I'm having grid1 current in EL34

[Elias]

Hello,

With my proto SE amplifier with Elektro Harmonix EL34 there is Grid 1 current going when the input signal is high level and V(gate-kathode) approaches zero. The current starts very abruptly and it seems that EL34 grid 1 will eat all the current I feed to it! How much it can handle? Is this anyway normal??

The problem is that the ECC83 driver stage I was planning to use cannot provide the neccessary current. Now I can only get 4W out from EL34

However when I feed EL34 with my solid state amp I can get much more (obviously it can push all the neccessary current). But I'm a little bit concerned of grid 1 current, because any of the data sheet I have do not mention anything about it's specification.

- Elias

 

[Ty_Bower]

What you describe is called Class A2. It is possible, but requires a driver stage which can source (potentially significant) current. As you have discovered, your ECC83 cannot.

 

[Elias]

Hello,

Yes. How much is the max grid 1 dissipation of EL34? Datasheets do not talk about such numbers.

How much will the expected life time be shortened with A2 operation?

- Elias

 

[Eli Duttman]

AFAIK, the EL34 is not rated for a positive grid current regime. THE multi-grid power O/P type to use under such conditions is the 6L6GC.

The 12AX7/ECC81 is a fine voltage amplifier, but it can't provide any sort of significant drive, let alone x2 operating O/P tubes. You can't have coupling caps. feeding x2 operating "finals". Either you drive them with interstage trafos or voltage followers that are DC coupled. Tubelab's Power Drive shows how it's done. The IRFBC20 is very well suited to the task. Unfortunately, the 'X7 is such a wimp that driving the gate of a IRFBC20 could overwhelm it. A solution to the problem is a cascade of FET source followers. DC couple a ZVN0545A's gate to the 'X7 plate. Cap. couple the little FET's source to the gate of the big FET.

 

[Ty_Bower]

Did a little search, and found some interesting articles:

Introduction to Class A2 Amplifiers:
http://members.aol.com/sbench/a2part1.html

Thread discussion of push/pull EL34 running class AB2:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=31772

 

[Jeb-D.]

How much is the max grid 1 dissipation of EL34? Datasheets do not talk about such numbers.

Don't worry about the el34. It's grid can handle more than the ECC83 can throw at it.


If your running triode wired, 4W output isn't that bad. You may be able to squeeze another watt or two out by messing with your operating points, but that's it. If your pentode or UL wired, that's quite underpowered and the design will have to be altered. Pentode wired should give you at least 10W out before the grid starts to draw current.

 

[Elias]

Hello,

AFAIK, the EL34 is not rated for a positive grid current regime.
...
Tubelab's Power Drive shows how it's done. The IRFBC20 is very well suited to the task. Unfortunately, the 'X7 is such a wimp that driving the gate of a IRFBC20 could overwhelm it. A solution to the problem is a cascade of FET source followers. DC couple a ZVN0545A's gate to the 'X7 plate. Cap. couple the little FET's source to the gate of the big FET.

Ok, so probably it's a risk on my own to try A2 with EL34 because it's out of it's specification.

Yes, Tubelabs power drive is one solution. I was just thinking simple FET source follower at first.

If your running triode wired, 4W output isn't that bad. You may be able to squeeze another watt or two out by messing with your operating points, but that's it.

My EL34 is in SE triode mode, and I should get at least 8W out of it with reasonable distortion because that's what I get when I drive it with a separate solid state amp. So far it's a driver issue.


- Elias

 

[Ty_Bower]

...it seems that EL34 grid 1 will eat all the current I feed to it! How much it can handle?

My reasonably is almost certainly flawed, or at the very least questionable. Regardless, the data sheet for EL34 states that grid 2 can handle 8 watts dissipation. I'd suspect grid 1 can probably handle a similar amount of power. The construction of the control and screen grids can't be all that different from each other, can they? How far positive do you need to drive it to get the output you want?

 

[Eli Duttman]

Re: Re: I'm having grid1 current in EL34

My reasonably is almost certainly flawed, or at the very least questionable. Regardless, the data sheet for EL34 states that grid 2 can handle 8 watts dissipation. I'd suspect grid 1 can probably handle a similar amount of power. The construction of the control and screen grids can't be all that different from each other, can they? How far positive do you need to drive it to get the output you want?

You'd best believe there's a difference in construction between g1 and g1. Something has to be done with the heat generated, when power is dissipated. If you get a chance, look at the construction of an EI KT90 or 6CA7. The screen grid posts have radiator fins attached to them.

 

[ray_moth]

I was the perpetrator of the second thread that Ty-Bower quoted - the one about EL34s in AB2. In the years since then, I haven't learned much more about the EL34's grid-current capability but I suspect it isn't great.

