SRPP Series Shunt Reg Anyone?

If they glow, this is the place to be
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

#16

Post by Cressy Snr »

Paul Barker wrote:sorry I wasn't meaning to sound authoritative about whether to use a cap on the SR130 I was just saying I don't have a clue about that.

In my case it doesn't seem necessary.

When it is used on the screen grid of a pentode or as smoothing in a preamp the noise of the VR probably matters more. In a power amp I don't find any noise I can hear.
Cheers Paul,

I'll just suck it and see when I get it built :)
I'll try the lot without the cap.
If there's no noise I'll leave it our altogether.
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

#17

Post by Cressy Snr »

Paul Barker wrote: More than once I have been tempted to look for my box of Zenners. If it was to hand I would definately have taken the short cut before now.
Well I've ordered some zeners just to be going on with; basically to see if the regulators work as I think they ought to.

If I get an improvement in the sound of the amplifiers, then I'll get some S130s ordered next week after the New Year close down.
There's also a nice loktal based 75V VR tube, that'll handle 60mA I've seen... the QS75/60.

I could of course stand the cathode of the bottom valve on a 4K7 25W resistor, with a suitable bypass cap, just to see like.......but that would just be silly and impatient of me :)
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8881
Joined: Mon May 21, 2007 9:42 pm

#18

Post by Paul Barker »

SteveTheShadow wrote:
I could of course stand the cathode of the bottom valve on a 4K7 25W resistor, with a suitable bypass cap, just to see like.......but that would just be silly and impatient of me :)
Ah but then it is no longer a regulator, as it needs a fixed voltage reference either on the cathode or the grid to regulate any change elsewhere. If you use autobias everything will change so regulation won't be tight.

It doesn't matter if the cathode or the grid has the fixed voltage reference the effect is equal. But if you cheat with a little bit of a voltage reference and resistors to make up the potential difference you compromise the tightness of response.

There are two 75C1's ear marked for you. But SR130's could be a problem, what precious few old BR 4 pin regulators I have I would like to keep. The one which is in the video could be lent for your experiments, it is not the gassy one of the two it is the one on the right which works very well. The other as I said is unmarked and stands at 108v so it is probably a 105v regulator. So it isn't what you need. If I have another SR130 and it works I can lend that also. You'll have to get bases though as I have vertually no BR 4 pin bases. The two in the picture took a lot of finding and that is it. Anode is the lone pin cathode is opposite it. Actually I better get to the post office with whatever I can.

Paul
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

#19

Post by Cressy Snr »

Paul Barker wrote:
SteveTheShadow wrote:
I could of course stand the cathode of the bottom valve on a 4K7 25W resistor, with a suitable bypass cap, just to see like.......but that would just be silly and impatient of me :)
Ah but then it is no longer a regulator,
Paul
Yes It was a rather stupid thing to say.

BTW thanks for the help and the loan of the bits.

I have sockets for the Brit tubes as I was experimenting a while back with globe AC/Ps and AC/P1s with an 801A amp I built a couple of years ago now.
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

#20

Post by Cressy Snr »

OK so here is the complete power supply setup.
That 61mA on the 12HG7 branch is wrong so take no notice of it :roll:
I can't be arsed to redo the drawing and upload it again

Image
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

#21

Post by Cressy Snr »

SteveTheShadow wrote: I have sockets for the Brit tubes as I was experimenting a while back with globe AC/Ps and AC/P1s with an 801A amp I built a couple of years ago now.
And here it is:
Image
Took some finding but I dug a picture out for old times sake.
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8881
Joined: Mon May 21, 2007 9:42 pm

#22

Post by Paul Barker »

Huston we have a problem.

I tested the one s130 in the video and to my absolute surprise it is holding 190v. Then I found one other popped it in place of first and it holds 90v.

So what I am now surmising is that also the unmarked BR 4 pin balloon VR's I have are also probably s130's.

As all these offer widely different volatges. I am left supposing that either VR technology of that era was very slack, or that these have all gone soft and are as much use as a voltage reference as a one legged man in a bum kicking contest. So I haven't sent any.

