Variable HV supply

[Vicboduk]

Hi all,
I’ve been lucky and found a box of a few NOS valves. Just a couple of each; GZ34, ECC83, EL84, EL34, KT66. This has prompted me into doing something with them. An amplifier is the obvious choice but alas I have no output transformer and I don’t need another amplifier! I do have to hand a mains transformer; 0-200-220-240 primary, 300-0-300 @120mA, 2x 6.3v @2Amp, 1 x 6.3v @1Amp windings. So I thought I’d have a go at a variable HV power supply. Something along the lines of; 50-300vdc @ a few tens of mA. I have a fair bit of experience of designing with silicon (I’ll go and wash my mouth out for using the S word!!) but very little with valves. Any suggestions would be welcome.

Vic

[ppppenguin]

Have a look here for how Solartron made a sensible HT PSU.
https://www.vintage-radio.net/forum/...ight=solartron

EL34 and KT66 seem a bit too precious to use in the rather rough and tumble environment of a bench PSU. I've got a few untested 6080 double triodes, ideal for PSU duty, that you could have for cost of postage. If they're any good you could even make a small donation to the Vintage Wireless Museum.

[Stockden]

Vic,

Somewhere I have a copy of a Practical Electronics design for a variable HT PSU (cunningly named the "Valstab"), which (from memory) uses a 6080 (as offered by Jeffrey) as the series stabiliser. If you'd like a copy let me know & I'll dig it out & scan it.

Hugh

[XTC]

I think you should save all those nice valves for something else, and go for a solid state design. No heaters, so the transformer will have an easier life and you can use the heaters for your experiments, probably better performance, less heat dissipation problems, a smaller and more convenient case, and cases are usually expensive. Cheaper, if you don't regard those valves as free; the replacements almost certainly won't be. Some people think there's something wrong about the idea of transistors working at hundreds of volts.

The 6080 was designed for this sort of application and the KT66, will be pressed into service and is a lot more expensive anyway, so if you must go hollow state, take the 6080 route. Although there certainly were commercially designed stabilised PSUs using EL34s and KT66s. 807 would be better IMHO.

A 6080 has a 2.5A heater requirement,and I'm not sure it's a good idea to parallel the heater windings. There were other valves designed as series regulators. A GZ34 doesn't really fit with that transformer either.

Go solid state on this one.

Pete.

[Vicboduk]

Thanks for the replies gents,

Jeffrey - The Solartron SRS153 is a more straight forward circuit than the heathkit IP-17 that I found (uses 2 x 6L6G's and a 6BH6 as an error amp) so may be better suited for my initial dabblings! Thanks for the offer on the 6080's I may take you up on that.

Hugh - Yes please, The more designs I can find the better idea I'll have of what I'm doing.

Pete - "Go solid state on this one. ", That's fighting talk on this forum
I just fancied a change and wanted to do something with valves - Now you've got me thinking how about a buck converter with valves - What is the switching time for a GZ34??

[ppppenguin]

If you want it regulated rather than just variable I've got some 85A2 voltage ref tubes. If you want them and 6080s then maybe we should meet at the Museum one day.

[Vicboduk]

I'm still undecided what I'd like to do. I'll gather some more info first so that I have a better idea. I intend to make it to the next table top sale at the museum (mid September) - so may well see you there if not before.
Vic

[XTC]

Originally Posted by Vicboduk

Pete - "Go solid state on this one. ", That's fighting talk on this forum

I can't see why. No one's ever seriously suggested building a capacitor reformer with valve rectifiers or that a standards converter is no good unless valve based. These things are utility devices, as is an HV PSU. If you look, you'll see that restoration and nostalgia are far less kind to test equipment, scopes and sig gens, than to receivers and hi-fi. Tragic when you look at some of the brilliant RF sig gens that struggle to fetch £10 and the hi-fi amps which fetch a fortune.

Originally Posted by Vicboduk

What is the switching time for a GZ34??

Who cares? A veteran scope designer dredging the details of a once popular timebase from his memory, and surveying a collection of tatty old scopes, finished off the conversation by saying, "Look ,who cares, better ways of doing things were found. I'm more interested in these PIC things these days .I can't see why anyone wants to waste time on this rubbish now".

Keep the valves for restoring vintage kit, which is nostalgia, or sell them and finance the PSU project, and have some pocket money on top.

Pete.

[Ed_Dinning]

Hi Vic, if you look in the back threads I think the farnell 300v regulated PSU was discussed some time ago, the circuit was shown and I believe I also have a copy of the hand book. It used EL34's and was well designed.

