Low output on JTM45 build

[tubby]

Just finished a JTM45 build, basically a clone with a few mods: elevated heaters & humdinger pot, "Frondelli" PPIMV (220k), variable NFB, 470k mixer resistors, 130R Rt resistors & dual 1n4407 diodes on PT HV secondary each side.
Amp sounds very good, just seemed a lacking in volume.
I measured the output (1 kHz sine into a 8R load) with barely 14VRMS at the onset of distortion. That eqals around 25W. From what I understand most JTM45s push around 30 to 35W clean.

So far I've tried new tubes (pre, power & rectifier), SS rectifier, more filtering, 1000pF snubbers at the grids of the pre & power tubes in case of oscillations. No discernible difference in power output.
Waveforms look good on the oscope. I'm using KT66s into a beefy 6,6k primary OT rated for 70 watts. Bias is set to 69%.

Here are some voltages, heater voltages are all spot-on.
V1: 209, 0, 1.67 / 211, 0, 1.89
V2: 171, 0, 1.2 / 321, 171, 173
V3: 229, 26, 41 / 237, 26, 41
V4 & 5: 435 plates and screens

When swapping power tubes to EL34 I got 17VRMS at the onset of distortion which is more what I would expect.
Is this due to the EL34's higher gm?

Thanks in advance for any ideas/input!

JTM45 schematic attached

[Helmholtz]

130R Rt resistors & dual 1n4407 diodes on PT HV secondary each side.

So you added 130R series resistors at the rectifier plates? If yes, why?
They would lower available output power by several watts.

What is your B+ at full ouput at the onset of clipping?

Does the output clip symmetrically?

EL34s will put out more because they have lower internal plate resistance and more importantly lower plate "saturation" voltage.
Chinese KT66s can have excessive saturation voltage, which limits available plate voltage swing and thus lowers output.
Gm only matters if your PI clips before the output stage with KT66s.

I assume you biased KT66s and EL34 appropriately and used the 8 Ohm output with the 8R load resistor.

[tubby]

Thanks for your input!

So you added 130R series resistors at the rectifier plates? If yes, why?

Most datasheets (e.g. Philips) specify a minimum transformer(?) resistance [Rt] at the plates of the GZ34 depending on voltage and first cap size. For 2x350V and 60qF the datasheet calls for 2x100R. One half of the scondary of the PT HV measured 70R, therefore I added some more resistance (admittedly a very conservative amount).

[QUOTE=Helmholtz;n948339]
They would lower available output power by several watts.
[/QUOTE]
To find out, I shorted out both of those resistors and measured the output again: 14,6 VRMS under the same conditions as already described. For voltages see further down.

[QUOTE=Helmholtz;n948339]
What is your B+ at full ouput at the onset of clipping?
[/QUOTE]
With Rt: 419V
Without Rt: 423V

[QUOTE=Helmholtz;n948339]
Does the output clip symmetrically?
[/QUOTE]
Yes, neg and pos peaks start to flatten out at the same time, in the same manner. Full square wave is symmetrical. If it helps I can post pictures.

[QUOTE=Helmholtz;n948339]
I assume you biased KT66s and EL34 appropriately and used the 8 Ohm output with the 8R load resistor.[/QUOTE]
That is correct. Forgot to mention that earlier, sorry.

[Helmholtz]

Most datasheets (e.g. Philips) specify a minimum transformer(?) resistance [Rt] at the plates of the GZ34 depending on voltage and first cap size. For 2x350V and 60qF the datasheet calls for 2x100R. One half of the scondary of the PT HV measured 70R, therefore I added some more resistance (admittedly a very conservative amount).

Ok, but the reflected primary resistance adds to the secondary DCR. Typically the PT itself provides enough source resistance.

What is your B+ at full ouput at the onset of clipping?
With Rt: 419V
Without Rt: 423V

With a B+ of 419V and a typical saturation voltage of 50V the theoretical output power is 41W. The OT typically eats around 10% or 4W, so we're in the ballpark.
Chinese KT66s can have a saturation voltage >100V. !00V saturation voltage would result in 31W before the OT.

[mozz]

This is what i usually do. Measure output voltage across 4, 8 and 16 ohm loads, calculate wattage, you will find which one it likes better.

[J M Fahey]

I find nothing "broken" in your amp, all is within tolerance.

Published tube amp power rating must always be taken with a grain of salt.