My opinion on VR's which I have formed over quite a few years using them is, that they are fairly loose anyway in their operation, but they can be relied on to hold a discovered by test voltage, not necessarily on the same page as the number printed on the side, at a known current. Vary the current and they will shift by single figures within a 5 volt variation.

I have found the tightest control of any VR only to come from the 75c1 which is the only one that glows orange. Obviously there is a different neon gas in these. I have sent you two with bases.

I think that your circuit would best be adapted to suit a single 75C1 on the cathode of the shunt section, and the position of the wiper on the voltage sense network altered.

When you think about it the shunt half of the 6080 has only it's own current to worry about so it can handle more voltage. At the maximum voltage of 250v it can handle 55mA to remain within it's limits. With your desired 320v B+ you have 278v at the anode of the shunt valve so with a 75v reference (which will porobably actually be 78v) your anode voltage is 200v at which you can pull 65mA and remain within 13watts for that section.

I was anyway already starting to wander why you had such a high voltage reference.

There will be no effect on the series pass element of this change.

I am also uneasy about the ripple voltage on the VR ref. I know that I instigated this thought process. What I really meant was, once you decide to use these methods reduce your reliance on the capacitor method to the point where the ripple is satisfactory as a whole. In one of my power supplies there are no capacitors there are only two stages of VR referenced Shunt valves and a lot of inductance. This power supply you are making is breaking new ground. But I do not sustpect it needs very much capacitance prior to it. you may be able to go back to venerable old ideas of a pi filter made of 8µF then choke then another 8µF or 16µF cap in good old paper in oil. don't get obsessed with psud or the resonant frequency of the supply. In my experience the venerable old designs even if due to resonant frequency bloom a little in bass, sound more organic than the "perfect" design. Less capacitance ion a valve amp always results in a better organic sound especially in voicing. The shunt regulation does not take from this. Shunt regulation sounds best of all, but I expect that preceded by an organic vintage sounding supply it will be better than preceded by a modern sounding supply.

If you go back through the years on forums you will see that early on these things were recognised, but not appropriated by the majority and mostly the voices who have experienced these things havce been silenced. It doesn't mean they were wrong.

I recall a day when Darren first came here. I built him an amp in my conservatory from scratch from my head during that day. He went home with a great sounding amp. the power supply which was of that old style, was tuned in entriely by listening. We tried all the paper in oil caps I had a one particular mix sounded out of the ordinary. It was clearly a much better sound than all the others. If memory serves me right it was better than a typical modern supply with Black Gates. I let Darren have those caps. they may well be the ones in his px4 amp today.

The reason for this tale is to show my personal experiences from which I form the opinions stated. This type of organic power supply formation makes a better overal sound.

In my most recent experience the all shunt regulated power supply far surpasess anything else I have done. It leaves black Gates for dead.

If I had a task to make something quite good in a hurry I could rely on the Black Gates to perform. I could also get a good sound from Paper in Oil. It wouldn't be as clean sounding as the Black Gates and it wouldn't have the levels of sustain in the bass resopnse that the Black Gates allow. But it would be a more organic sound which you would neither prefer nor see as inferior to Black Gates.

Why this matters to you is that one thing I know. Unless your capacitors sound like black Gates or the best of paper in oils (some are nasty, and the only way to know is to buy many and try all.) then you want as few of them as you can get bye with. Because capacitors influence sound greater than everything else.
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8881
Joined: Mon May 21, 2007 9:42 pm

#23

Post by Paul Barker »

When i was building this little Darling test bed I re-read the Darling Fest article from Sound Practices, because so ltitle has been published about the valve since then.

The very significant point Jeremy draws upon was that in his two stage design retaining the same node in the power supply for the output valve and the VA/driver sounded best.

I too have done this on my little Darling, it is a 1626 driving a 1626 sharing the same power supply and the same cathode resistor. If I had two chanels they would all four valves share one cathode resistor. I have done this since my 112a days. There I had four 112a's sharing one resistor. This gets your cathode resistance down so low that it has a tolerable effect on valve impedance so as not to require a nast capacitor across it, which always sounds horrible in comparison to unbypassed resistor. But that is an aside.