The screen (g2) is rugged enough to tolerate 8w dissipaton and is sufficiently spaced from g1 and g3 that a small amount of physical distortion due to occasional overheating will probably not cause it to crash into the other electrodes.

G1 is not at all similar to g2. G1 is made of much finer wire, is very small and is very close to the cathode. I believe that overheating of the EL34's fragile g1 could easily occur and the resultant distortion of its structure would most probably wreck the tube.

I feel sure that the EL34 was NOT designed with grid current in mind and could easily fail if used for that purpose. I've never been game to try it. It's horses for courses and other output tubes exist (like 6L6, 807 and 6550) that seem to be far better candidates for Class A2 or AB2 operation..

 

[Elias]

Hello,

For normal music signal of course A2 operation would be very seldomly happening, because of music crest factor is usually between 10dB-20dB.

Maybe EL34 grid 1 can handle the occasional short peaks, maybe not. No data available, cannot tell.

On the other hand when the amplifier is driven to clipping then A2 is a 'normal' mode and maybe that's the end of that tube then.

So much guessing. Hmm.. I could sacrifice one EL34 on the altair of science for the purpose

- Elias

 

[revintage]

Hello Elias,
In what load and B+ are you only getting 4W? If you are in the ballpark of 3k load and 3-400V B+, something is wrong! You should not need to go A2 until after doubling the 4W.

If you are using your surplus 230V:16V toroid (guess ca 18V unloaded) this represents below 1200:8ohms! If so, unwind until you have ca 12V unloaded.

Personally I do not believe in A2, that shouldn´t be needed with EL34, as it presents a signaldepending load to the driver after going past 0V Ug .

 

[Elias]

Hello,

In what load and B+ are you only getting 4W? If you are in the ballpark of 3k load and 3-400V B+, something is wrong! You should not need to go A2 until after doubling the 4W.

If you are using your surplus 230V:16V toroid (guess ca 18V unloaded) this represents below 1200:8ohms! If so, unwind until you have ca 12V unloaded.

EL34 is biased with CSS at kathode (LM317). I've tried currents between 50mA-80mA. B+ is 340V.

Yes I'm using the toroid as seen here
http://www.diyaudio.com/forums/showthread.php?postid=1620017#post1620017
I made an extra winding on top with several taps (between 4V-16V). The best seems to be 12V as you guess.

4W is the max power without grid 1 current

Maybe this russian EL34 has too low gm? I have only tested one sample of EL34.

- Elias

 

[Merlinb]

EL34 is biased with CSS at kathode (LM317). I've tried currents between 50mA-80mA. B+ is 340V.

[/i]

Hang on, you mean a CCS under the cathode, in a single ended EL34? No wonder you're getting nothing out! How can you generate power in the OT if a 317 is trying to keep the current in it constant! You'll get almost no gain! I say ditch the CCS and use a zener, or something.

 

[Elias]

Hello,

Hang on, you mean a CCS under the cathode, in a single ended EL34? No wonder you're getting nothing out! How can you generate power in the OT if a 317 is trying to keep the current in it constant! You'll get almost no gain! I say ditch the CCS and use a zener, or something.

CCS is bypassed with big capacitor (470u).

- Elias

 

[Elias]

Hello,

I did some more measurements and it seems that I can get 9 Wrms output power with a grid 1 peak current of 13 mApeak.

Without grid 1 current only 4W.

This is triode mode SE.

So it seems I can get more than 3dB improvement.

Should I be worried about 13mA peak?

In real life such a peak would be occuring very rarely taking into account the crest factor of music signal.

- Elias

 

[revintage]

I was wrong about output power of teh triode-strapped EL34. According to the data sheet you should only get 6W(8%THD), 4W(6%THD) or 2,5W(5%THD) into 3k with B+ 370V and Iq 70mA. This is with 0ohm RDC in the OPT.

 

[Elias]

Hello,

According to the data sheet you should only get 6W(8%THD), 4W(6%THD) or 2,5W(5%THD) into 3k with B+ 370V and Iq 70mA. This is with 0ohm RDC in the OPT.

I measured from EL34 alone 2.3W 6%, 4W 8% and 6.3W 8%.

Datasheet forgot to mention you need to feed 4mA grid 1 current to get 6W At least in my case.

When using ECC83 as a driver, distortion improves 1-2% due to distortion cancellation. But with ECC83 max output is 4W, it cannot feed the grid current.

Feed back improves distortion even more of course, I quickly measured 3% with 4W.

There is 9W output possible with 13mA grid 1 current. Now I'm thinking if inserting FET interstage buffer is really worth of the +3dB power increase? With 20dB crest factor of music that seems as a small improvement.

- Elias