0-350V 0.2A Bench Power Supply

[oldway]

Using a linear power Mosfet to build a programmable resistive load and using it as pass transistor in an HV power supply are quite a different thing....In the first case, safety problems are not involved

Your link is only confirming what we all knows: There are real troubles with Power Mosfet forced into linear mode of operation.

[oldway]

What happens with you ?
It's the fourth time that you change the title of this Topic, you can't do this. 

The original title was:
looking-for-yours-opinions!-hv-stabilized-power-supply

Than you changed for
Solved. Tread to delete. maybe

I advised you that you can't delete a thread and loose all the informations whe provide for you.
You changed the title again to :
0-350V 0.2A Bench Power Supply

That was good because this is realy the subject of the topic.

And now, again, a new title :
Power supply gonzo holywar  

This Topic is full of schematics of HV bench power supplies and has a lot of usefull technical informations, you can't destroy all this with a so stupid title who has nothing to do with the subject of this topic. 

I will claim to the moderators.

[Simon]

Lets keep a sensible topic name shall we. Power is indeed to be used responsibly

[001]

I do not understand anything after page #6. 
Some holywar or rocket sience detected 

[Simon]

I do not understand anything after page #6. 
Some holywar or rocket sience detected 

So don't read it or ask for clarification.

[Ian.M]

There's no holy war here, there's just a clear consensus that a robust MOSFET based pass element PSU is going to be difficult and expensive for an amateur without prior high power high voltage MOSFET experience, to design, build and construct as a one-off, hence the various valve proposals.   I think just about everyone who's done their homework is ruling out BJT pass elements due to secondary breakdown SOA limits.    There has been a lively debate on which valve to use and the most economic way to power and bias it.   Give it another week or two and we may yet see a complete design proposal for the PSU.

[oldway]

We are not vacuum tube fools, we are only looking for the simplest and safest solutions, even if they are not modern.
The real solution, both in terms of safety and performance, would be a SMPS with variable output voltage and adjustable current limit.

But it is clear that it requires a great mastery of power electronics to plan and build such a power supply and that it is not within the reach of 001

[T3sl4co1l]

There's no holy war here, there's just a clear consensus that a robust MOSFET based pass element PSU is going to be difficult and expensive for an amateur without prior high power high voltage MOSFET experience, to design, build and construct as a one-off, hence the various valve proposals.   I think just about everyone who's done their homework is ruling out BJT pass elements due to secondary breakdown SOA limits.    There has been a lively debate on which valve to use and the most economic way to power and bias it.   Give it another week or two and we may yet see a complete design proposal for the PSU.

And thus proving the fallacy of the forum: shout louder and more often, and your voice becomes truth.

Well, at least a wise person reading this thread will be able to see reason.

I've stated the facts.  Others have simply continued beyond the point of reason...

Tim

[james_s]

How about a buck regulator? Then you are operating the switching element as a switch, the voltages involved should not be too difficult to deal with.

I still have an ancient Heathkit IP-32 power supply I use for HV work, vacuum tubes really do have some nice advantages when high voltages are involved. The thing must be around 50 years old now and has never had any problems, despite getting plenty of abuse when I was a kid.

[Ian.M]

@Tim, James and anyone else who wants to see the end of tubes,

If there's an easy robust and reliable way for a relative novice to do it with a MOSFET that's in-stock at Digikey or Mouser, that you will guarantee to survive the output being shorted when set for the full 350V, 0.2A the O.P,. requested and already loaded to 100mA, then yes, state the facts again.   *NO-ONE* wants to have a big fragile glass 'bottle' in their PSU, with its heater dissipating an extra 10W one doesn't want, let alone more than one of them, and the two extra transformers required to provide their heater supply and bias.   

I wouldn't go quite as far as Widlarising a working PSU containing tubes - I'd rather sell it off on EBAY - but I'd certainly listen if an easy to build MOSFET based PSU was proposed - no custom transformers or magnetics , no handwaving about the regulator design - just a schematic, and a list of semiconductors required all in stock at a major distributor, and hints about any pitfalls for the unwary, e.g. where poor layout or heatsinking can cause disaster.

[T3sl4co1l]

@Tim, James and anyone else who wants to see the end of tubes,

If there's an easy robust and reliable way for a relative novice to do it with a MOSFET that's in-stock at Digikey or Mouser, that you will guarantee to survive the output being shorted when set for the full 350V, 0.2A the O.P,. requested and already loaded to 100mA, then yes, state the facts again.

