Understanding "grid drive" to drive, 2, 4, 6 power tubes

[nickb]

Using which method?

The average of I(t).V(t).

[Helmholtz]

The average of I(t).V(t).

Thanks!
Problem is that there doesn't seem to exist information about relevant integration /averaging time. I tend to think that screen wire temperature is the limiting factor, which shows best in glow color.

[nickb]

Thanks!
Problem is that there doesn't seem to exist information about relevant integration /averaging time.

I chose an integer number of cycles.

I tend to think that screen wire temperature is the limiting factor, which shows best in glow color.

There's certainly the virtue of simplicity in that approach

[Helmholtz]

There's certainly the virtue of simplicity in that approach

..and the correct integration time is guaranteed.

[nsubulysses]

This is where we went over this a couple years ago https://music-electronics-forum.com/...t=42959&page=2

[Helmholtz]

I chose an integer number of cycles.

Sorry by causing some confusion by inquiring about "integration time". I forgot that we are dealing with a steady state situation here and all information necessary about screen dissipation is contained in one full cycle at equilibrium: One half-cycle heating, the other cooling. Integration time would only be critical at dynamic/varying level operation.
Seems 3 a.m. was too late for me last night.

[J M Fahey]

No, you weren´t wrong.

At least not if we are talking a Guitar amp being played loud, live, and where input signal is a guitar and not an oscillator.

In that case there is no steady state, so we must consider another factor: screen will have a thermal time constant, like any physical object with mass.
The screen wire mass will turn electrical power into heat, we already know that, but since power signal is not constant, instant wire temperature will vary .... after a short but real time constant and of course it will integrate power input.
Consider it an RC net, it will integrate any and all pulses sent to it.
Can´t give an exact value (somebody must have studied it of course) but hink it must be around a second or so .... the point is that such time constant is way longer than any guitar signal frequency can have.

We are worried about temperature peaks, of course, which cause meltdowns.
Thanks God, at least temperature peaks can be seen, so screen brightness is an accurate representation of temperature, so at least for me, first roughly calculating screen dissipation to be in the ballpark of datasheet recommendations (even if within 2:1 tolerance or worse), I could then fine tune it by simply having a friend play guitar full blast and me watching screens in a semi darkened room and adjusting screen voltage with a homemade regulator until they glowed orange on peaks but not bright white ... which happened at the beginning and was SCARY.

All the work I did was only on 6DQ6, the *only* power tube available when Argentina got into Default in 1972 and all imports were stopped cold turkey.

6DQ6, (plus 5U4 and some other "TV tube" I don´t remember) were locally made at Deltron, an RCA subsidiary, so we had as many as we wanted.

[Helmholtz]

Yeah, that's exactly what I meant originally.

But then I realized that the question here was how to measure screen dissipation at full power, which is a steay state/equilibrium situation. This measurement makes sense, as screen dissipation is maximal at full power (other than plate dissipation).
But a real speaker load is different as its impedance increases at low and high frequencies. A higher load impedance means that plate and screen currents will be high for a longer part of the cycle, thus increasing average screen dissipation compared to nominal impedance.
As screen temperature is the relevant parameter (integration included), its glow color is the best indicator in dynamic operation.

It is not advisable trying to measure screen dissipation using a real speaker or a speaker impedance emulating load doing a frequency sweep at full output!

[uneumann]

I don't recall where I've seen this before, but I use a panel LED to visually monitor screen currents with a circuit like this



The R53 resistor sets a low threshhold for turning on the LED - say at some measured power level like 30 - 50% max power.
R50 sets a high current threshold to keep the LED from burning out by turning on the transistor to bypasses current over the threshold.
A ~40ma diode passes most screen current under most conditions so it's an effective visual monitor of real time behavior.
Your eye integrates the variations easily. It's not a precise numeric readout, but you can see problems immediately and you get a good sense of what "normal" behavior looks like.

[Zozobra]

Thanks for all the help along the way

Anyone know primary impedance for 6xEL34 sound city 120 OT or 6xEL34 DR201 OT by chance?

I guess my recollection of the older 6 power tube amp I made doing 196W with 6X6CA7 was a fluke or I am remembering it incorrectly (but then again isn't it common for amp builders to stretch the truth a bit with these types of things similar to how they say people who like to go fishing do !!??!?!?! whoops)

Not quite what you're after, but InMadOut do a 6x 300W OT that has a 1k2 primary:
http://www.inmadout.com/download/pdf/TU300.pdf

I'm sure they used to do a 6xEL34 DR201 OT but they only have the 4xKT88 version on offer now which has a 2250 ohm primary.


As for the screens, rather than using a resistor on the filter node have you considered a MOSFET regulator? I'm trying the circuit 3/4 of the way down the page on my current build with a pair of KT88s:
http://education.lenardaudio.com/en/...ar_amps_4.html

I've built the circuit at tested it without tubes but I'm yet to get any further with it as I've been busy with other things

[SoulFetish]

I don't recall where I've seen this before, but I use a panel LED to visually monitor screen currents with a circuit like this

[ATTACH=CONFIG]51788[/ATTACH]

The R53 resistor sets a low threshhold for turning on the LED - say at some measured power level like 30 - 50% max power.
R50 sets a high current threshold to keep the LED from burning out by turning on the transistor to bypasses current over the threshold.
A ~40ma diode passes most screen current under most conditions so it's an effective visual monitor of real time behavior.
Your eye integrates the variations easily. It's not a precise numeric readout, but you can see problems immediately and you get a good sense of what "normal" behavior looks like.

Most LEDs (standard through hole types) have pitiful current ratings. But this may helpful in simplifying your design. There is a cheap simple red LED: MV50154 which is rated to handle 100mA of continuous operation which I found a few years back. I remember doing some experimenting on a design using a simple LED biasing arrangement so I could switch between true cathode biasing on a pair of EL84's, and a emulated "fixed bias" effect using these LEDs. This was possible because of the MV50154 which had a remarkably linear operating range and extremely low internal resistance. Since both cathodes were biased through a series string of these, I could hold the cathodes at fairly constant DC operating point without worrying about one half entering cutoff, turning off the LEDs, and floating the cathodes.
Anyway, the added benefit to this was being able to visually see when the amp would shift from the class A portion of operating in AB, to class B where the peak currents would noticeably increase the LED output. Perhaps these might work well in this application as well.
Here is a some data I had on them:

[Jime]

nsubulysses - where is your shop in Chiraq? I live way north (near WI/IL state line). I need an electronics technician to fix an amp. I'd rather talk one on one about my situation vs. in a thread...I'm a complete noob and although I'd like to try and DIY...I'm pretty sure it is way beyond what i could reasonably figure out on my own. I don't want to bog down expert sites like this with noob questions like "what is impedance"...you can learn a lot on google but for my fix...i am pretty sure it is beyond simply replacing the power transformer and need I need an expert. Thanks! JimE

ps - apologies if asking this sort of thing in an ongoing thread is a nogo here...I don't see a way to message members individually (like I said...I'm a noob).

[g1]

Jim, the private message (PM) function is suspended awhile for new members. Maybe a moderator can comment on how long that lasts.
Meanwhile, if nsubulysses has an alternate method for contact, maybe they will post it here.

[Jime]

Jim, the private message (PM) function is suspended awhile for new members. Maybe a moderator can comment on how long that lasts.
Meanwhile, if nsubulysses has an alternate method for contact, maybe they will post it here.

cool, thanks G1

[nsubulysses]

Sorry I can't take on repair work. I would recommend starting a thread if you have some questions and me or other people way more knowledgable than me will answer.