Finished and working - 12 valve all DHT PP 26-71a-45

[RhythMick]

Well it was a good effort. All 6 bias trimmers are now out of circuit, top resistors changed to set bias. Interstages mounted back in amp.

Squeals like a stuck pig.

I recognise I'm pretty tired. Stopping now. Will check wiring tomorrow and make sure no mistakes.

Bloody knackering this kind of thing. Did you have the same bias arrangement on the other one with the wings?

Did you decouple the grid bias from the cathode; something like pic below?


26bias.JPG


Cheers,
Stephen

Steve looking again at your sketch you have both a 1uF cap to ground and a 100k resistor which I don't have. Could you help me understand the purpose of these please ?

The cap I can understand, to stabilise the bias against fluctuations - as it's a balanced long-tailed-pair I'd expected there to be little need for that but it's something I can try.

What's the 100k for ? Grid resistor ? Without it, each grid will see a coil and then the lower half of the cathode stack to ground ? Again something I can try easily.

Geoff's amp has neither and is singing beautifully.

I think first course of action will be to bypass the balancer pot with clip leads so the virtual cathodes connect directly to the top resistor. Quick and easy. If that doesn't do it I'll try both the 100k and the cap in the bias.

[izzy wizzy]

I think you still might have an imbalance signal sitting on top of the choke being fed back to the grids which is positive feedback. To rule out this idea, RC decouple the bias to the grids and see what happens.

Or get rid of the balance pot and make it like Geoff's. It's only DC balance anyway and may make it unstable or more different at AC. The balance pot may be causing more imbalance in the tail.

The RC network effectively shorts out the AC feedback while keeping the DC bias adjustment. The balance trimmer maybe causing more imbalance signal to appear in the tail AC wise.

Cheers,
Stephen

[Nick]

High AC impedance to improve the balance

But will it? The inductor in the anode will do that. Electrically it would be the same if you moved the choke to between B+ and the top of the interstage primaries. Two inductors like that in series will have all sorts of resonances. But also, sitting the grid circuit on that (while the valve wont see it as the grid is taken out after the resistor before the choke) it will raise the impedance of the secondary of the driving interstage and grid to your 0v, so allowing all sorts of parasitics to get into the act. Its what Izzy was getting at with the cap after his experience with grids being decoupled from 0v. But in this case will I think even further confuse things as it will allow the inductor in the cathode to be see by the grid.

[RhythMick]

OK latest.

1. I shunted the bias points to ground with 20uF tubecaps. Seemed to have solved the squeal, but one channel down so couldn't fully tell. Once I fixed that (wire not reattached last night) I found that the squeal had turned into a muted audible whine over the music in both channels. We're obviously in the right area, but not quite there yet.

2. I inserted a 100k (cheap little metal films I had handy) between the cathode stack and the bias, following Steve's diagram above. Now have a loud hum both channels. Music also sounds a bit muted, could just be from the loud hum being in the middle of it.

I think I'll try bypassing the balancer pot and see if that helps.

[RhythMick]

No it didn't make any difference.

I think the next step is to make the circuit exactly like the known working version. I suspect there's something else at play though, but let's see what happens.

[RhythMick]

Before I rip it all out, I took out the 100k and used 30uF caps on the bias points. It definitely cures the squeal. and music now playing nicely both channels.

Not out of the woods as I'm left with an audible hum. Audible on the Sony speakers means unacceptable on the Lowther DX4s so must be cured, but I'm wondering if that's unrelated to the squeal I had before. I'm going to check all voltage points again.

[RhythMick]

Just listened to half a CD while measuring voltages. Almost pleasant, except for both a hum and a whine.

Before recording measurements for each valve pair I used the balancer to balance the current through the valves. It did help to reduce the hum, a bit like a humbucker works across a single valve. Whether this is of benefit or not I'll defer judgement on, but the lowest hum point was with the valves in balance. With all pairs in balance the hum was much reduced but left with the whine.

The 30uF caps are clip-leaded and placed outside the amp. Whether the clip leads are contributing i don't know.

One point of note, I found that when touch my multimeter (initially measuring DC volts, but even when off) to the B+ connections (top of the 10R resistor) the note of the whine changed. Different note for each valve. Almost like an instrument, which reinforced what Steve said. I'm tempted to try much larger caps - good idea ? If I try this, what voltage do they need to be rated for ? Obviously no voltage drop across them, but the 45 grids are at 120v so I assume the rating needs to be higher than that ? Correct ?

For interest, the voltages on the valve pairs are...

26 : Vak = 142, Vgk = -11, 5.5mA through each valve
71A : Vak = 123, Vgk = -32, 14.4mA
45: Vak = 220, Vgk = -52, 30mA

Minor variation between the 2 channels.

[Mike H]

Trying again.

This is Geoffs
Screenshot_20181228_082358.jpg

This is mine as it stands now

The attachment 12 valve all DHT PP 26-71a-45 - Page 11 - audio-talk.png is no longer available

That is very close to what a common anode Hartley oscillator looks like, at least if there is some coupling between the tail choke and the input transformer -

Common Drain Hartley oscillator

Common-Drain-Hartley_osc.jpg (8.7 KiB) Viewed 5865 times

Before I rip it all out, I took out the 100k and used 30uF caps on the bias points. It definitely cures the squeal. and music now playing nicely both channels.

Yeah, following on from th above (if true), will stop the cathode(s) feeding positive feedback into the input transformer.

HTH (?)

[izzy wizzy]

This is the sort of thing I was trying to convey but for pp. https://sites.google.com/site/yourtube ... e_bias.gif

Cheers,
Stephen

[Nick]

Just remove the choke in the cathode, its not doing what you think it is.

[Nick]

will stop the cathode(s) feeding positive feedback into the input transformer.

May be missing something, but I don’t see how you can get positive feedback from the cathode to the grid without something adding 180 degrees phase shift.

[RhythMick]

will stop the cathode(s) feeding positive feedback into the input transformer.

May be missing something, but I don’t see how you can get positive feedback from the cathode to the grid without something adding 180 degrees phase shift.

Yes I don't understand it either. And by definition what is positive to one Grid will be negative to the other.

[RhythMick]

Just remove the choke in the cathode, its not doing what you think it is.

The working amp has tail chokes and sounds awesome.

Surely that is exactly what a choke does, provide high AC impedance with relatively low DC resistance. And the higher the impedance in a long tail pair the better the balance. I will at some point try a constant current sink but I don't know enough about designing those yet.

The only difference now between my circuit and Geoffs is the balancer pot. I'm going to rewire to match the working circuit.

[Nick]

Yes I don't understand it either. And by definition what is positive to one Grid will be negative to the other.

That's only true for a signal that is induced in the transformer core, DC shifts of the common point will raise both grids at the same time in the same direction. Current has to flow through the coil to generate flux, the grid ends are in effect disconnected so no current can flow in the secondary.

[Nick]

The working amp has tail chokes and sounds awesome.

Well maybe the layout means there is less coupling between the tail choke and the interstage?

And the higher the impedance in a long tail pair the better the balance.

Yes, if the anode loads are resistive, but if the anode loads are inductive and coupled, they will provide the balancing action.

The only difference now between my circuit and Geoffs is the balancer pot. I'm going to rewire to match the working circuit.

Fair enough, its your amp, so you do as you wish.