How hot should an EL84 run in a class A?

[Randall]

I've got a Trace Elliot Velocette 12R in for losing steam during gigs. I found one EL84 running about 150 degrees hotter than the other, so I put in a fresh pair. It sounds great, but it's getting hot. Both EL84's read around 500 degrees F soon after being switched on. This leads the the chassis and PT getting real warm. All the PS supply V's are within tolerance, nothing looks suspect.

Is this too hot for a class A EL84? The schem from Trace Elliot looks to be a not so subtle nod to a Gibson GA-15RV "GoldTone" amp.

Is this fixed, or do I have a problem?

http://rudn.nodevice.com/preview/big/373/373432-2.jpg

[Leo_Gnardo]

Haven't seen a Velocette since about 20 years ago. I thought they all burned up. NO, 500F is ridiculous! I'm surprised the plates haven't gone white hot. I do remember they were hell on the output tubes and all I could do is dial down the bias by increasing the cathode resistor, try to find some spot where they weren't punishing the output tubes but still sounded good. These days you can order EL84's with low Pc figures from Ruby, CE/Antique and I'm sure other suppliers, that will also help keep the bias current and heat under control.

I don't know if they're still made but at one time you could get finned tube coolers that fit EL84's. I use some on an old Knight kit amp I use on my back-to-mono hi fi system. If I'm not mistook I got 'em from Pearl, an outfit in Canada.

[Enzo]

I have to believe 500 degrees means you are measuring the internal structure temp through the glass.


There is a very famous book titled Fahrenheit 451. It is about a future society that outlaws books. 451 degrees is the average ignition point of paper. Fahrenheit 451 - Wikipedia, the free encyclopedia

So I tend to think 500 degree tubes would be starting fires, let alone amplifying.


Class A power tubes do run very hot. Calculate the dissipation rather than relying on a thermometer.

[potatofarmer]

500F doesn't seem hugely unreasonable. Some data sheets list maximum expected bulb temperatures - I have the 7868 sheet handy and it specifies up to 240C (464F) "at the hottest point on the bulb surface". I bet Enzo's right about the measured temp reflecting the plate temperature. IR temperature guns really only work accurately with high-emissivity materials, like flat dull steel. Even shiny metals can read inaccurately. Red plating doesn't start until the 500C (900F) ballpark in plain steel.

So yeah, apart from "don't touch it," temperature readings aren't giving you any useful information.

[Dave H]

What is the EL84 cathode resistor value? I can't quite read it. If it's 120R and with the B+ at 360V I'd change it to 150R. That's what I used in my 18W and it sounds fine and doesn't run too hot. It won't be class A even if the EL84s are running at 14W each.

[Alf]

Just looked if I had a schem of the thing. Just a handdrawn one but there the resistor is 120 ohm and the cap is 220 uf.

I once measured the temperature on the outside of a EL84 tube glass envelope of a Vox AC30 and it was 220 degrees centigrade which is 428 Degrees in fahrenheit.
I upped the resistor of 47 ohm to 68 but the owner wasn't happy with the sound so I put the original one back.

[Randall]

Enzo is correct, I am measuring the internals thru the glass, so so much for that.

When I calculate my dissipation, I get 17 watts, which I don't understand. Even looking at the values on the print:

12 volts across the 120R cathode resistor 12v / 120R = 0.100A / 2 = .050A per tube.

12v cathode voltage subtracted from plate voltage Vp = 350v - 12v = 338v

338v * 0.050A = 16.9 watts

Is this correct? I thought an EL84 was rated at 12 watts?

[g1]

Yes that's correct. EL84's are often run way beyond their max power rating. You should be able to reduce the bias and that should cool everything down a bit.

[Dave H]

Is this correct? I thought an EL84 was rated at 12 watts?

EL84 is rated at 12W Plate dissipation. You have used the cathode current which is plate + screen current so you have calculated the plate + screen dissipation at 17W. The screen current will be about 1/5 of the cathode current. The plate dissipation is therefore 4/5 of 17W = 13.6W and the screen dissipation 3.4W.

[g1]

I thought screen current would be more like 10%, rather than 20% ?

[Dave H]

I thought screen current would be more like 10%, rather than 20% ?

So did I until I looked at the EL84 data sheet. It says 46mA plate and 11mA screen.

Edit, You are correct. I was reading the data sheet values for 17W output and Randall is measuring at idle so plate dissipation at idle is 9/10 of 17W = 15W, even worse for the EL84s and as it's a class A/B amp they could get even hotter with an input signal.

[oc disorder]

All the PS supply V's are within tolerance, nothing looks suspect.

My tuppence worth is to re-check the heater supply as I can recall a similar model with high heater voltages and fitting wire wound resistors to bring it down.
The supply here is 230+ so it may have just been an odd one.

[Randall]

OC, The heater supply looks good.

Dave H. this is a Class A amp, is it not? It says so on the chassis.

So what is the verdict? It just runs real hot, and maybe has always done so? Give it back and suggest running a fan on it?

[pdf64]

Unless the PT primary impedance is much higher than the normal 8k used for EL84 p-p pairs, from Is the Vox AC-30 Really Class A? with 338V across the power tubes, I can't see that it could operate in class A.

[Dave H]

Unless the PT primary impedance is much higher than the normal 8k used for EL84 p-p pairs, from Is the Vox AC-30 Really Class A? with 338V across the power tubes, I can't see that it could operate in class A.

You are right the AC30 isn't class A but it was clearly the best sounding amp in the UK back in the '60s which makes it first class to me.