Mystery amplifier

[PJL]

So both in series gives you a ratio of roughly 6.5/0.3=21.67.
Your 6ohm speaker gives a load impedance of 6*21.67*21.67=2818 ohms, a bit low...
If the speaker is wired across just one of the secondaries we get 6*43.3*43.3=11249, too high leading to early clipping.

[Kokotoni Will]

Originally Posted by PJL

So both in series gives you a ratio of roughly 6.5/0.3=21.67.
Your 6ohm speaker gives a load impedance of 6*21.67*21.67=2818 ohms, a bit low...
If the speaker is wired across just one of the secondaries we get 6*43.3*43.3=11249, too high leading to early clipping.

Judging from your chart, if VA = 290V, which my PSU is capable of, then wouldn't RL equal about 10-10.5K?

[PJL]

The probability is the two secondary windings are interleaved between blocks of primary windings. It might be possible to parallel them but then you would need to change the values of the feedback network. The speaker could be attached to one of the secondaries leaving the feedback network connected across both, but I suspect this would lead to increased distortion.

Maybe someone else can comment?

[(EC8010)]

There's little reason to suppose that the feedback compensation is set correctly even now; published designs in the 1960s specified all manner of output transformers as being suitable, yet it's unlikely they were all the same. "Building Valve Amplifiers" 2nd Ed. describes how to set feedback compensation on P413 onwards.

[Kokotoni Will]

Originally Posted by (EC8010)

1M is too high a value for a volume control because it causes too high an output resistance that interacts with loading capacitance to cause an HF roll-off; 100k is a much better value directly at the input of a valve amplifier. And if it needs to drive a bit of cable, then you need to go down to 10k.

If your hum is due to the RIAA stage, then no need to modify the amplifier. But if the amplifier hums with no input, then you may have to consider the challenge of star earthing. Sadly, it's not guaranteed to eliminate the hum; your hum could be magnetically induced. You really want to eliminate all other possibilities before taking such a radical step.

I'm now using 22k pots and it sounds great. There is a little bass attenuation - I'll test with a stereo 10k pot tomorrow.

The hum seems to be due to the RIAA stage. I'll probably try building a battery-powered phono preamp at some point and see if that helps.

[Kokotoni Will]

The following text and the attached (corrected) diagram are from "Building Valve Amplifiers" 2nd Ed., p413-14:

"The input stage’s anode load RL is shunted by a capacitor C1, which may be in series
with a resistor R1.
The global feedback resistor (Rfbk) is bypassed with a capacitor C2, which may be in
series with a resistor R2."

Which parts of my amp schematic does this diagram correspond with?

[Kokotoni Will]

I'm guessing the upper part of the diagram corresponds to the parts circled in blue in PJL's schematic, but I'm not sure about the lower part.

[trobbins]

Your C6(330pF) +R8(10k) relates to Morgan-Jones' reference diagram R1-C1. However, the assessment of that first-stage step network is based on using a triode, whereas your circuit uses a pentode, so the step/shelf frequencies and magnitudes are likely noticeably different, including that the loading on your anode has the following stages 1M grid-grid, which is also quite different from Morgan-Jones' example (although I only have 3rd Ed).

Your feedback network R14(33k) and C7(100p) relate to Morgan-Jones Rfbk and C2 (R2=0). The use of your C7 may have been made using appropriate assessment of squarewave response to enhance stability margin for resistive loading tests. However that form of tweaking with C7 may not adequately cover stability concerns, especially for no-load and small capacitance only loading scenarios, but was and still is a commonly used technique.

Wrt stability margins, the choice of parallel or series connections for OPT secondary can significantly affect high frequency resonance characteristics, and as such the form and values of feedback compensation could well influence stability margins. That may be a mute point if your speakers are 'benign' at frequencies where instability could arise, or you never disconnect your speakers (accidently or by purpose).

[Kokotoni Will]

Originally Posted by trobbins

Your C6(330pF) +R8(10k) relates to Morgan-Jones' reference diagram R1-C1. However, the assessment of that first-stage step network is based on using a triode, whereas your circuit uses a pentode, so the step/shelf frequencies and magnitudes are likely noticeably different, including that the loading on your anode has the following stages 1M grid-grid, which is also quite different from Morgan-Jones' example (although I only have 3rd Ed).

Your feedback network R14(33k) and C7(100p) relate to Morgan-Jones Rfbk and C2 (R2=0). The use of your C7 may have been made using appropriate assessment of squarewave response to enhance stability margin for resistive loading tests. However that form of tweaking with C7 may not adequately cover stability concerns, especially for no-load and small capacitance only loading scenarios, but was and still is a commonly used technique.

Wrt stability margins, the choice of parallel or series connections for OPT secondary can significantly affect high frequency resonance characteristics, and as such the form and values of feedback compensation could well influence stability margins. That may be a mute point if your speakers are 'benign' at frequencies where instability could arise, or you never disconnect your speakers (accidently or by purpose).

Thanks, trobbins. To me, the amp sounds great, so I'll leave the negative feedback as it is.

[Kokotoni Will]

The other filter capacitor in the same can as C23 failed, so I ended up restuffing the can with two new capacitors. I used a dremel with a mini diamond cutter to open the can, and drilled holes right up against the lugs to feed the new capacitor leads through. I used a cross section of a beer can with some contact adhesive to join the two halves of the can.

[Richardgr]

Dare I mention the new caps are rated for 400Vdc, and the originals were 450Vdc? Maybe there is not an issue with surge in this amplifier.

[Kokotoni Will]

Originally Posted by Richardgr

Dare I mention the new caps are rated for 400Vdc, and the originals were 450Vdc? Maybe there is not an issue with surge in this amplifier.

I know it's not best practice to go lower than the specified voltage, but the PSU valves need to fully warm up before they can deliver 290V to the amp.