0-350V 0.2A Bench Power Supply

[001]

ATTENTION!

THIS TREAD IS FULL OF HOLYWARS AND FREAKS!

My conclusion:
1 - NO WAY to buid ruged HV bench power supply
2 - any VARIABLE  HV power supply is piece of art
3 - Nowday PS is not affordable for HV output mods

[Ian.M]

Most power MOSFETs are not specified for operation in their linear region and will fail at a small percentage of their nominal rating.  The situation with BJTs is somewhat better, (at least published SOA data with a DC line is more common) but even so it will be extremely difficult to find devices that can operate in their linear region dropping >350V @ 200mA reliably.       

[Zero999]

Don't make the transistors drop the whole voltage. Use a tap switcher or pre-regulator to reduce the voltage to something more sensible, before the pass transistors.

[001]

Thanx a lot!

High voltage MOSFETs now very common and cheap http://www.shaoguang.com.cn/pdf1/mosfet/FQA10N80C.pdf

The question not about pass transistor. But about sense & regulation circuit

[Zero999]

Thanx a lot!

High voltage MOSFETs now very common and cheap http://www.shaoguang.com.cn/pdf1/mosfet/FQA10N80C.pdf

The question not about pass transistor. But about sense & regulation circuit

You seem to have misunderstood Ian's post.

Yes, plenty of cheap, high voltage MOSFETs are available. The trouble is, they're typically not rated for linear operation, only switching. This means that they cannot withstand the full voltage and current rating for very long before blowing up.

[Ian.M]

Don't make the transistors drop the whole voltage. Use a tap switcher or pre-regulator to reduce the voltage to something more sensible, before the pass transistors.

Obviously that helps with the dissipation issue, but the pass transistor(s) still have to be able to drop the full voltage at the maximum current for long enough for the preregulator to respond and drop its output voltage or they'll probably blow if the output is ever shorted.

[001]

Why so many words about pass transistor?
Is  MOSFET rated to 800V 10A 240W is not ok for 350V 0.2A linear power supply?  Short current limitter is not difficult thing anyway
It is more intersting how to minimise ripple voltage, what type of NFB is ok, etc

[mk_]

Why so many words about pass transistor?
Is  MOSFET rated to 800V 10A 240W is not ok for 350V 0.2A linear power supply?  Short current limitter is not difficult thing anyway

Search for hotspots on the die when FETs optimised for switching are used as linear pass-element. Search for a pdf called "New 500V Linear MOSFETs
for a 120 kW Active Load"

Take a look at the IXYS LinearL2 TM Series or any other one with Extended FBSOA. This Extended FBSOA is usualy a sign that these FETs can be used linear too.

It is more intersting how to minimise ripple voltage, what type of NFB is ok, etc

short detection should be fast (I mean really fast, at least in the 10us-region) but thats easy...

good luck

michael

[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. 
....

Perhaps not the cleverest thing you did today. 
Using vacuum tube as series regulator is very safe, safer than Mosfet.

[mk_]

Thanx A Lot!

I see what in 350V 0.2A range where are many affordable mosfets
So it is not a problem

What You can say about this diagram?

Try it, there is no regulation, just some kind of setting...

Anyway, it should work but is not something I would call "HV bench power supply".

michael

[oldway]

Even modern design HV power supply as Power designs 1547 still use tubes as seies regulator.

[oldway]

Some other HV power supplies schematics....

[001]

You misunderstand me
Power designs 1547 is a 2500V supply from 1970th
but I need no more 350V 0.2A
is it problem for modern mosfets? 

[001]

And someone made HV differencial error amp 20 years ago (pass transistor is MOSFET):

[Ian.M]

You try finding those HV MOSFETs from 20-30 years ago in stock at a reputable distributor.  The same goes for high voltage BJTs that used to be used as CRT drivers.  First it was increasing integration leading to specialised HV linear driver chips, then as flat panel prces dropped, with the death of the CRT TV, most high voltage linear discrete semiconductors became obsolete and were discontinued.

There are still a few parts around that will make you life easier e.g. Microchip's (ex Supertex) LR8K4 400V three terminal adjustable regulator,  but its only good for a few mA so you'll still need a pass device that can handle 400V at 0.2A without hot-spotting and failing.   The easy option there is still a valve cathode follower.

