Low output on JTM45 build

[tubby]

With onset of clipping I mean just a tiny portion of flattened tops. Typically goes along with some ripple modulation at tops.
So we know we're just above max. clean output.

Okay, here are the new measurements:

[tubby]

Post #25 shows the output at hard clip. It's a little fuzzy but I read the p-p indicated as 48.24 V. If that is correct, it should be able to do that much clean. That would work out to about 35W.

That was with EL34s and mismatched (halved) load impedance, though.

[g1]

That was with EL34s and mismatched (halved) load impedance, though.

Ah. Can you tell us where the measurements in post #62 are taken? The p-p voltages displayed are all quite different which seems odd.

[tubby]

Ah. Can you tell us where the measurements in post #62 are taken? The p-p voltages displayed are all quite different which seems odd.

Sure! Same measurement points as per post #55. Or easier, if you click on the pictures it should say right below!

[Helmholtz]

Okay, here are the new measurements:

I don't understand the triangular grid signals. Looks like slew rate distortion.
I have seen that effect only with large (2nF) grid to ground caps

Are you sure the cap between the PI plates is actually 47p?

[pdf64]

Waveforms inside an NFB loop don’t look as expected, they amalgamate the input signal and the error correction signal. ie I think the extended sharp peaks at the output valve control grids are the NFBs attempt to correct the anode clipping.
I suspect that if the NFB loop is opened, the control grid waveforms will look more sinusoidal (once gain is normalised).

[Helmholtz]

Waveforms inside an NFB loop don’t look as expected, they amalgamate the input signal and the error correction signal. ie I think the extended sharp peaks at the output valve control grids are the NFBs attempt to correct the anode clipping.
I suspect that if the NFB loop is opened, the control grid waveforms will look more sinusoidal (once gain is normalised).

Some nipples are normal, but not an almost pefect triangular shape with straight slopes.
I just scoped my Vibrolux with similar clipping. No triangular grid signals, i.e. since I removed the 2nF grid caps.
Straight slopes indicate slew rate limiting.
A large value cap between the PI plates could be an explanation.

But would be interesting to see without NFB.

[pdf64]

…I just scoped my Vibrolux with similar clipping. No triangular grid signals ...

The slew rate thing may be a suspect, but I think a Vibrolux will behave differently to a JTM45. Due to the 8k OT, it’ll probably clip at the anodes earlier than the grid. That would give scope for the NFB to attempt to correct the clipping.
Whereas a 4k Vibrolux will clip at the grids before the anodes, so the nipples would get chopped off (ouch!)

[tubby]

But would be interesting to see without NFB.

I have made the NFB loop switchable, so I did a quick measurement without NFB.
Signal drive to the PI was readjusted to compensate for increased gain without NFB.
At about the same level of clipping of the output as per post #61,
the signal at the control grid on one of power tubes looks like this:

At the PI's higher gain side output the signal on the control grid of the other power tube looks similar,
with signs of slight clipping.

No triangular slopes on both tubes' signals, so I'd think the distorted waveform
was due to NFB as mentioned by pdf64 in post #66.

[pdf64]

Did the GT labeled KT66 maybe look something like this?


In that case they'd be of chinese origin (Shuguang, same as the Tube Town label).

Yes, the ‘C’ suffix is a strong indicator of the origin of those.
Mine are a bit different -

[tubby]

Yes, the ‘C’ suffix is a strong indicator of the origin of those.
Mine are a bit different -

Interesting, I've never seen those GT KT66 before. What makes you think they're russian if I may ask?
The only KT66 tubes with getter flashing on the bottom of the glass envelope I know are Shuguang and NOS (GEC, Osram, etc.) tubes.

[Helmholtz]

The slew rate thing may be a suspect, but I think a Vibrolux will behave differently to a JTM45. Due to the 8k OT, it’ll probably clip at the anodes earlier than the grid. That would give scope for the NFB to attempt to correct the clipping.
Whereas a 4k Vibrolux will clip at the grids before the anodes, so the nipples would get chopped off (ouch!)

Right you are.

I didn't think of the 6.6k OT.
So we should see different grid signals with halved secondary load.


Still no clue as to the low power.
Maybe try a pair of new 6L6/5881s for reference?

@tubby: Are we sure the variable scope input attenuator is in calibrated/arrested position?

[pdf64]

I seem to recall it from publicity blurb / articles around their introduction. And there may have later been a snippet or 2 more about them on Myles Roses’ site. Something along the lines that a lot of effort was put into getting them right. Which my experience at least bears out.
I don’t think they were in production for long.
Back then, Chinese output valves weren’t really a serious thing, they had a very poor rep, though some of the 12AX7 were well liked, especially for high gain stuff.

[Helmholtz]

I seem to recall it from publicity blurb / articles around their introduction. And there may have later been a snippet or 2 more about them on Myles Roses’ site. Something along the lines that a lot of effort was put into getting them right. Which my experience at least bears out.
I don’t think they were in production for long.
Back then, Chinese output valves weren’t really a serious thing, they had a very poor rep, though some of the 12AX7 were well liked, especially for high gain stuff.

Maybe early russian made TungSols? No longer available, though.

[Helmholtz]

That was with EL34s and mismatched (halved) load impedance, though.

While I understand what you mean with "mismatched" here, it should be noted that 3.3k (3.2k....3.5K) is the optimum plate load for EL34s, used with all EL34 Marshalls allowing them to put out in excess of 50W per pair (i.e. with SS rectification).
Maybe your Mullards are no longer like new?