Help me understand measuring/calculating tube plate dissipation under load

[tubby]

I just finished a master volume Bassmann 6G6B build, set the bias on the 5881s to 70% via the 1R resistor method and tested it live -> plays and sounds fine!

Out of interest I wanted to measure if and how the plate current changes when the MV is turned up, so i hooked up multimeter to the amp with no signal applied.
The result: When I set Bass,Treble, Pres to around noon and Gain to max, I can crank the MV and plate current rises a few mA per tube when cranked as measured by the the 1R resistor method.

However, when I apply a 1 kHz sine signal to the input, dime all the tone controls and Gain and start cranking the MV, measured plate current per tube rises well beyond 170% calculated plate dissipation when the MV is turned up all the way.
I accounted for the 40V sagged B+ under full load in the dissipation calculations. None of the tubes were redplating, which doesn't seem right...

Can't seem to get my head around this. Is it because the plate current per tube is fluctuating in push-pull according to the signal applied (in this case sine wave) and the multimeter across the 1R resistor is too slow to measure that?

[Helmholtz]

However, when I apply a 1 kHz sine signal to the input, dime all the tone controls and Gain and start cranking the MV, measured plate current per tube rises well beyond 170% calculated plate dissipation when the MV is turned up all the way.
I accounted for the 40V sagged B+ under full load in the dissipation calculations. None of the tubes were redplating, which doesn't seem right...

With signal present tube dissipation cannot be calculated from B+ times average DC cathode current. Reason is that actual plate voltage strongly varies with signal. Looking at the loadline shows that when plate current is high, plate voltage is low. So actual tube dissipation is much lower than your result.

Assuming your meter correctly measures average DC with AC present, your calculated power of maybe 50W would be the power drawn from the power supply per tube. This power must equal tube dissipation (plate + screen) plus 50% of the power delivered to the load.

[tubby]

Thank you Helmholtz, that makes perfect sense. I totally forgot about the power delivered to the load...duhh...

Now, suppose I want to try 6V6 tubes in that amp. Are there measurements that I could take to make sure the 6V6 I have laying around are not overdissipating under any loading condition (provided bias is set correctly)?
Since the OPT is not matched for 6V6, and there is no tube data published (that I know of) for 6V6 with 420V on plates and screens. There's got to be another method than checking for red-plating...

[pdf64]

Try playing with nickb’s ivds http://bmamps.com/ivds.html

[Helmholtz]

Are there measurements that I could take to make sure the 6V6 I have laying around are not overdissipating under any loading condition (provided bias is set correctly)?

Direct measurement of plate dissipation with signal is not easy as both plate voltage and plate current contain AC and DC. Also AC components are out-of-phase.
It could be done with a DSO that allows for pointwise multiplication of voltage and current signals followed by averaging of the resultant power function.

But you could use what I explained above for indirect measurement:
Measure the DC power (Pdc) through tube + OT just like you did. Also measure output power Pout (will require a true RMS meter for non-sinusoidal output).
Tube dissipation will roughly correspond to Pdc minus 0.5 Pout.

As Aiken shows here https://www.aikenamps.com/index.php/...why-70-percent , max plate dissipation does not occur at max output.

Too low OT primary impedance can significantly increase plate dissipation.

I would not use 6V6s in a BM without modifications (suitable OT, power supply sag). Otherwise expect severe overdissipation and short tube life.

[tubby]

Try playing with nickb’s ivds http://bmamps.com/ivds.html

Thanks pdf64, didn't know that site, great tool!
I've been using this online loadline calculator so far: https://www.vtadiy.com/loadline-calc...ge-calculator/

I don't understand, however, why when playing around with grid bias voltage on nickb's ivds calculator the loadline angle is changing so much. From my limited understanding I would expect mainly the operating point (which isn't shown in the graphs)
to change.

When compared to the vtadiy calculator mentioned above the resulting difference when changing screen grid voltage with otherwise same paramters seems pretty drastic. Different modeling approaches?

[tubby]

Direct measurement seems very intriguing, however that probably exceeds the functionality of my scope...

The indirect measurement method as you have explained seems like a good method for the amateur. Thank you for sharing!

And yes, after putting in the actual values of my build (420V on plates and screens; 4,2K OT primary), I can see that it's not a good idea to try out a pair of 6V6 regardless of bias settings.
However, if I double the reflected impedance by hooking up a 8R load to the 4R output I might get away with it from the looks of the loadline calculator.

[Helmholtz]

420V on plates and screens

If 420V is your full load B+ with 5881s, it will be higher with 6V6s as these higher impedance tubes will produce only about 20W to 25W output power.

if I double the reflected impedance by hooking up a 8R load to the 4R output

Remember that high load impedance is likely to cause excessive screen dissipation.


Why do you want to use 6V6s?

[tubby]

If 420V is your full load B+ with 5881s, it will be higher with 6V6s as these higher impedance tubes will produce only about 20W to 25W output power.

