Low output on JTM45 build

[Helmholtz]

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

Judging from the DC voltage at idle, secondary voltage must be close to 2x350Vrms.

A typical primary DCR of 10R would add 23R to the secondary DCR.

[tubby]

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.

Interesting, while testing I indeed used a pair of chinese Shuguang KT66. However, I also measured with a pair of JJ KT66 and got the same results in output power.

[tubby]

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.

This is what I did. I expected most power at half the load impedance of the selected output impedance. However, this was not the case, measured output was roughly the same (~ 1 VRMS variation).

[tubby]

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

Very interesting! Seems I succeeded in building a true original then...

It was only after researching output power for the JTM45 on the web I got curious because most amp kits seem to put out 30-35W clean. But then again not many people actually measure their JTM's output and probably go by kit manufacturer's data.

[tubby]

I'm confused:

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 ?

Yes, that's the one. PT HV SEC is 2 x 345V as estimated above by Helmholtz.

As mentioned in my first post I used a "very conservative value of 130R" instead of 100R.
Reasons for that being that I
a) had this value of power resistor on hand
b) expect longer service life out of a GZ34 that way even though I didn't incorporate the standby switching.

Also, as mentioned in post #3 the voltages with or without those resistors don't differ that much.
B+ at the onset of clipping with Rt: 419V, without Rt: 423V.

I dared to omit the shared (1,5k) and the individual (470R) screen grid resistors completely for a test run after rebias with KT66. This resulted in +1 VRMS output compared to the orignal value of barely 14VRMS into 8R load.
Don't think I wann try this with EL34s though!

[Helmholtz]

Interesting, while testing I indeed used a pair of chinese Shuguang KT66. However, I also measured with a pair of JJ KT66 and got the same results in output power.

Saturation voltage is the residual plate to cathode voltage drop at max. plate current. The lower, the more available primary voltage swing.

Do your power stage and supply component values correspond to the original schematic?

What are screen voltages at full output?

[Helmholtz]

This is what I did. I expected most power at half the load impedance of the selected output impedance. However, this was not the case, measured output was roughly the same (~ 1 VRMS variation).

That's strange. I would have expected a significant power increase with Raa = 3.3k.
Zollner measured Pout = 40W at Raa = 3.5k with chinese KT66s and Pout = 48W with TungSol KT66s.

Did you try that with EL34s?

Generally I don't recommend to omit the 470R screen resistors.
Did you see any signs of instability/oscillation?

[loudthud]

Yes, that's the one. PT HV SEC is 2 x 345V as estimated above by Helmholtz.

As mentioned in my first post I used a "very conservative value of 130R" instead of 100R.
Reasons for that being that I
a) had this value of power resistor on hand
b) expect longer service life out of a GZ34 that way even though I didn't incorporate the standby switching.

I can appreciate you wanting to be conservative with today's failure prone GZ34s.

There is a little equation in upper left of the series resistor graph. Rt is the total resistance, you must subtract the secondary and effective primary resistances to get the actual resistor value. In this case it seems like the transformer has nearly 100 Ohms per side all by itself.

Rt = Rs + n^2 * Rp

Rt is the total effective plate resistance
Rs is the resistance of one side of the secondary
n is the turns ratio to one side of secondary ( V sec / V pri )
Rp is the primary resistance

Check the resistors in the screen network. You may have toasted that 1K 2W.
What is the idle current of the power tubes ? From the graph, B+ should be up around 450V with no signal and about 400V with a rail to rail square wave.

[tubby]

Saturation voltage is the residual plate to cathode voltage drop at max. plate current. The lower, the more available primary voltage swing.

Do your power stage and supply component values correspond to the original schematic?

What are screen voltages at full output?

Power stage and supply components are as specified in the the schematic in post #1 with the following exceptions:
- Inserted Rt=130R, 2x1N4007 and fuses as rectifier protection/failsafe per side in the PT HV
- Filter cap nodes for preamp and PI stages are 26 / 33µF instead of 16 / 16µF respectively
- First voltage drop resistor in B+ chain is 8,2k instead of 10k

Here are the screen voltage measurements (onset of clipping / full tilt square wave) : 415 / 325
1k shared screen omitted: 424 / 330
1k shared screen and Rt omitted: 430 / 370

[tubby]

That's strange. I would have expected a significant power increase with Raa = 3.3k.
Zollner measured Pout = 40W at Raa = 3.5k with chinese KT66s and Pout = 48W with TungSol KT66s.

Did you try that with EL34s?

Generally I don't recommend to omit the 470R screen resistors.
Did you see any signs of instability/oscillation?

Here are the output Vrms measurements with NOS Mullard EL34s and load mismatch as suggested (16R output into 8R load).
Bias however at only 50% which was the hottest I could get from the bias range available with that pair of tubes for some reason.
16,1 / 18,2 (onset of clipping / full tilt square wave)
1k shared screen resistors omitted: 16,1 / 19,78 (onset of clipping / full tilt square wave)
The last measurement is now in the range of output power which I would have expected from a stock build with KT66s and shared screen resistor as per schematic.

I didn't see any signs of oscillation. Maybe a slight transformer ringing (?) at full tilt output (see below).