Studying your supply most of your humungous values of capacitance is to smooth the driver. I wander if it is not worth trying to take the driver supply from the series/shunt reg portion and backing off on most of those pie stages. I would just try 8 L 16 if you have the capacitors.

Read what Jeremy says.

Basically you get cancellation because each valve that shares the supply is out of phase.

The reasons that power supplies became so complexed was when pentodes feedback and push pull became prevalent. For various reasons mostly motorboating, decoupling between multiple stages was essential.

In simple few stage Se designs without feedback older technology is a better match. The constraints which created the need for elaborate pi filter networks no longer exist. Old habbits die hard.
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8881
Joined: Mon May 21, 2007 9:42 pm

#24

Post by Paul Barker »

Looking at your supply there is nothing to lose trying the 4 then choke the 100 ditch the other caps move the VA take off to the output of the reg and possibly leave the three series resistances in place if they are required to arrive at the correct voltage or if that can be eliminated leaving the series element to control the volt drop better still.

You may even find that after the choke you could go straight to the regulator. Brave I know but worth a try.

The totempole series shunt regulator is in itself a great innovation which may be all that is required.

Maybe it would be better in the event (in spight of the cleverness of what you have come up with) to sepetrate the two sections. Control the series pass element with an EF86 in pentode mode or ECC83, then place a choke before the shunt element.

In any case it may be advised (without me looking up the data) to use two 6080's due to the heater cathode potential differences. OR looking at data you may be able to voltage reference a suitable floating point for the heater supply.

What you are working on is great. But it could be (short of a fully shunt regulated supply) the greatest.
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

#25

Post by Cressy Snr »

Paul Barker wrote: I think that your circuit would best be adapted to suit a single 75C1 on the cathode of the shunt section, and the position of the wiper on the voltage sense network altered.

When you think about it the shunt half of the 6080 has only it's own current to worry about so it can handle more voltage. At the maximum voltage of 250v it can handle 55mA to remain within it's limits. With your desired 320v B+ you have 278v at the anode of the shunt valve so with a 75v reference (which will porobably actually be 78v) your anode voltage is 200v at which you can pull 65mA and remain within 13watts for that section.
You know I really like this evolving approach to designs.
When you mention things like the single 75V VR tube on the bottom of the shunt valve I ask myself "why the hell did I not think of that?" It simplifies the thing further and puts the SRPP in a situation where it is performing "conventionally" in terms of having about half the HT being presented to each valve.

I'm usually good at seeing the big picture but in this instance managed to get hung up on details that were not worth bothering about.

I'll have a redesign and post it up for review later tonight. :)
I'm enjoying this process.
I'll have a look at the hk potentials and see if I can come up with a floating point for the heater. The 6080 can stand 300V so it ought to be reasonably simple to raise the heater to somewhere where both sections would be happy.

Re this totem pole style reg, surely somebody must have thought of it somewhere in the past, they usually have in this game. I mean not everyone can come up with the Aikido.

Steve :)
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8881
Joined: Mon May 21, 2007 9:42 pm

#26

Post by Paul Barker »

SteveTheShadow wrote: puts the SRPP in a situation where it is performing "conventionally" in terms of having about half the HT being presented to each valve.
Actually it isn't convensional because the current through the series pass element contains the shunt current and the amplifier demand, whereas the shunt element contains the shunt current.

Not being pedantic but a successful SRPP is a very complicated beast most of the so called SRPP's seen ins chematics are not liekly to poerform at all well. usually a different resistor is required to make an SRPP work as nature intended. It is a mythical and complicated valve stage. Best left to copying of venerable examples closely and not just thrpowing together any old thing. Take the one from Ongaku or from Loftin and White copy them exactly or don't call your design an SRPP at all.

