EL95 push pull amp tube failure

[octal]

I built a prototype of a little push pull EL 95 amp. I'm using pretty much the datasheet values - 255v b+, 10k opt. Cathode bias. Screen resistors are a conservative 2.2k. During load testing (see the end of the thread on "measuring power output" ) I had an output tube fail.


I changed the tube and it's working again, but I'm wondering why it failed and I'm kind of afraid to take it up to clipping again. EL95 has a max Pa of only 6W, though the screens are actually beefier than an EL 84 (2W EL84, 2.5W EL95)


Are these tubes known to be fragile? I know that you can "bend" the ratings pretty hard on typical guitar amp tubes like EL84s, 6v6, etc,  and I thought that by doing a datasheet design I'd be safe.  Now I'm not so sure.

[jojokeo]

I like these little tubes and not had one fail on me. Your B+ is fine and I'd lose the screen resistors using something like 470r should be just fine. I don't know what you did doing "load testing" but it's suspicious. I'm assuming you biased everything properly not exceeding the 6w Pdiss rating as you don't mention anything regarding this but do speak of the screen resistors? You also didn't mention the newness of the tubes or if you tested them prior to putting them into use?

[tubeswell]

Could just have been a dud tube

[octal]

Yes, I'm hoping it was just a dud tube.  I do have the bias pretty hot. I need to do some math and subtract the screen current out to really calculate what's going on. Maybe I'll increase the cathode resistors a little bit to cool off the bias.


Load testing= running the amp up to clipping into an 8ohm, non inductive power resistor with a 400Hz sinewave input while monitoring the output with a scope and voltmeter. Got a little over 6W out at clipping. Data book claims 7W at 5% THD.


Turns out the screen resistors are 1.5K. Why do you prefer 470ohms, jojokeo?


Thanks!


Nathan

[trobbins]

What failed in the tube?

[tubeswell]

...I need to do some math and subtract the screen current out to really calculate what's going on. Maybe I'll increase the cathode resistors a little bit to cool off the bias.

Looking at the EL95 data sheet http://frank.pocnet.net/sheets/030/e/EL95.pdf, the Ia:Ig2 ratio is 5.33:1. Check that your screen current and plate current aren't exceeding this ratio.

[jojokeo]

Turns out the screen resistors are 1.5K. Why do you prefer 470ohms, jojokeo?

I don't go higher than 1k for a screen resistor because it begins to adversely affect the tube imho. This tube is no more sensitive to run than any other. Just stay within the Pdiss is all. After checking my schem I was thinking about in my prior comment. My amp is single ended and meant to use either 6dl5, 6ar5, or 6aq5. I have 250Vdc on the plate w/ a 1k on the screen at 242Vdc and never a problem with any of the tubes I've run in it. I biased the amp with 470r on the cathode after initially using a 330r with no issues. Since there was no power or tonal differences there was no reason to push it.

[octal]

Oh, you're doing single ended. Kinda apples and oranges then for the operating conditions, yes?

I bumped up my Rk a little.  B+ is a bit lower than I remembered. Subtracted out the screen current and it looks like 20mA per tube. At the 238V B+, subtracting out the 8.5V across the Rk, it looks like I'm idling at 4.6W, so that's about 75%.  Lightly load tested the amp again and nothing happened. Mullard databook calls for 250V, 22mA.

I never figured out the failure mode of the old tube, but it was drawing a boat load of current and the output of the amp was hugely distorted. Maybe it _was_ just a dud tube.

[jojokeo]

Oh, you're doing single ended. Kinda apples and oranges then for the operating conditions, yes?

NO, it's not really as far as if you have a tube failure due to improper use or how it's ran affecting the tube negatively. When a tube's got the same voltages present what's the difference on how it's effected "in general terms" to the screen, plate, and cathode? Also the fact that in single ended use the tube's "always on" rather than pushpull where it alternates cycles and is only "on" half the time and lastly, a single ended tube is biased (on average) to 90% of Pdiss rather than 70% as in pp operation. So I'm actually putting more stress on my single tube that you are on both your tubes - all things being equal.

[kagliostro]

May be there is interest to give a look to this schematics (note that ELL80 is a pair of EL95 in a single bottle - now a very expensive tube)













K


Some images to be seen in full resolution must be saved on your PC