I laid out the design process, multiple times.

I think most novices are better at reading...

You also seem to be implying tubes are indestructible, which is an amusing implication.

but I'd certainly listen if an easy to build MOSFET based PSU was proposed - no custom transformers or magnetics , no handwaving about the regulator design - just a schematic, and a list of semiconductors required all in stock at a major distributor, and hints about any pitfalls for the unwary, e.g. where poor layout or heatsinking can cause disaster.

Show me where in this thread a:
- Schematic
- Semiconductor with full SOA
- Hints
was wholly absent.

There is much thread to go through; I'll wait.

Tim

[james_s]

I don't want to see the end of tubes, I quite like tubes, they have a nostalgic charm and for some applications they still have advantages. The extra load and heat presented by the heater(s) is not too big a deal in a *linear* power supply which is already going to dissipate much of the energy going into it as heat. I'm not advocating one way or the other here though, it's certainly possible to do this entirely solid state, though it may not be quite as trivial as it first seems. I don't really understand the source of conflict here, there are any number of ways to build such a power supply with no one definitive "right" way to do it.

Some very old HeNe laser power supplies used a linear current regulator, in those it was common to have a string of several power BJTs in series to get the required voltage. Another possibility is an IGBT, I've only ever used these as switches but they may work ok as a linear pass element, there are certainly some high voltage/high current parts available these days.

[Ian.M]

@Tim, James and anyone else who wants to see the end of tubes,

If there's an easy robust and reliable way for a relative novice to do it with a MOSFET that's in-stock at Digikey or Mouser, that you will guarantee to survive the output being shorted when set for the full 350V, 0.2A the O.P,. requested and already loaded to 100mA, then yes, state the facts again.

I laid out the design process, multiple times.

I think most novices are better at reading...

You also seem to be implying tubes are indestructible, which is an amusing implication.

but I'd certainly listen if an easy to build MOSFET based PSU was proposed - no custom transformers or magnetics , no handwaving about the regulator design - just a schematic, and a list of semiconductors required all in stock at a major distributor, and hints about any pitfalls for the unwary, e.g. where poor layout or heatsinking can cause disaster.

Show me where in this thread a:
- Schematic
- Semiconductor with full SOA
- Hints
was wholly absent.

There is much thread to go through; I'll wait.

Tim

You agreed that some MOFETs have adequate DC S.O.A for the job (e.g FQA10N80C)
You posted a schematic of a switcher that unfortunately was well above novice level (custom wound magnetics).
You posted hints on various pitfalls.

However all of the above were for different PSUs.

Others posted some links to 'ready to build' linear MOSFET HV PSUs
e.g.

I think I found what you are looking for... 

http://hpm-elektronik.de/projekte/nt350-400/hv-nt-350v-schaltplan.jpg

http://hpm-elektronik.de/ng350-0400-netzteil.htm

The SPW20N60S5 MOSFET it uses is listed as obsolete/NRND, 0 stock at Digikey and Mouser - It looks like they are still  available from other reputable suppliers + the manufacturer: Infineon Technologies, so its still in the game,  but Oldway who proposed it said he didn't trust its current limiting. 

Found this the other day https://linearaudio.nl/t-reg-tube-voltage-regulator It's a nice design and the idea is based on the obsolete MC1466L "floating" voltage regulator. The original design used tubes or enhancement mode mosfets, DN2540, and it appeared in the March 2009 edition of Elektor and very recently audioxpress http://www.audioxpress.com/article/t-reg-a-high-voltage-regulator-for-tube-amps. It had a habit of blowing up the enhancement mode mosfets when shorted out but I think the design has been updated for depletion mode mosfets with a much larger SOA. Hmmm... think I've got a spare EL34 somewhere will have to spice it up to see if it's short circuit proof. EDIT: Oops got enhancment and depletion mixed up

Red emphasis added by me.   Worth investigating but that's a pretty large pitfall right there. 

Then at least half the thread went to pot (or rather 'bottle'. You even linked to a 'bottle' design on your own site, albeit 'jokingly') and the remaining solid staters didn't seem to converge on anything a novice would be able to build from.

Several of the diagrams above are already close enough that I don't need to draw one up myself.  A reading of my posts shows this: I have noted that some are going in the wrong direction, while some are very nearly complete.  None have been exactly satisfactory, but there's no accounting for taste.

I'd gladly draw up a proper design -- simulation and BOM included -- if sponsored for it.