[tubby]

I can appreciate you wanting to be conservative with today's failure prone GZ34s.

There is a little equation in upper left of the series resistor graph. Rt is the total resistance, you must subtract the secondary and effective primary resistances to get the actual resistor value. In this case it seems like the transformer has nearly 100 Ohms per side all by itself.

Rt = Rs + n^2 * Rp

Rt is the total effective plate resistance
Rs is the resistance of one side of the secondary
n is the turns ratio to one side of secondary ( V sec / V pri )
Rp is the primary resistance

Check the resistors in the screen network. You may have toasted that 1K 2W.
What is the idle current of the power tubes ? From the graph, B+ should be up around 450V with no signal and about 400V with a rail to rail square wave.

PT sec half is 63R. V sec / V pri is 345/230= 1,5 and PT prim is 6,2R.
If I'm not mistaken, this equals Rt of ~77R which means I'd need only another 30R per side to be within spec according to GZ34 data sheet.
On the other hand, the measurements taken so far seem to indicate that those 130R are not the single big issue for the lower power output. At least AFAICS.

Screen resistors measure all ok, I used 5W screen resistors throughout.

Idle current is 43mA per tube which equals 70% of plate dissipation at 430V if a plate current reduction of 5.5% for grid current is considered.

Yes, I was also expecting my B+ to higher, more along the values you mention above. On the other hand, I've got zero experience of how accurate these rectifier datasheets are in that respect.
I tried two other GZ34 and got 10-15V lower voltages. Thus I considered a B+ with no signal of 430V instead of 450V as ballpark.

[pdf64]

That's seems a high idle current, mine's fine with them in the low 30s.
KT66 anodes are 25W under the design max system, which equates to 22-23W under the design centre system, ie the system under which EL34 anodes are rated 25W, and makes sense, as it puts it on a par with 5881.

https://frank.pocnet.net/sheets/084/k/KT66_GEC.pdf
https://frank.pocnet.net/sheets/010/e/EL34.pdf
https://frank.pocnet.net/sheets/049/5/5881.pdf
http://www.one-electron.com/Archives...on%20Tubes.pdf

[Helmholtz]

Here are the output Vrms measurements with NOS Mullard EL34s and load mismatch as suggested (16R output into 8R load).
Bias however at only 50% which was the hottest I could get from the bias range available with that pair of tubes for some reason.
16,1 / 18,2 (onset of clipping / full tilt square wave).

So you're getting 32.4W of sine power with EL34s into 3.3k.
That definitely looks wrong. Two good EL34s should give in excess of 50W with both screen and plate voltages above 400V at clipping.

What OT do you use?

What is the primary DCR of your OT?

Is your load resistor actually 8R?

Do you measure output voltage directly at the OT secondary wires?

[tubby]

That's seems a high idle current, mine's fine with them in the low 30s.
KT66 anodes are 25W under the design max system, which equates to 22-23W under the design centre system, ie the system under which EL34 anodes are rated 25W, and makes sense, as it puts it on a par with 5881.

https://frank.pocnet.net/sheets/084/k/KT66_GEC.pdf
https://frank.pocnet.net/sheets/010/e/EL34.pdf
https://frank.pocnet.net/sheets/049/5/5881.pdf
http://www.one-electron.com/Archives...on%20Tubes.pdf

I would agree, normally I'm biasing in the 50 - 65% range. However since this thread is about low power output I wanted to make sure it's not a bias issue with that certain set of tubes.
And from what I understand power output rises with increasing bias current. At least up to a certain point and provided the power supply can keep up.

[tubby]

So you're getting 32.4W of sine power with EL34s into 3.3k.
That definitely looks wrong. Two good EL34s should give in excess of 50W with both screen and plate voltages above 400V at clipping.

What OT do you use?

What is the primary DCR of your OT?

Is your load resistor actually 8R?

Do you measure output voltage directly at the OT secondary wires?

I'm using the 6,6k primary of a dual-primary "Tube Town" JTM45 OT from www.tube-town.net
Specs according to spec sheet:

• Pri1: 8 kOhm / CT
• Pri2: 6,6 kOhm / CT
• Sec: 4-8-16 Ohm
• Power: 70 Watt
• Made in Europe
• Inductance Primary 8 kOhm: 660 Henry (@240V,50Hz open circuit)
• Inductance Primary 6,6 kOhm: 540 Henry (@240V,50Hz open circuit)
• Minimum Reproducible Frequency at max power (70W): 20Hz/-3dB
• Re(a-a) 8 kOhm: 185ohm +/-5%
• Re(a-a) 6,6 kOhm: 165ohm +/-5%

Measuring the primaries' DCR to CT I get 76R resp. 78R so I'd say thats within spec. I have another, beefier Hi-Fi grade 70Wrms OT with a 5,5k primary here, so I hooked that one up for comparison. At onset of clipping that gave me ~14Vrms (as before), full tilt square wave output was 1,5Vrms higher compared to the Tube Town OT. I would explain this with tranny construction and build quality but I'm definitely no expert.

Just double checked my load, it's 8R, cable is about 1m length and is 0,5R. I hook up the scope where the wire coming from the impedance selector is soldered to the speaker output jacks.