac PP across two grids folds up the tent?
Man something is broken.

[ValveMan]

No - 10Vac PP per grid. (which would be 7V rms)
Looking at one grid, it clips at the +5V swing. - same on the opposite side when the signals swings around & goes +5V on that side
I guess the problem is that the gain is now only 12-15 depending on the bias point, whereas for the same +/-5V signal level previously it would give a gain of 20 providing +/- 100V swing per tube

[anson]

Just checked the DC levels on the screens - they're actually 2V higher than the plate (both idle & working) - could that be a problem?

Correct me if I'm wrong, but the screens should never be higher thatn plate voltage. they should be around 42% of that in UL configuration

[GregS]

Could the value of the grid leak resistors have anything to do with it? What are they?

Greg

[ValveMan]

I've been out of town for the last week - thanks for the continued support!
Anson - the voltage swings are 42% of the plate swings. I'm not sure about the implications of being 2v higher at idle, but I suspect it's normal (lower dc resistance from taps)

Greg - I'm using 150K right now (was trying to mitigate grid clamping), but I'll kick that up to 470k and see what happens

[ValveMan]

Simple question (hopefully . . . .

Ok . . . . changed grid leaks to 475K; backed the coupling caps down to .022uF; experimented with various levels of B+ between 270V and 340V; adjusted the bias between 7V and 13V at all those settings; also switched to pentode mode (284V tap to 1K on screens) . . .
. . still not getting any more than 150Vpp per side, with clipping starting at +5V signal on the grid.

What's a reasonable upper limit to the signal (peak-peak on a single side) for the max 'clean' voltage swing at the plate for an EL84 PP hooked to a 7K6 25W transformer (Hammond 1650F)?

[imaradiostar]

1 volt rms sine = 2.8 volts peak to peak. Double check your math- count the divisions on the scope. Any cheap-o digital volt meter should be within 5% or less of the true RMS value when measuring a clean sine wave.

I may have missed it but I didn't see a post where you disconnected the output grids from the driving tube and verified that the cathodyne isn't the cause of the problem.

You're not going to get much wattage out of an EL84 at 270 volts. It's still gonna take screen voltage close to plate voltage to get them to draw much current. You can draw all the current in the world and it won't matter if you don't have an appropriately low output impedance. Have you tried mismatching the impedance up or down?

Surf over to Randall Aiken's web site- he had a pair of El84's running at some pretty high voltages and he had to use huge screen resistors to keep them from blowing up and he still didn't get much wattage. I'd say a big part of why EL84's sound good is that they can't produce much output without distortion. Their real benefit is the lack of drive required to produce full output but this still assumes screen voltages around/above 300 volts.

Of course, this is my experience. Your mileage may vary!

jamie

[ValveMan]

Thanks Jamie!
"1 volt rms sine = 2.8 volts peak to peak. Double check your math- count the divisions on the scope. Any cheap-o digital volt meter should be within 5% or less of the true RMS value when measuring a clean sine wave."
Yup - using a Fluke 75 (non-RMS) & a scope, so when I say 10V p-p that'd be 7V pp RMS or +/- 3.5Vrms. That's all it takes to clip.

"I may have missed it but I didn't see a post where you disconnected the output grids from the driving tube and verified that the cathodyne isn't the cause of the problem."
Did that - also switched back to the original LTP too

"You're not going to get much wattage out of an EL84 at 270 volts. It's still gonna take screen voltage close to plate voltage to get them to draw much current. You can draw all the current in the world and it won't matter if you don't have an appropriately low output impedance. Have you tried mismatching the impedance up or down?"
Yeah - did that after verifying the 8ohm was configured right (voltage ratio of 16:1 with 7K6 on primary a-a)

If my math's right, to get 15W at the speaker would take 11V RMS p-p into an 8ohm load - which would be 15.7V pk-pk sine. The most I get with the +/-150V p-p each side on the primary is 9.5V p-p which would give ~5.5W
(9.5 x 0.707 = 6.7Vrms 6.7 x 6.7 / 8 = 5.6W)

"Surf over to Randall Aiken's web site- he had a pair of El84's running at some pretty high voltages and he had to use huge screen resistors to keep them from blowing up and he still didn't get much wattage. I'd say a big part of why EL84's sound good is that they can't produce much output without distortion. Their real benefit is the lack of drive required to produce full output but this still assumes screen voltages around/above 300 volts."

Agreed - but I'd hoped 15W of clean power could be squeezed out of the 84's before clipping
I've 10 matched pairs - bought from a reputable & helpful dealer - haven't tried them all yet because I wanted to make sure they weren't getting killed by something stupid in the circuitry.
I need to clean up the schematic so I can post it & ask for opinions.
Cheers!
Gerry