[Dave H]

I remember reading somewhere (I think it was in "Guitarist") that Marshall said the original JTM45 measured 28W
The difference between 28 and 35 W is only 1dB

[g1]

I remember reading somewhere (I think it was in "Guitarist") that Marshall said the original JTM45 measured 28W
The difference between 28 and 35 W is only 1dB

Also there was some back and forth between 6.6K and 8K OT's. I believe 6.6K was the original so that would correspond with the lower figure.

[Helmholtz]

Also there was some back and forth between 6.6K and 8K OT's. I believe 6.6K was the original so that would correspond with the lower figure.

8k (as used with drake OTs) will give even lower output.

Max. output should be achieved with an 8 Ohm load connected to the 16 Ohm output, making Raa = 3.3k.
I'd expect > 50W with EL34s.
But that would no longer be a typical JTM45 (sound).

According to Zollner's measurements, russian Tung Sol KT66s or JJ KT66s should give more output than chinese KT66s because of lower saturation voltage.

[g1]

Max. output should be achieved with an 8 Ohm load connected to the 16 Ohm output, making Raa = 3.3k.

So he should follow mozz test from post #5 then.

[mozz]

So he should follow mozz test from post #5 then.

Well, when i test like that, it just shows me which load gives max power. Best sound? That's subjective. I can usually sneak another watt or 2 by using a 12AT7 P.I. but i always tell them that incase they want to change.

I have had some amps come in best at 6 ohms, 4 and 8 were lower but that shows me not to be worrying too much which speaker i was going to use. I generally only do that wattage test when i only have 1 output tap and it's a unknown scrounged output trans using it with different tubes than it was designed for. Also sometimes i use the big LCR meter and put a resistor on the secondary, say 8 ohms, then hook up the primary at 1khz and see what the reflected impedance is. If that is close to what tubes i planned on using, it's a go.
First JTM45 I built for a guy was a butchered fender head that had basically nothing original left inside. When i was done i measured about 30 watts, having never looked it up I kind of figured it was 45 watts, but they are 30 watts, i was happy, customer was happy.

[Helmholtz]

Output power of a class B PP amp can be calculated from P = 2* (B+ - Vsat)²/Raa , where B+ is measured at full output and the saturation voltage Vsat is the plate voltage where the loadline meets the Vgk = 0 plate curve.
Note that Vsat increases with lower Raa.

Of course the method only works if the tubes correspond to their datasheets and some current production tubes don't.
I also assume that Vsat increases with tube age, thus lowering available output.

Anyway, difference in power output between different tubes in the same amp is most likely due to different Vsat.

Apart from Vsat, actual ouput can be lower than calculated by up to maybe 20% because:

1) The screen voltage drops at high output. Lower screen voltage tends to increase Vsat.
2) Actual Raa can be up to 10% higher than nominal, mostly due to primary and secondary DCRs.


Different plate loads:

From the formula above there is a general tendency of power increasing with lower Raa, at least below and up to the "knee" of the Vgk = 0 plate curve.
(Above the knee Vsat rapidly increases and thus counteracts the power increase.)

But lower Raa also means higher plate currents and dissipation. Also PT and OT currents will be higher at full output.

So the plate load which gives max output may not always be a safe solution.

Apart from power considerations, a different loadline might change the amp's sound and feel.

[loudthud]

I'm confused:

Most datasheets (e.g. Philips) specify a minimum transformer(?) resistance [Rt] at the plates of the GZ34 depending on voltage and first cap size. For 2x350V and 60qF the datasheet calls for 2x100R. One half of the scondary of the PT HV measured 70R, therefore I added some more resistance (admittedly a very conservative amount).

Is this the data sheet you based you design on ? http://tdsl.duncanamps.com/dcigna/tu...ex/5ar4-4h.gif

It seems to me the R for each plate of the GZ34 should be 30 Ohms, not 130 Ohms. When they say 2x100R, they mean 100R per side. Your B+ seems low.

What are your screen resistors ?

[Helmholtz]

It seems to me the R for each plate of the GZ34 should be 30 Ohms, not 130 Ohms. When they say 2x100R, they mean 100R per side. Your B+ seems low.

Where do you see 30R?

I read 2x 100R with 2x 350Vrms or 2x 75R with 2x 300Vrms

BTW, the datasheet shows the formula how to take into account the reflected PT primary resistance.

[loudthud]

In post #3 tubby says he measured 70 Ohms on one side of the secondary. That doesn't take into account the Primary resistance. Never did say what the secondary Voltage was.

100 - 70 = 30