pwr tube cathode question

[daz]

no go there either. But i'm happy as is. the switch allows me to A/B the 2 tones instantly and really hear the difference. the loud clean setting definatly sacrifices some tone. But it's fairly close with those small caps. I used 50v caps by the way....hope thats enough. The low position is what i'll be using all the time most likely since i rarely play out anymore. but when i do the cleaner setting may be the ticket.

[daz]

just have to put a bigger signal in from the phase splitter to the power tubes

Can you tell me how i could do that? My current PI is 470R/1M/1M/10ktail and NFB. Isn't that about as gainy as it will get? If not, let me know what to change and i will happily try it.

Tubeswell....good idea there to. i may try that but first i want to investigate whether Peter's notion that i just need to bump the signal at the PI is right. if so that would be the ultimate solution.

[pdf64]

Are you not getting full power with the MV full?
You could get a bit more voltage swing from the phase splitter by reducing the 10k B+ dropper which feeds it, so that it will have more B+ to work with. Try 4k7 or 2k2.
Re redplating, my understanding is that redplating might not occur till 150 or maybe 200% of max dissipation. So just because they aren't redplating, you still need to check dissipation.
Note of caution - the OT primary winding might not be rated for such a high continuous current through both legs. So check it isn't getting warmer than normal. Peter.

[daz]

The PI node r is already 2.2k, and the OT is a 50 watt marshall clone from Hammond. One of the newer guitar series. it actually sounds better in this amp by the way than a heyboer i first had in it.

[daz]

One thing thats beginning to worry me is that from all the reading on the subject i have been doing, i am now unsure whether i am running within safe specs as far as plate dissipation. I'm afraid to rely on the weber calculator for several reasons, so can anyone tell me a reliable way to calculate this? preferably a method a monkey would have no problems doing because when it comes to math my IQ is low double digit. Right now the weber calc shows 25 which is the EL34's limit, but i am not confident it's correct, not to mention the fact thats considered too hot in the real world, no? Doesn't the 70% rule apply to cathode bias too?

[pdf64]

70% is a guideline, not a rule. Check the operating conditions in tube data sheets for suggested circuits, I remember seeing them running at 80%. Class A / hot AB often run pretty much at 100%, ie AC30, 5E3. Avoid going over 100%.
It's safer to run cathode bias hot, as it self regulates ie as the cathode current increases so does the bias voltage, which acts to reduce cathode current.
EL34 is a 30 watt max plate dissipation tube.
Under no signal conditions:-
Measure the plate to cathode voltage V(PK).
Measure the voltage across the cathode resistor V(Rk). Divide V(Rk) by Rk (in ohms) = cathode current Ik.
Plate (+ screen dissipation) = IK x V(PK)
To remove the screen dissipation from this, measure the voltage across the screen resistor V(RScG). Divide this by the value in ohms ofthe screen grid resistor RScG
V(RScG)/RScG = IScG
IP=IK-IScG
Plate dissipation = IP x V(PK)
Peter.

[tubeswell]

Fixed bias is the one to watch out for with idle dissipation. Cathode bias is self-correcting due to the voltage drop across the cathode resistor changing with increase or decrease in voltage, thereby changing the bias, and regulating current flow (as pdf says). With fixed bias, there is nothing to counteract the surges in current that accompany the surges in voltage when signal is applied because the grid is set at an absolute negative voltage with respect to ground, so you need a safey margin for current increases once the voltage on the plate starts to swing.

[daz]

Well, like i said, when it comes to numbers i'm literally a complete moron. I just can't seem to get the final result from that Peter. If you or anyone else want to try it, here are my measurments...

cathode R value=450 (remember, there are two 450's, one per tube)

Voltage across cathode R=29v

plate to cathode=399v

screen drop=8.7

screen R value=1k (there are 2 screen R's)

I have a switch that adds or removes two 25uf 50v bypass caps, but i intend to use it mostly with them switched out.