Ed

[glowinganode]

Just to add my two penneth, solid state bridge rectifier across ht winding, ct to earth. Should give you +/- 400v to play with. Zener diode or gas regulator to give you -150v bias at a few mA and reference voltage for error amp. Use KT66 or EL34 as series regulator, you want at least 100v acroos s/r so 300v out is top limit. Output current limited by anode (and screen if triode connected) dissipation of s/r tube. At 50v out you will be limited to about 60mA (20W/350v=57mA). EF80 pentode connected makes good error amp, as has high gain, high anode voltage rating and a short tail. Cathode to earth, g1 to middle of the divider chain. Since the resistor between g1 of the error amp and -150v ref always has 150v across it (give or take), output is propotional to resistor between reg o/p and g1 of error amp. May have trouble acheiving 50v at small output current, as the series regulator will need to be well cut off, leaving only a few volts on the anode of the error amp. If you put a neon between the anode of the error amp ang g1 of the s/r this will ensure you always have a reasonable a-k voltage across the error amp. However this may limit the max output when you have minimum voltage across the s/r a-k.
However....this leaves you with no spare LT windings, so unless you're going to use a seperate LT transformer, an alternative (unstabilised) design would be to use the s/r tube as a cathode follower (should be ok to use the 1A LT winding for this) and a divider chain from the unreg HT to earth, tapping to g1 of s/r. You then have 2A of LT spare, but the HT output voltage will change with varying load.
Sorry to ramble, but have recently designed and built an HT psu using an old "Tektronix" scope transformer, so I've been through all this. Not as easy to design as it first appears. Without more windings on the mains transformer, you are limited to what you can acheive. If you want any more help or circuit suggestion, feel free to PM me. Rob.

[love kt3]

hi ive been thinking about building something simlar,i was looking at the ps-2 and ps-3 heathkit desines(found by google).they use 1619 valves-can be replaced by el34,but you will need another fill xfmr.

[KeithsTV]

Here's a stabilised PSU from the July 1967 PW. Looks as if several alternate valves can be used for the series pass device. It uses a 6X4 as the rectifier which may not be the best one to use. It should be possible to substitute a better rectifier.

The PE Valstab is in the September 1970 issue. If Hugh can't find his copy I can scan it.

Keith

[glowinganode]

Well found Keith, this is what I had in mind in my ramble. Rob.

[Stockden]

Well, as promised I've had a dig in the archive and located not only the Valstab article but also another design from the May 1962 edition of SWM. I've scanned both and, in my usual incompetent way, failed to produce sensible PDFs from them. Rather than delay things further here are the first 5 pages of the Valstab article in JPG. I'll send the other 3 plus the PDF of the SWM article in a 2nd post.

Hugh

[Stockden]

and here's the remainder as promised.

Hugh

P.S. I think I'll start a new thread elsewhere on "How do I produce decent PDFs from a scanner?".

[Stockden]

Originally Posted by KeithsTV

Here's a stabilised PSU from the July 1967 PW. Looks as if several alternate valves can be used for the series pass device.

Keith

Yes I remember it well, I built a PSU from this article at some point in the early 70's. The first version used a single 807 as the series regulator and it worked very well indeed. A later attempt using a a pair of EL84's was a complete failure and I could never decide whether I'd simply made a mistake while building it or there was a fundamental flaw in using EL84's in this way.

I kept the EL84 version for a decade or 2 but, having never quite got round to sorting it out, dismantled it for parts about 6 months ago.

Hugh

[Vicboduk]

Wow! What a response. Thanks one and all for taking the time and trouble. I now have a better understanding of how to go about things. I like the look of the Valstab circuit. The transformer is very similar to what I have although I’m a bit down on LT power. I’m tempted to parallel the two 2Amp windings but Pete(XTC) states that this might be undesirable. I may go with one winding just to start things off. Unfortunately I won’t have anytime this weekend to get things going…
Rob – your “two penneth” worth must make you the best value for money going! A complete circuit in one paragraph

Vic

[XTC]

From what I've read, some transformers are designed so that secondaries can be connected in parallel, but most are not. If the windings were slightly mismatched you could easily have half a volt or so across a low resistance winding. That's certainly losses and heating you don't want. Maybe a low value resistor is a good enough fix? Maybe there are more subtle effects than voltage mismatch which have to be designed out.

I've always understood that it isn't a good idea to parallel similar windings and hope for the best, so it's certainly something to check out.

Something else to watch is that the Heater/Cathode voltage should not be exceeded for the series regulator. It looks as if this has been thought of in the designs above, but it's an easy enough hole to fall into.


Pete.

[kalee20]

As XTC says, don't parallel the windings unless you are sure they are matched. This means having the same number of turns.

You can check this by connecting one end of the secondaries together, and a voltmeter between the other ends. If it reads less than 0.1V then you should be OK.

Many transformers may have deliberately different turns - for instance, if the transformer has two windings at 6.3V 2A, then the winding which is wound first may measure 6.7V off load (dropping to 6.3V at full load). The other winding, wound later, will have a slightly longer length per turn, so greater resistance, so may be have an extra turn or two to compensate - again dropping to 6.3V at its full load.

[Variometer]

Some time ago, I posted a circuit on here that I made for a variable PSU, 0V to almost 300 VDC if I remember correctly. I used a variable transformer from MAPLIN and a normal bridge rectifier (silicon) and smoothing circuit. Very simple, very effective and I find it extremely useful. You can find it by looking through my previous posts. It is on the bottom of the second page!

It is by no means a technological miracle, but it does work very well indeed!

Bob