420V was with 5881s at idle, sagging some 30-40V under full load, IIRC...

Remember that high load impedance is likely to cause excessive screen dissipation.

I was under the impression excessive screen dissipation isn't (as much of) a problem with Beam Power Tubes due to beam forming and aligned grid structures. Could you elaborate?
I'm using 1K screen grid resistors on each tube now. Do you think I should I use a higher value with 6V6s?

Why do you want to use 6V6s?

The amp is quite a bit louder than I expected with 5881s and I'd like to use the sound of driven pre AND driven power tubes in my playing. And I always wanted to try 6V6s for a slightly different tone flavor. I was looking at tube data for a tube which would get me a little less power than the 5881s and still be fine with around 420V and a 4k-ish load. Can't use 6L6Gs due to their max voltage ratings, modern 6V6s like the JJ or the TAD-STR seemed to fit the bill. NOS 6V6s might be an option, they probably could take it, but I haven't got money to burn. From what i've heard the JJ isn't a real 6V6 and doesn't sound like one, so I ruled those out. The TAD-STR (which are Chinese Shuguangs) come with 450 max plate volts and 400 max screen volts on the data sheet, so I thought I'll give those a shot. Could be marketing BS of course...

I could use an attenuator, but to me that's just doesn't feel right. Proper power scaling would be another option, but I'm sorta out of space in the amp chassis.

Well, I thought changing to lower power powertubes would be a quick fix, turns out it ain't so easy....

[g1]

Can't use 6L6Gs due to their max voltage ratings,

Typo? Which tube do you mean?

[Helmholtz]

I was under the impression excessive screen dissipation isn't (as much of) a problem with Beam Power Tubes due to beam forming and aligned grid structures. Could you elaborate?

Screen dissipation always is a concern with excessive screen voltage (like ~400V with 6V6s) and high load impedance. Real pentodes tend to be even more critical. Measuring screen dissipation with signal isn't any easier than measuring plate dissipation. Might be easiest to look for "redscreening".
Higher value screen resistors will help somewhat.

I could use an attenuator, but to me that's just doesn't feel right.

What do you mean with "doesn't feel right"?

Remember that substituting 5881s with 6V6s will reduce max output power and sound level by only 3dB.

lower power powertubes would be a quick fix

Lower power tubes are not meant for higher power amps.



Maybe one of these could provide a solution: https://www.yellowjacketstc.com/product-details

[tubby]

I was considering the 6L6G (no typo) because I thought this tube - it being an earlier 6L6 type of tube than say the 6L6GC variant - would have less power. And according to the data sheet a 6L6G has 19W of max plate dissipation which would actually be somewhat less than a 5881 with 23W of max plate dissipation, however max plate voltage for the 6L6G is 360V and max screen is 270V. Didn't think that would go well with 420V...

Here's the data sheet for the 6L6(G). The "G" suffix just means glass envelope. Internally the 6L6 and the 6L6G are the same AFAICT from the data.
http://www.r-type.org/pdfs/6l6g.pdf

[Helmholtz]

I was considering the 6L6G (no typo) because I thought this tube - it being an earlier 6L6 type of tube than say the 6L6GC variant - would have less power. And according to the data sheet a 6L6G has 19W of max plate dissipation which would actually be somewhat less than a 5881 with 23W of max plate dissipation, however max plate voltage for the 6L6G is 360V and max screen is 270V. Didn't think that would go well with 420V...

Lower limits just mean the tube can't take as much. Output power depends on voltages, plate load and tube characteristics - not on limiting values. And characteristics should be the same for all 6L6s/5881s.

A lower limits tube won't limit its own operating conditions. It's up to the circuit designer to make sure that tube limits are not exceeded.

[tubby]

What do you mean with "doesn't feel right"?

I can't put my finger on it, but the sound is somewhat different with the few loadboxes I've tried. And with some of the prices of those attenuators.... Sheesh, I'd rather buy a new set of tubes every 6 months...

Remember that substituting 5881s with 6V6s will reduce max output power and sound level by only 3dB.

I totally trust you on the theory there. You'll also know that a 3 dB change of SPL equals a factor of 0,816 in loudness, that's a reduction of ~20%. And loudness (being the subjective perception of sound pressure) is what we hear, not SPL and not max output power.

Lower power tubes are not meant for higher power amps.

That is true, indeed. But you and me know what Fender did to those poor 6V6s in some amps waybackwhen...Still we love some of them for their tone, right?

Maybe one of these could provide a solution: https://www.yellowjacketstc.com/product-details

Thanks, I'll check those out, although I'm afraid they won't fit since the amp's a head with tubes mounted upright.

[tubby]

Output power depends on voltages and tube characteristics - not on limiting values.

Not sure I can follow... Last time I checked voltages were limited values in tube data sheets?!