HT Supply (Test Set)

[kellys_eye]

I've checked the most recent site entries for information on ht supplies and only found references to valve regulated supplies (with various broken links to schematics )

I'd like to build a PSU suitable for various valve projects and am considering modifying an old PC SMPS to provide 6.3V, 12.6V and a variable 100-300V DC output (also perhaps a negative bias voltage ?? ). The ex-PC SMPS idea only comes about because I have dozens of them in a scrap box! Not quite the same situation applies when it comes to basic valve-circuitry type transformers though....

Reconfiguring/rewinding the SMPS transformer should be fairly straightforward but has anyone got a diagram for a reliable ht voltage regulator using hv MOSFETs?

Anyone else ever done this with an old PC power pack?

[Radio Wrangler]

If you vary the duty cycle of the switcher to vary your 100-300v output, then the 6.3 and 12.6 will vary as well.

If you put post regulators on those supplies, then they will drop a lot of volts at a lot of amps when the HT is wound up. However you do it, three supplies need three regulators.

You'll need mains isolation, so your HT arrangement could be one PC SMPS with its transformer giving the usual low voltage high current winding, then a second transformer stepping things back up again. Re-winding a transformer and getting it impregnated to meet the flash voltage requirements is getting rather involved.

Most reg HT supplies used heavy iron 50Hz transformers and then linear regs afterwards. They also used beefy power valves that the audio fraternity lust after, but power Mosfets fix that latter problem. Beware that most Mosfets offer high voltage ratings and high current ratings, but they're made for switching and the power ratings can be disappointing.

The key to designing a good linear bench supply is to have an auxiliary supply running to power the regulator control stages, there's a lot of bad compromises to trying to run everything from the main supply.

For inspiration, I think there's at least one of those huge tomes of collected circuits loose on the web. John Markus was the editor.

David

[AC/HL]

As always, the search box above will find a few threads on this (or any other) subject. Try "HT power supply" for starters.

[Bazz4CQJ]

Could you say a little more about what uses you anticipate for this PSU?

I've been playing with valve equipment for more decades than I want to remember, but I've never owned a 'real' regulated HT supply. The unit which has met the majority of my needs is a very basic 240V transformer supply, and various combinations of resistors and (gas) stabiliser tubes have been attached as and when required.

Although you could probably save money by re-working the computer PSU units, I'd be suspicious about the amount of time and effort you could spend doing so. Have you looked at the "Instructables" website; that seems to be the home of hardcore pc-smps re-workings.

B

[Ed_Dinning]

Hi, have a look at the Sussex valve tester circuit. This uses a variable HT supply; if you do not need constantly variable HT, say from 0 to 300v at 200mA, then use a transformer with tappings so that each "range" is about 50v. The Farnell valve based HT power supplies did this and could be translated to a transistorised circuit. This greatly reduces the dissipation in the series pass element.

A further thing, SMPS units are very noisy unless filtered correctly; so unless you get the filtering right whatever you are powering will "hear" a lot of hash.

Ed

[dave cox]

A variac plus current limited/tripped (but otherwise unregulated) supply might be the answer for designing valve circuits. A full featured HT lab supply is a much more complex beast and probably not all that useful anyway for the projects you have in mind. I would try and avoid anything with SMPS on the bench as they can be rather load dependant ...

dc

[Refugee]

Have you considered using the PCB out of a VGA screen.
They tend to regulate the tube heater supply and often have a second regulator for the HT running from about 150V.
You could rewire the choke in the secondary regulator to get 300V.

[kellys_eye]

OK, thanks for the various feedback.

As mentioned I have many such SMPS boxes looking to be repurposed beyond the usual stabilised supplies workshops might use for general analogue/digital experimentation and would even like to attempt a 12VDC-to-ht unit for mobile use (the FT200 I have has such a device available based on a HEAVY pack! - it would be interesting to see a lightweight replacement).

Time, expense and trouble are not my considerations in any of this - curiosity is the greater driver for me.

Baz makes the right suggestion in that I probably need less in the way of regulation/variation than just a simple source of a basic 150/250 DC supply and external regulation if/when the occasion arises.

Failing all else I will resort to a simpler SMPS solution using a reversed low voltage secondary mains transformer although this means driving it with 12V to begin with! Still, I DO have the PC power packs for that!

I saw a recent article by someone who designed a valve tester that used a simple PIC arrangement for the SMPS side in a reversed-mains transformer design and there were no issues regarding noise.

Once again I thank you all for the feedback and the prompting to think along different lines.

[Diabolical Artificer]

I've recently built a valve tester using the schematic below. I used this for the anode and screen grid supplies using two 250V transformers. These had heater taps for 12V valves. I used a switch to connect two 6.3V heater secondaries in series.

