50 se shuntregulated supply

[vegard]

I have been working on a 801a se a2 amplifier during fall and winter.
When searching for 801a amps on www I stumbled upon this page, and
coincidentally they have shunt regulated supplies as well (!?)

some minor faults in schematic. Initially buildt with lclc shuntreg, because I wanted to compare lclc and lcshuntreg..
One lc was skipped as I do prefer the lc shuntreg supply.
Also z729 was replaced with 841, the latter sonically superior.Both rather quiet about 110 uV on output

attached is schematic, spectrums as amp enters a2, and spectrum at clipping upper trace is g1 of 801 lower is output.
also attached a preliminary schematic of the 50 (have some questions about regulators , maybe this is the right place too ask..

All this led into a desire to uild an it coupled all dht amp (50 tubes)
I would like to try shuntregulated supplies without erroramplifier:
suggestions ?

[vegard]

50 it

[andrew Ivimey]

very impressive indeed!

[IslandPink]

this led into a desire to uild an it coupled all dht amp (50 tubes)
I would like to try shuntregulated supplies without erroramplifier:
suggestions ?

Hi
Nice amp , good taste in valves . Do you have NOS 50's then ? ( if so you are in a select group ! ) .

You talk about shunt reg without error amplifier. What kind of circuit do you have as an example ? It would seem that without the error amp, the Zout would be that of the shunt valve anode resistance - so quite high . Is this amp used full-range ?

[vegard]

Hi
Thanks for positive feedback.
It is often mentioned as one of the major advantages of many dht's
That one can make amps without feedback .
The one I show has one positive feedbackloop backwards from the cathodefollower, and one negative in the psu .
As you can see the shuntregulator in the psu is borrowed from Steve Bensch.
I wonder if it is possible and if so , beneficial soundvise to make the psu for the 50 se with a shuntreg as one of the elements , but without the error amp?
The amp is supposed to be used fullrange.
Yes obviously zout will be higher from the psu, remembering ra of the 50 is
About 1800 how "high" zout from psu allowed before benefits of a shuntregulator vanishes?
Yes I was lucky to get some 50's from a norwegian dealer about to retire.
Could not resist them.
Datasheet says about 50 ma at 450 v , but Jack Elliano/Electraprint told me that according to his experience one should up the current to 70-80 ma.
As the bias supply of the 801a amp is variable : I swapped in a used 50
Running the 50 for several hours at 75 ma without signs of redplates or faillure.
Maybe the shuntreg without errorampm"path" leads astray.But as I am curlous...
I have some 10's, some 10y's and a quad of the tango interstages.gathered over time: better to use these.
I will have the mains , chokes custom made

[IslandPink]

Ah, I had missed the circuit last night - the situation becomes more clear.
I had a good look at the main HT p.s. and it's nice. I like the use of the Pimm CCS feeding the shunt section - hadn't thought of doing that for a power amp ( in my case the main PS needs to supply 250-300mA ! ) but have used simple PimmCCS / VR tube shunt on my phono for many years.
The neg.supply for the 12AX7's is a good idea, to get the EL519 cathode at ground and avoid by-pass caps. Good old Steve/Stephie Bench !
The only thing I'm confused about is the 120V screen supply which also seems to feed the error amp and go to ground .... maybe I will look at 'Steve's tube pages' for the circuit .

I'm not sure what sort of Zout value from the PS is really a problem, but I wouldn't want to go much beyond 20-30 R myself , even if it was constant with frequency . My concern is that signal demand is going to cause 'bounce' of the supply if it has higher Zout ; and the 'bounce' feeds into the sub-supplies for the driver stages , to a small extent , in an out-of-phase way that may not be helpful .
I don't believe the error amp in a shunt reg. is as audible as it is in a Series-reg supply . It's likely the coupling cap quality is important though - I hope to find out more when I build mine soon .

Anyway, I would not stop you from trying what you suggest !

Ps. I envy your 50's .

[IslandPink]

Having looked at Steve's tube pages, I understand the 801SE shunt supply much better .
So, when you say, you are going to try the supply 'without error amp' are you just going to use the simple circuit with VR tube (s) and the EL509 , which has 40R output Z ?

One thing I would say, if you haven't already tried , is to put a small cap or a small cap/resistor Zobel across the OA2's feeding the screen & error amp . I know that in my phono stage , this makes a noticeable improvement to the HF transparency . I think I use about 300R and 0.1uF in parallel with a pair of ( stacked ) OD3's in that unit .

[vegard]

Stephies shunt reg is a nice performer!
I am curious about the audibility of the erroramp.
I guess I have got "hooked up" on psu for se amps, they seem to be just as or more important than the rest of the amp..