We recognise that you are in the design business and expect to make a living, so aren't expecting an actual design from you unless you have cold cash in your hand, but, if you can spare the time for continued participation in this thread, your comments on the only two MOSFET designs that are at a novice buildable level (see quotes in this post), or any other third party design you care to link to, that is at a similar level of buildability, would be most welcome.    I don't think I've missed any other ready to build MOSFET based linear designs - please feel free to correct me if I have.

[oldway]

Hi!

I`m working for a long time using my old tube bench anode supply (0-350V 0.2A max)
But it is so bulky and actualy crap. I sell it to crasy vintage collector today for about $200

Now I see around a lot of cheap power MOSFETs
It looks not so difficult to make HV linear stabiliser.
But I`m not "bicycle inventor"  May be someone was made similar project? Any info highly wanted

Thanx 

P.S.: I do not understand anything after page #6. Some holywar or rocket sience detected

I think it's good to remember the level of knowledge in electronics of 001 .... In fact, the answer to his post was given to him immediately: no, it's not so easy that you believe to do a linear series regulator with Mosfet.

Why make a linear regulator? There is no technical specification other than ajustable from 0 to 350V 0.2A .... Nothing about the how much % regulation, nothing regarding the level of ripple or noise, .....

The HV linear regulator is dangerous if the pass device can fail in short circuit.
A linear regulator is a central heating, an unacceptable loss of energy.

It is obvious that the linear regulation solution is not the right one.

The right solution is a switched mode power supply.

[Ian.M]

I think it's good to remember the level of knowledge in electronics of 001 .... In fact, the answer to his post was given to him immediately: no, it's not so easy that you believe to do a linear series regulator with Mosfet.

Why make a linear regulator? There is no technical specification other than ajustable from 0 to 350V 0.2A .... Nothing about the how much % regulation, nothing regarding the level of ripple or noise, .....

The HV linear regulator is dangerous if the pass device can fail in short circuit.
A linear regulator is a central heating, an unacceptable loss of energy.

It is obvious that the linear regulation solution is not the right one.

The right solution is a switched mode power supply.

IMHO you are 100% right if you add two little words!  At the O.P's experience level, "The right solution is to buy a switched mode power supply."

[oldway]

Not exactly....For 001, the right solution was NOT TO SELL the vacuum tube HV power supply he had....

[Ian.M]

Unfortunately he didn't post here before closing that deal . . . .

[forrestc]

I`m working for a long time using my old tube bench anode supply (0-350V 0.2A max)
But it is so bulky and actualy crap. I sell it to crasy vintage collector today for about $200

I don't see anyone who has pointed you toward this direction, although maybe they've hinted at it:

There are lots and lots of 8-32V in to 45-390V boost converters available for around $5US on ebay, for example, https://www.ebay.com/itm/DC-DC-8-32V-to-45-390V-Step-up-Booster-Module-Power-Voltage-Boost-Converter-ZVS-/263052186020?_trksid=p2349526.m4383.l4275.c10

In the US (and possibly other countries) you can also get them from amazon for around $10, for example https://www.amazon.com/Qianson-Converter-±45V-390V-Adjustable-Capacitor/dp/B01IVMU2XI/ref=sr_1_1?ie=UTF8&qid=1512021716&sr=8-1&keywords=390V+boost

Most of them seem to be rated to 0.2A on the output.   If the output is too noisy there are filtering circuits you could probably add to the output.

I'm sure someone has warned you about the dangers of high voltage.   

[oldway]

High quality bench magic smoke generator ? 

[Ysjoelfir]

I have a few of those linked boost converters. They kind of work. I use them for Nixie and tube projects with a fixed voltage and not very high needs for ripple and noise rejection.
I have also blown up some of those already. They can deliver 200ma, but they can also deliver quite some more - if shorted.
Oh, and some of them require a minimum load.

So... not realy suitable for a bench power supply.

Sadly my knowledge about the quirks and pitfalls in PSU design isn't good enough that I would say that i could build the beginner proof perfect solution...  :/
that is why I thought about making a new thread (since the discussion about valve vs FET distracts quite a bit from the actual goal of this thread) with the mission to build an "eevblog forum "high" voltage power supply" step by step, collecting ideas, working with everyone together to get a good result that is suitable for a beginner and so on. Any comments on that? Could be a quite interesting group colaboration project.

