audio-talk

Thermionic Idler wrote: ↑Thu Jun 03, 2021 10:19 am

steve s wrote: ↑Wed Jun 02, 2021 12:14 pm I have a decent stock of gz32 and 37s ? Let me know if you are stuck

Thanks Steve - if you have any Mullard coke-bottle GZ32's that might be useful, as I only have the two and no spares - though hopefully they'll last a good long time especially after the 1N4007 mods are done. They seem to do a nice job of running the input stage, where the current demands are that much lower.

No problem dave, I'm away for a few days will sort through them once I'm home

OK - I finally tracked down the post from Lynn Olson about 300B heater voltages that I vaguely recall reading some months back.

From this, I reckon I don't need to target 5 volts exactly. Any range from 4.5V - 5V should be ok, and the resistors I installed a few months ago should cover use with the range of mains voltages encountered in the UK. And I think the same applies to the input valves. It's over-volts that's the problem. I'm leaving the heater supplies alone.

All I'm going to do is remove the PS resistors and replace with either 2x 1N4007's or equivalent that can do 2kV. I nearly said "then I'm done" but you know and I know that further fettling is inevitable.

Or you could just get four of Andrews regulators and forget about it.

True, sadly I don't think there's enough space in the chassis...

Latest schematic - revision 3 showing the diodes to be added (I went for GP10Y's as they had the 1.6kV PIV rating) and the different rectifiers. Also uprated the output stage fusing to the much more common 500mA.

Given the CCS's why C7?

Nick wrote: ↑Sun Jun 13, 2021 1:31 pm Given the CCS's why C7?

Yeah, that's actually a very good question, to which I don't know the answer!

Lynn O. had included a cap in that position on the schematic for the Symmetric Reichert that this was derived from, but had unfortunately not specified its value. I installed an 80uF cap as an "informed guess", after cross referencing to the Karna Kay which was a variation on the same theme, but yeah, it wasn't obvious when doing the Spice model what that cap actually did - taking it out gave me identical results, but I didn't feel confident enough to second-guess the designer.

So I suspect it's redundant, although I suppose it gives me the option of trying the amp without CCS's if I wanted to do such a thing.

I think he has the tail resistor bypassed in the Karna Kay as that stage has resistive loads and is not being fed a truly balanced signal so a very short tail will need a bypass. As you have CCS in the anodes, you shouldn't need the bypass.

He was also fond of the "ultrapath" cap thing and I know some like it. I never heard anything different FWIW the times I tried. I think Mark rated it in his Amity.

Or you could try resistive loads and put the CCS in the tail to force balance.

I'd second the Lehane VCCS for the fils if you can shoehorn them in somewhere but it might upset your beautiful build underneath

SPICE is great but as the models are perfect, we don't see imbalance signals as much as real life. The short tail bypass is there for the imbalance AFAIK. A very long tail or CSS removes the need.

I've had my head deep in LTSpice and PSUD today, in an effort to predict what happened to the operating points after the house move and rectifier change. I figured if I got the model to behave exactly per my measurements last taken on 14th Feb, I should be able to predict the deltas that have occurred since then.

It was easier than physically hauling 40kg of amplifier up and down the stairs...

Suffice to say that my 300B operating points look a little off - the current seems high relative to the plate voltage (72.5mA against 327V). I'd like to try one of the operating points that Thorsten Loesch (and indeed Western Electric themselves) recommend - 350V / 60mA into 4k (half of the 8k p/p transformer). This will give me a 21W plate dissipation and open the door to mesh plate 300B's if I'm feeling rich in the future! Looks like all I need to do is remove the power supply resistors that I added in February and connect a 120R 10/12W resistor in series with the main 500R cathode resistor. If it works, the valves will last longer too.

On to HiFicollective to order a pair of 120R's........

On the input stage, I've decided to increase the PSU resistance a bit - I'll just swap the 68R's from the output stage to the input stage (currently there are 39R resistors there).

The 6V6 operating points won't be affected because of the current sources on them, but it will de-stress the 12SN7 which is currently operating with quite a high plate dissipation - and I really don't think it needs to be that high. Turning the wick down on it to a shade less than 2 watts per plate will mean I can use any 12SN7, rather than specifically the GTA versions.

This will also put less stress on the GZ32 - the steady state peak current pulses will be 66% of maximum, according to PSUD.

EDIT - having done more analysis today, it looks like I still need another 1.2k resistance in the line to the 12SN7's to get them to the 180V / 10mA operating point I'm after. The 80uF cap on the 6V6 cathodes has turned out to be redundant, so how about re-using it to make a third RC smoothing stage for the 12SN7 anodes? Might be overkill but anything that reduces power supply ripple has to be good, right?

Thermionic Idler wrote: ↑Sun Jun 20, 2021 3:04 pm ...anything that reduces power supply ripple has to be good, right?

How about a regulated supply, like the Maida, Dave?

I use one for the B+ supply (12au7 and 300B) in my Transcendent Masterpiece clone HPA (along with a set of Andrews filament supplies) and it is silent - sounds bloody marvellous too.

If you are unsure about hearing a regulator I found that a small (say 22r) resistor after it and a good cap hides it nicely.

Ray P wrote: ↑Sun Jun 20, 2021 7:21 pm

Thermionic Idler wrote: ↑Sun Jun 20, 2021 3:04 pm ...anything that reduces power supply ripple has to be good, right?

How about a regulated supply, like the Maida, Dave?

I use one for the B+ supply (12au7 and 300B) in my Transcendent Masterpiece clone HPA (along with a set of Andrews filament supplies) and it is silent - sounds bloody marvellous too.

Interesting idea, I'll look into that. That should also help isolate the two stages as well. Taking out the redundant 80uF cap will give me some space in the chassis.

Nick wrote: ↑Sun Jun 20, 2021 7:34 pm If you are unsure about hearing a regulator I found that a small (say 22r) resistor after it and a good cap hides it nicely.

I think Lynn used the same trick (although sans cap) with the CCS's - that's why the 820R's are there.

I think Lynn used the same trick (although sans cap) with the CCS's - that's why the 820R's are there.

That may be more of the case that CCS's tend to look like a wire at very high frequencies.

I've just looked up the Maida regulator (I presume we're talking about this one?). The use case is perfect, I need to drop about 25 volts at 20.5mA, so I expect heatsinking to the chassis will be enough.

Nick wrote: ↑Sun Jun 20, 2021 7:34 pm If you are unsure about hearing a regulator I found that a small (say 22r) resistor after it and a good cap hides it nicely.

About what value of cap?