Recommendations for EL34 for 520V on the plate

[abbottamps]

I'm wondering if anyone has a recommendation as to what could be put into an Orange OR80. It keeps eating tubes because it's putting around 520 on the plates. The plates *aren't glowing red, but the grids are when the amps cranked (which the person usually has it full volume)... to try and help the tubes, I put in a large 2k chassis mount resistor in series with the two grid stoppers (2.2K's)... seemed to help the tubes a bit and they lasted longer than previous tubes, but it probably needs a bit more resistance to stop the glowing grids. The Winged C's I put in eventually gave out and catastrophically failed

If anyone has any idea as to a good strong EL34 that could withstand this beast of an amp, your suggestions are appreciated. I'm going to increase the resistance for the grids sake, but I still think it's a fairly harsh environment.

Thanks,
Travis

[Greg Robinson]

Are you serious?
The plates are glowing red at idle, have you tried adjusting the BIAS yet?
Also, which grids are you talking about? An EL34 has three of them. The control grid, the screen grid, and the suppressor.
Normally when you just say "grid" you are referring to the control grid. Adding series resistance here will not help.
I assume you're referring to the screen grids. Increasing the series resistance here will help to limit current at clipping, but you still haven't dealt with the underlying problem.
Put th screen grid circuit back to stock, adjust the BIAS for the correct idle dissipation. If you get premature screen grid failure, disconnect the suppressor grid from ground, and tie it in the the raw bias supply. Still getting premature failures? Then increase the screen grid resistors.

You run a business? Really?

[abbottamps]

*Plates aren't glowing red... that was a really bad typo. Screen grid is what I'm talking about, that's what I added resistance to. It seemed to help after adding the extra 2k in there, as this round of tubes lasted a relatively long period of time, but it still doesn't appear to be enough. I tried taking the amp back after seeing the tubes after a few minutes of being completely cranked, so I could add more resistance, but my friend insisted on just playing it because he needed it for recording. The grid only appears to glow red when it's cranked (which it always is). The OR80's are infamous for killing power tubes quickly, and it's running two EL34's and outputting 80 watts. Apparently newer production tubes can't really survive.

I think that I'll have it under control after adding some more resistance, and I'm just asking if anyone has anything more reliable in the realm of new production tubes that would be a solid choice for an application like this one. I'm biasing them around 70%.

[Greg Robinson]

Ok, if it's not already done, disconnect the suppressor grid from the cathode, and connect it to the raw bias supply (the bias supply filter cap). This will help limit screen current. Then try increasing screen resistance.
No new production EL34's meet the true specs and will struggle in guitar amps that push the boundaries. In my experience, JJ's are probably the most reliable. This amp might be a candidate for NOS only.

[soundmasterg]

You could also adjust the bias down lower to 65% for example. The amp will likely sound just as good but will operate in a more forgiving range that will extend the tube life.

Greg

[jpfamps]

Ok, if it's not already done, disconnect the suppressor grid from the cathode, and connect it to the raw bias supply (the bias supply filter cap). This will help limit screen current.

Won't this increase screen current as it will repel electrons?

[jpfamps]

JJ E34L is probably your best bet of modern production.

I've used these in Laney Super Groups, 600V plate and screen, with 2.2k screen grids and they seem to survive OK.

[Greg Robinson]

Won't this increase screen current as it will repel electrons?

No, the cause of excessive screen grid current in overdriven guitar power amps with true pentodes like the EL34 is that when the anode potential (voltage) drops below the quiescent screen grid voltage, the space charge that has accumulated around the anode is drawn back to the screen. It is limited by the suppressor, but it is a lossy electrode and is not very effective when the valve is heavily overdriven. Lowering the potential of the suppressor does not have a measure able effect on quiescent operating point, but makes it more effective at repelling the space charge during overload conditions.
Beam power pentodes, beam power tetrodes and kinkless tetrodes are all inherently less likely to develop screen grid overcurrent conditions. This is why it's unusual to hear someone talking about melting screen grids in a 6L6GC amp, but very common for EL34's.

[soundguruman]

Yes, the screen current repels electrons, if you believe in those little things..
the more there is, the less sensitive the plate becomes.

[soundguruman]

I'm wondering if anyone has a recommendation as to what could be put into an Orange OR80. It keeps eating tubes because it's putting around 520 on the plates. The plates *aren't glowing red, but the grids are when the amps cranked (which the person usually has it full volume)... to try and help the tubes, I put in a large 2k chassis mount resistor in series with the two grid stoppers (2.2K's)... seemed to help the tubes a bit and they lasted longer than previous tubes, but it probably needs a bit more resistance to stop the glowing grids. The Winged C's I put in eventually gave out and catastrophically failed

If anyone has any idea as to a good strong EL34 that could withstand this beast of an amp, your suggestions are appreciated. I'm going to increase the resistance for the grids sake, but I still think it's a fairly harsh environment.