[floobydust]

I think it's proven group collaboration on forums such as this don't do well and rarely finish something when the project is poorly defined, or there is no leadership on the project.
The requirements have to be complete enough and nailed down, and that was not clear here so the solutions were all over the place but it was a good discussion   

Forum format is not much different than the 1980's BBS, a long thread of opinions with free and paid experts, noobs and trolls mashed together to try hammer something out among the flames.
You can start a new thread with "solid-state" added to the title but OP is the one to ultimately build and finish it- as they initiated the thread.
Commanding to be given a finished design and bill-of-materials, as some do - sorry but you'll actually have to do some work. Or instead, purchase the product you need.

Many here know how to make a solution, but community member's time, money, and energy to build and test... compared to spew out some text in a post, there are limits and no real incentive to design and build for free.

I think the forum format needs to evolve to successfully build things, it hasn't really been found on the Internet the best way to crowd source something.
So many open source "let's build X" threads that die and fizzle out. This all should be another discussion, a thread somewhere else.

[oldway]

Generally, the one who needs a bench HV power supply for his hobby and wants to make it himself is someone who makes projects and experiments for vacuum tube amplifiers.
For this purpose, 350V is not enough, it needs up to 450V  output voltage.

So, of course, the HV tube power supply option is the best option for such hobbiests.

The schemas and the realization are simple as demonstrated by the many schematics that I posted.

The only real problem is finding the necessary components, including the transformer.
It is difficult to make an "universal" project because everyone will use the components they already have.

For those who need an HV power supply for professional use, they will not waste their time making it themselves, the only option is the purchase of such a power supply

I remain convinced that the hybrid HV power solution I proposed is the best. (tube pass power stage with op. amplifiers control)

Two ranges of output voltages, manually switched, to reduce the dissipation of the pass tubes, a pass tubes stage, normally blocked by a G1 voltage of -100V and controlled by an optocoupler, voltage regulation and current limiting being made by operational circuits.

As designed, the tube pass stage has intrinsic current limiting which is the result of the characteristic curves of a pentode.
All this makes this power supply very safe

NB: I complete this post by EDIT to avoid multiple posts.

The hybrid solution has several other advantages:
1) the impossibility of overshoot at start-up because the tubes must heat up first.
2) Voltage regulation and current limiting by operational amplifiers allows high accuracy in settings and high stability.
3) this hybrid solution allows to add features like:
- adjustable soft start (switchable).
- Voltage / Current dropping characteristic with adjustable slope to simulate unstabilized power supply.

[chris_leyson]

Thanks Ysjoelfir, fobbydust and oldway for your comments. A linear tube supply is probably the easiest way to go for hobby use. You can still find transformers these days with a variety of secondaries, for example, Oxford Electrical Producs in the UK have some transformers with 240-0-240 and 290-0-290 volt secondaries and a good range of output transformers for tube amps. I already have a transformer and some EL34s but finding a chassis isn't so easy, I'm looking at something like RS Components 754-6017 but can't see how to mount a base plate, their data sheet is less than helpful.

[oldway]

To make the chassis is very simple: a rectangular aluminum plate 3mm thick folded in U.
The size of the aluminum plate should be chosen according to the area needed to install the main components.
Both sides of the U will be 5 to 6 cm high.

The front and rear panels are aluminum plates also 3mm thick and dimensions corresponding to the width of the frame.
The height is to be chosen according to the height of the components.
The case can be made of wood with 3mm grooves to fit the chassis + the front and rear panels.

Provide cutouts in the upper and lower part of the housing for ventilation, with aluminum grid glued.

[Zero999]

I agree that a linear regulator isn't the best idea, as it will dissipates 80W to 100W and the high voltage complicates things but I've seen equally high power dissipating linear regulators discussed here.

There seem to be plenty of MOSFETs available, which can dissipate 250mA at 400V.

The circuit doesn't have to be complex, just use a big, generously overrated, MOSFET. Here's a design with the TL431. On reflection, it could do with some output capacitance (100nF ceramic, in parallel with 10µF electrolytic) and a 15V zener between the gate and source would be a good idea too.

EDIT:
I've just remembered this needs to be adjustable down to 0V! The attached circuit won't work down below 30V, unless R4 and R5 are also changed and even then, the minimum voltage with the TL431 is 2.5V, unless a negative supply is used. It seems a bit of a crazy requirement though. Use a separate 0V to 30V PSU for lower voltages! As mentioned earlier in the thread, semiconductors can fail short circuit, so don't expect the output to be safe to touch, even if it's set to a non-hazardous voltage.