With a simple patch panel and using two DVM's to read anode V and I on a PC chassis. Not elegant but works.

[k_yller]

Originally Posted by kellys_eye

/.../Time, expense and trouble are not my considerations in any of this - curiosity is the greater driver for me./.../

it explain everything
first of all is not big problem to design switched mode psu, reworking of old PC PSU is wrong direction

[Ed_Dinning]

Hi Plenty of app notes on the web for SMPS chips, some even include transformer design data. You could also add a PFC front end, if running from the mains as part of the desigh exercise.
Try Nat Semi, Fairchild, Microchip or Power Integrations for data.
Ed

[kellys_eye]

Dr Wobbles link is certainly useful - thanks!

k_yller - the 'expense' part may have been an exaggeration since I'm known to be 'careful' with my cash . It's just a shame those PC SMPSs are lying there doing nothing useful. Well, one of them is doing the 5/12V duty for analogue/digital work and another is used to power marine electronics equipment (repair work) and LCD monitors. That means TWO whole units have been repurposed! That only leaves about eleventymillion to find a use for..... shame their scrap value is less than zero.

Ed - Yes, I have a lot of the online SMPS design software/links but I see no advantage in re-inventing the wheel. The basic PC SMPS is well established (design wise) and pretty much reliable. Generic circuit diagrams exist to allow the experimenter to find and adjust the feedback circuitry, and voltages more than 10% outside the norm are best achieved by rewinding the secondary of the xfrmr anyway - so if it's coming out, a ht overwind isn't much of a troublesome addition (the original 5/12V secondaries would be rewound with twisted, smaller-csa wire as the desired output current requirement is in no way of the 30A+ order!)

I'll get started on a quick mod and see how it goes, posting any positive (and probably quite a few negative) results for anyone else interested in having a lightweight, multi-purpose lt/ht supply.

[daviddeakin]

I had the same idea of rehashing old ATX (PC) power supplies, but I wanted to avoid having to rewind the transformer, so I have been trying to drive the transformer "backwards", to step up 12V to about 100V, which could be voltage doubled. It is sort of working, but I may need some help! Details here:
http://forum.allaboutcircuits.com/sh...ad.php?t=95301

(It's actually based on a design that appeared in Elektor, which had a hand-wound transfomer and was used for a phono amp, so SMPS don't have to be noisy anymore.)

[woodchips]

What about a simple light dimmer feeding a bridge rectifier, cap, choke, cap arrangement? Simple, cheap, available, more than enough current. The choke based filter is useful for many variable supplies, variac and transformer?

[kellys_eye]

DavidDeakin - rewinding the transformer isn't difficult and is a quick solution to the problems you're having.

Perhaps the hardest part is separating the core halves but I do this by standing it on top of my woodstove for an hour (probably too long but, meh.... it works!) then using the gloves we use to maintain the fire and tugging the halves apart. The core doesn't seem to suffer in any way and the low coil turns numbers mean winding your own secondaries is a quick process.

[daviddeakin]

Originally Posted by kellys_eye

DavidDeakin - rewinding the transformer isn't difficult and is a quick solution to the problems you're having.

It does mean buying two gauges of expensive enameled copper wire, on rolls I'll probably never finish, urg!

[k_yller]

"Everything should be made as simple as possible, but not simpler."

Albert Einstein

converting ATX PSU into high voltage SMPS for tube purposes could be a nice brain excercise - I like this kind of gimmnicks , kind of fallout atmosphere
but cheaper will be to follow "old good ideas"
to tell the truth what You need to build "test PSU" ?
one or two transformers, two or three valves and less then 10 other components and like usually, most expensive part will be case

[Diabolical Artificer]

Couldn't agree more K yller. When I started to think about making a valve tester.HT PSU I looked into making a SEPIC which is a kind of buck/boost convertor. Linear do some great chips that would do the job.

In the end as above I used two transformers and a few other bits and bobs and a variable PSU using a mosfet,as in the link in my previous post. Cost very little but time.

Andy.

[kellys_eye]

Ahhh, though...but.... (copyright some TV show or other) - "The Likely Lads" - Terry Collier in fact

1. I don't have the transformers
2. I don't have the valves
3. I probably DO have the other 10 components
4. I don't like 'expensive' anything

other than that, you described the perfect solution!

On the other hand (I have five fingers - copyright Tommy Cooper)

1. I have the SMPS
2. I have the transformer
3. I have the enamelled copper wire
4. I don't even need 10 other parts!
5. 'Cheap' is a word I can work with

[Bazz4CQJ]

Terry was more laughed at than admired; Jack Ford was more impressive. I'm sure that what you're proposing has been done before, but as mentioned earlier, it's more the sort of thing you'll find on Instructables.com.

B