[tubeswell]

"cathode R value=450 (remember, there are two 450's, one per tube)"

"Voltage across cathode R=29v"

So: 29V/450R = .064A (64mA tube current)



"plate to cathode=399v"

"screen drop=8.7"

"screen R value=1k (there are 2 screen R's)"
(So assuming you have 1K screen grid each tube)

therefore 8.7V/1000R = .0087A (8.7mA) used up at each screen (which if you take it off your 64mA tube current leaves you with .055A (55mA) plate current

so at last we get to:

399V x .055A = 21.94W per tube

22W/30W = 73%, which is about right for fixed bias

[Robert M. Martinelli]

To re-enforce what Peter, Steve and TW said, if you take your measurements at the cathode ( which is THE tube's reference point ) you're measuring the TOTAL current flowing through that tube, so you have to take into account the screen grid current as well ( something many people forgets ) which, as you have seen from the calculations, can typically go up to 20 percent of the bias current, so neglecting it makes for a big measuring error.

I'll add an example : if your +B ( +HT ) or plate voltage is, say, 400 VDC, and your cathode voltage is, say, 25 VDC and your cathode resistor is, say, 500 Ohms ( I'm using rounded numbers to make it easier ) then you have 50 mAmps flowing through the cathode. This current is the sum of plate and screen grid currents, so to state the actual plate current you need to determine the screen grid current first. To do this simply measure the voltage drop across the screen grid resistor and divide it by the screen grid resistor value. If you have a 1000 Ohm screen grid resistor and the voltage drop across it is, say, 10 VDC, then your screen grid current will be 10 mAmps.

Subtracting this 10 to the total of 50 means your actual plate current is 40 mAmps. 40 mAmps multiplied by the plate-to-cathode voltage
( 400-25=375VDC ) gives you 0,04 A*375V=15W ( quiescent plate dissipated power ).

The cathode resistor will have to dissipate 25VDC*0,05A=1,25W.

- End of the example -

This is the reason why you didn't see your tubes redplating at 30W, the actual plate dissipated power was less than that, as some of the power was dissipated by the screen grids.

If two ( or four ) tubes share a single cathode resistor, you simply need to add the two ( or four ) currents, and, since you need the cathode voltage to stay the same, your single resistor will have to be half or one fourth ( Ohms ) of the two or four resistor you were using before. Since the voltage drop ( or cathode voltage, as you like ) will be the same and the current will be twice of four times, the single resistor will have to be capable to dissipate twice or four times the power.

As wisely noted, a very useful feature of cathode bias is its ability to perform a self-adjusting function, ( sort of NFB ) unlike external bias networks.

Bypassing it with a large cap prevents the cathode voltage from swinging and limiting the gain over the whole spectrum of frequencies of interest.

Hope this makes it a little bit clearer.

Best regards

Bob

[Steve Conner]

Did you drill a hole for a switch? I thought you would hook up a relay so that when you footswitch to the dirty channel, the cathode bypass caps get disconnected.

[daz]

Thanks all !!!! Steve...no, not switchable via relay. to be honest i probably will never use that switch, but i just want louder/cleaner if i ever need it. It just sounds so freakin' good w/o the caps.

ok, so i'm dissipating 21.94W per tube....anyone think i should adjust the resistance to bring it up to 25, or is it best where it is. Or simply put, do you thinks it's at an optimum place right now or still too hot or not hot enough? Also, do the figures change at all with caps or does that only affect the A/C signal?

[tubeswell]

I say you could try 390R in each side, and measure everything again (and see if you like the tone) - up to you

the bypass gaps will increase the gain (by counteracting the cathode voltage swing thus holding the cathode voltage relatively constant) - they don't do anything to the overall bias point - bias is determined by the cathode resistor

[daz]

But whats the verdict on dissipation? i mean, if that takes it to 25 will the tubes be getting slammed too hard for any sort of longevity? from what you and others have said i'm hearing that with cathode bias, or "self adjusting" bias that it's ok to reach close to or at max dissipation. But how much longer will tubes last where it is now if at all? Also, someone said 30w is max for EL34, but i always read it's 25. What the truth?

[tubeswell]

Well the EL34s should be able to take increased dissipation in cathode bias. (I guess the OT is capable of handling the extra power?)