Thanks,
Travis

Well you gotta have a 1 K screen grid resistor at least, in series with the screen.
For EACH tube.
Or it will for sure blow chunks.

Sounds like your voltage selector is wrong or something.
You are trying to compensate for a malfunction, I would venture to guess.
There's something else wrong.
There is a distinct possibility that you are fighting a high frequency oscillation, causing too much current draw.

[Steve Conner]

You could try an impedance mismatch of one step (8 ohm speaker on 16 ohm tap). Loading the plates heavier diverts electron flow from the screens, at the cost of increased plate dissipation.

Or you could go to 6550 or KT88-type tubes: being beam tetrodes these draw less screen current. The Tung-Sol reissue 6550s might be a good bet.

I'd also check that the customer isn't running with an impedance mismatch in the opposite direction: a blown speaker in a 4x12 or the like.

I don't believe changing the voltage on the suppressor grid is good for anything.

[jpfamps]

No, the cause of excessive screen grid current in overdriven guitar power amps with true pentodes like the EL34 is that when the anode potential (voltage) drops below the quiescent screen grid voltage, the space charge that has accumulated around the anode is drawn back to the screen. It is limited by the suppressor, but it is a lossy electrode and is not very effective when the valve is heavily overdriven. Lowering the potential of the suppressor does not have a measure able effect on quiescent operating point, but makes it more effective at repelling the space charge during overload conditions.
Beam power pentodes, beam power tetrodes and kinkless tetrodes are all inherently less likely to develop screen grid overcurrent conditions. This is why it's unusual to hear someone talking about melting screen grids in a 6L6GC amp, but very common for EL34's.

Yes I know this.

The lower screen grid currents in beam tetrodes are due to the grid and control grid being wound concentrically, so the screen grid is in the shadow of the control grid. I expect this is why beam tetrodes are generally more expensive than pentodes as better tolerances are needed to be held in the grid winding.

My logic in my assumption that tying the suppressor grid to the bias supply would increase screen grid current is that when the plate voltage drops significantly below the screen voltage any electrons that pass through the screen grid will be slowed more by negative suppressor than one at ground potential, and thus are more likely to be attracted back to the screen grid, and hence produce higher screen currents on clipping.

[Greg Robinson]

Yes I know this.

The lower screen grid currents in beam tetrodes are due to the grid and control grid being wound concentrically, so the screen grid is in the shadow of the control grid. I expect this is why beam tetrodes are generally more expensive than pentodes as better tolerances are needed to be held in the grid winding.

My logic in my assumption that tying the suppressor grid to the bias supply would increase screen grid current is that when the plate voltage drops significantly below the screen voltage any electrons that pass through the screen grid will be slowed more by negative suppressor than one at ground potential, and thus are more likely to be attracted back to the screen grid, and hence produce higher screen currents on clipping.

The screen grid has a very small surface area compared to the anode, and really only helps accelerate electrons towards the anode, while drawing only a small proportion of the anode current in normal quiescent operating points. Remember it's the space charge that flows through the screen during clipping, and lowering the voltage of the suppressor helps prevent this. This is why pentodes were developed after tetrodes, beam/kinkless tetrodes were developed later to avoid royalty fees/patenet infringement. This is detailed in RDH4, or try it yourself. It works.

[jpfamps]

The screen grid has a very small surface area compared to the anode, and really only helps accelerate electrons towards the anode, while drawing only a small proportion of the anode current in normal quiescent operating points. Remember it's the space charge that flows through the screen during clipping, and lowering the voltage of the suppressor helps prevent this. This is why pentodes were developed after tetrodes, beam/kinkless tetrodes were developed later to avoid royalty fees/patenet infringement. This is detailed in RDH4, or try it yourself. It works.

The screen grid doesn't "help" accelerate electrons from the cathode, it is almost completely responsible for accelerating electrons from the cathode.

This should be self evident from the near horizontal portion of a pentodes load lines; increasing the anode voltage has little effect on anode current.

The anode can only really exhert any influence on electron flow once the electrons have passed the screen grid.

If the anode voltage is significantly below screen grid, the screen grid current rises (as you point out).

I have always assumed that this is because some electrons, once past the screen grid, are accelerated back to the screen grid as the screen grid is at a significantly higher potential than the anode.

I can't see a mechanism whereby under these circumstances increased screen grid current is caused by greater secondary emission from the anode as the decelerating effect of the screen grid should reduce secondary emmission.

Tying the suppressor grid to the bias supply would only enhance the deceleration of electrons under these conditions, ie increase screen grid current.

[Steve Conner]

I'm sure I remember busting the suppressor grid myth about 10 years ago on the old Ampage, with lab tests. Unfortunately I can't remember exactly what myth I busted. Under peak signal conditions (heavy conduction with low plate voltage) something was optimal with g3 at 0v and got worse with either positive or negative bias. But I'm not sure if it was a maximum of plate current or a minimum of screen current.