There is a whole article in Glass audio which honours this vbenerable stage design. After having read that article a few times I couldn't design you one, but I came away in the knowledge that there is more going on in a successful SRPP than I can explain to you. You would have to read the article. in a nutshell there is a fine balance which is perfect and everything else which you see thrown together all over the web is nothing of the kind. this is probably the reason it was mothballed as a stage. Anyway there have been better ideas since. No your thing is in no way shape or form a perfect SRPP. It is an anathema by very nature of having two entirely different currents to handle.

What it actually is, is a neat package which contains a series pass and a shunt regulator which shares one role (the shunt reg acts as the error amplifier for the series pass). the cleverness of it is only the neatness of the package. A proper series pass reg would do that job better. the shunt reg would be a good one anyway and well worth retaining.


SteveTheShadow wrote: Re this totem pole style reg, surely somebody must have thought of it somewhere in the past, they usually have in this game. I mean not everyone can come up with the Aikido.

Steve :)
It's probably quite new, mainly because hardly anyone has thought of series and shunt regulating. hardly anyone regulates anyway these days.

Had it been thought of. (Yes I thought of doing both. For a while I had in my all shunt reg supply a 13e1 series element, but found it unnecesary) Yes someone somewhere thought of it maybe. but have they done it and posted a schematic? NO!
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8881
Joined: Mon May 21, 2007 9:42 pm

#27

Post by Paul Barker »

I suppose that SRPP article is now to be found scanned and emailed :lol: just so you know I am not Billy Bullshit. What I know that I don't know is more than some people think they know. But I know I don't know much at all. What I do know has told me there is a lot more that I don't know than I thought before I knew as much as I now know. The less people know the more they think they know and in their opinion what they know is all there is to know.
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

#28

Post by Cressy Snr »

As you say Paul, a proper series reg, which I already have, followed by a shunt reg will probably perform better than the totem pole circuit but I just
have to try it now.....one never knows.

I'm going to have to be careful though that I don't alter too many things at once. I'll do the pair of regs first, see how it sounds.

If it sounds better, I'll then start on the caps in the smoothing filter and see how that pans out with the driver fed from the regulated end of the supply.

I've got all the appropriate sockets on the chassis so each change is reversible if it doesn't work out.
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8881
Joined: Mon May 21, 2007 9:42 pm

#29

Post by Paul Barker »

Yes.
See PM for article. I will take the pictures off photobucket when you have them.

In the conclusion it states

"
Despight the apparent simplicity the SRPP presents hidden qualities not fully understoo."

This was said in 2000. I took it as read. At the time I read the article I was not able to comprehend it. If I had the interest I could grasp it now, but it is very long winded and writen in the style of an agrophobic autistic university lecturer with little hand hold on the mind of the normal person.
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8881
Joined: Mon May 21, 2007 9:42 pm

#30

Post by Paul Barker »

The basic problem of the amateur SRPP is demostrated in this statement in the article.

"furnish the solution to a nonlinear problem starting from a linear assumption. In fact the upper triode receives an input signal already corrupted in it's harmonic content by the lower tube; therefore you don't necessarily obtain the minimum THD by differentiating two out of phase currents with the same amplitude."

You brother have no hope of calling your circuit an SRPP. It is dead in the water.

What the writer of the article then proceeds to do is design a 6sn7 srpp which he optimises the blend of harmonics to what he measures as optimum. There is a different optimum for each input voltage. such an excercise would have to be performed for every different valve for every different signal amplitude for every different supply voltage.

The SRPP is best left to the professionals.

Hence the shift to Fets and pentodes for the top valve and triodes or even dht's for the bottom.

BUT the modern trend is not seeing the beauty of the correctly designed SRPP. Done right it is simple easy and excellent. SO as I said before copy a well known quality example or study the article and start using the tools he demostraints to blend your own harmonics to perfection. He even talks about altering the harmonic blend of a stage to suit the overall.

This is exactly what Kondo di with the SRPP. He employed it as one of his harmonic blends. Ongaku is packed with entirely differing blends which Kondo chose to suit his transformers and nobody elses. He was a lot cleverer than anyone who thinks they can copy his design and put it together with different transformers etc and expect it to sound the same.

But they won't have a bad amplifier. They will have yet another also ran.
Post Reply