I do not why I did not solder in the 0,1uF caps , I know they should be there.
easy fix though.. Shall have a look at the zobel suggestion as well.

I am thinking lc- series reg (tube)-shunt reg without erroramp. Depending on the pass tube what zout from psu.

What are your experiences with separate supplies for input/driver/outputstage ?

[andrew Ivimey]

'What are your experiences with separate supplies for input/driver/output stage ?'

groan. The horror is that, in my experience, it makes a difference, it sounds better but how big should an amplifier be? Having done the basic SE amp with a shared PSu and built out and out and out with the power supplies I now want to work backwards and find where it isn't so different, where it doesn't matter so much.

I note you use Rod Coleman's modules. There seems to be something magic about these modules, for example, and yet I still want to use ac heaters ... (I still do with vt4cs) because simple is best - only maybe it isn't!

[IslandPink]

Stephies shunt reg is a nice performer!
I am curious about the audibility of the erroramp.
I guess I have got "hooked up" on psu for se amps, they seem to be just as or more important than the rest of the amp..

Yes, agreed !

I do not why I did not solder in the 0,1uF caps , I know they should be there. Easy fix though.. Shall have a look at the zobel suggestion as well.

Should help , these two .
I have ( strangely ) been discussing this PS with another guy, on e-mail this evening . It does seem to me there is potential for a significant amount of white noise in the PS as a result of the VR-tube supply to the error amp being amplified . Have you 'scoped' the PS output ? - Would be interested to know if this problem exists .

I am thinking lc- series reg (tube)-shunt reg without erroramp. Depending on the pass tube what zout from psu.
What are your experiences with separate supplies for input/driver/outputstage ?

I have never gone that far - as Andrew probably - space, heat , cost !

I am hoping a really good shunt supply should make the separate supply option less important - which is why I want to get down to eg.2mV for the PS noise/ripple, and I can supply the driver and input valves direct ... hopefully !
Certainly I have heard some very good results from separate PS's . One Paul Barker ( on our forum here ) built had 212 output valves, but the 211 drivers had a shunt-reg supply . Very very nice.

[Paul Barker]

If you are talking about the first Eggfest it was a resistorless capacitorless power supply modled on Stephies. the output valve was the PX25 and I demonstrated the improvements in sound of output valves from px4 shoulder then px25 shoulder and finally px25 balloon. The px25 balloons two broke that day and another back home so the pair which remain are kept in cotton wool in a drawer. Only Steve remains bold enough to use them. He must have more than just one pair.

I can't honestly remember what the 212 power supply was on that day but shure it was monoblock. But the 212 amp people heard at my home a few years prior was it's superior. Darn garned direct coupled 6em7 Cathode Follower driver for goodness sakes beat all my further efforts for years until I gave up alltogether for many years trying to chase an impossible target. But if you want power the 212 can be tamed. But it is a tricky beast to get the best out of. however a pair of 211's in parallel which electrically are about the same, do not sound as good. Neither does the GM70.

Jono kind of put paid to it that day with his lovely 45 amp and a private chat in which he said about my 212 efforts "is it worth it?" No it wasn't. I go for lower powered targets these days.

[Paul Barker]

Never heard a 50, or touched one. Would be nice to. but 801a's are as close as I'll get. Those are nice enough for me anyway. I don't need the ultimate these days.

[vegard]

It would be very nice to keep the psu on the "smallish side"
Without compromissing sound.
The shunt reg /psu for the 801a is rather quiet:
See picture.
"I have ( strangely ) been discussing this PS with another guy, on e-mail this evening . It does seem to me there is potential for a significant amount of white noise in the PS as a result of the VR-tube supply to the error amp being amplified . Have you 'scoped' the PS output ? - Would be interested to know if this problem exists "
I have the scope on the picture, but I am not that experienced using it.(but I am very eager to learn)
If you can guide me on the settings I'll check.

[IslandPink]

I'm assuming this is a probe of the power supply output ?

Hmm...
It's difficult to see enough for me to say what the numbers are on there , maybe someone else is familiar with that instrument . I'not quite sure how the three sets of numbers on the 'set level' scale work .
The waveform trace looks like there's quite a lot of sine-wave hum , because there's none of the tell-tale 'fuzz' on the trace that you get when you are getting down into the single-mV range and the general white/RF noise starts to appear .

[vegard]

The distortionanalyzer is set in voltage mode an shows just below 100 microvolt ac on speaker terminals.

The scoope trace is 1khz sinus and 7,2v at output terminals.

Sorry if I was unclear : to evalute the white noise you mentioned
What settings shall I use on the scoope