[jmelson]

Why so many words about pass transistor?
Is  MOSFET rated to 800V 10A 240W is not ok for 350V 0.2A linear power supply?

No, no no!  It is called safe operating area.  Most MOSFETS are designed for switching operation, and are intended to be either fully ON, or fully OFF.  If operated in the linear region (between on and off) they do not share current evenly within the transistor's area, and the hot spot burns out.  There ARE MOSFETS designed for linear operation, mostly in audio amplifiers.  But, they are expensive, special parts, and not likely at all to be rated for 350 V.  (You'd expect to find them with 100 - 150 V ratings.)

Check the data sheets for the SOA rating of any FET you plan to use.  If not listed, try another part #.

OK, so I looked up the FCP650N80Z 800 V, 10 A continuous-rated FET.  The safe operating area chart shows that at 350 V, it can handle about 400 mA continuously.  Now, this would only be in the short-circuit condition, but a bench supply will experience that sort of incident on occasion.  So, as long as your current limit circuit will be able to pull the gate bias back fast enough, this one should work.

Definitely have more than one transistor handy while building your supply.

Jon

[001]

There are still a few parts around that will make you life easier e.g. Microchip's (ex Supertex) LR8K4 400V three terminal adjustable regulator,  but its only good for a few mA so you'll still need a pass device that can handle 400V at 0.2A without hot-spotting and failing.   The easy option there is still a valve cathode follower.

And what about 10M45? 0-450V 100mA 3pin regulator
By the way I prefer use discrete components

[Ian.M]

That's a resistor programmable current source, not a voltage regulator. 

[T3sl4co1l]

Thanx a lot!

High voltage MOSFETs now very common and cheap http://www.shaoguang.com.cn/pdf1/mosfet/FQA10N80C.pdf

The question not about pass transistor. But about sense & regulation circuit

You seem to have misunderstood Ian's post.

Yes, plenty of cheap, high voltage MOSFETs are available. The trouble is, they're typically not rated for linear operation, only switching. This means that they cannot withstand the full voltage and current rating for very long before blowing up.

The quoted MOSFET shows a straight line DC SOA.  It's fine.

FYI, BJTs are worse than MOSFETs, and IGBTs are worse still.  MOSFETs vary, with some being worse than the best BJTs, but many are available (sometimes at a premium, as the "Linear operation" types are) which specify DC SOA.

I've tested a QFET before, at voltage, and it failed just a bit past ratings (RthJC), no 2nd breakdown.

Tim

[oldway]

And what about using HV bipolar transistors like BU208 ? (Vcex = 1.500V min.)
They where intended for use in horizontal deflection in color TV's....

EDIT: No, not possible, only 30mA at 400V... ...

[T3sl4co1l]

Anyway, as for control, it can be done just the same way the vacuum tube unit did it.  Indeed, you need at least a few "tube replacements" to handle the high voltage: one is the pass transistor, one is its driver (the voltage gain element).  Everything else can be at low voltage, you can use 15V and op-amps if you like.  All you're doing is scaling up op-amp voltage to B+ voltage, then scaling it back down again with a voltage sense divider.

Current sense should be done with a shunt resistor in the GND return path.  No need to touch high voltages, and it's an isolated, stand-alone power supply (right?) so you can do ground-return trickery no problem.

It may be worthwhile to use a source degeneration resistor on the pass transistor, and gate protection zener diode, to limit transient current to safe values.  When the transistor can destroy itself in 20 microseconds, you want to use it carefully.

Tim

[Marco]

So, as long as your current limit circuit will be able to pull the gate bias back fast enough, this one should work.

Yep, he was right to say his MOSFET was specced for his use. It's an old style planar MOSFET as well, so it might actually live up to those specs. That said, the specs also say it requires quite good heatsinking at this power.

[T3sl4co1l]

And what about using HV bipolar transistors like BU208 ? (Vcex = 1.500V min.)
They where intended for use in horizontal deflection in color TV's....

Saturated switching.  Check the SOA:
http://pdf1.alldatasheet.com/datasheet-pdf/view/2595/MOSPEC/BU208.html
15mA at 400V.

Its descendants are still available today, e.g. MJE13008 and its brethren, commonly seen in CFL inverters (and probably LEDs now, too).  Relatives are commonly seen in old computer power supplies, VCRs, DVD players, and so on.

Tim

[oldway]

Vacuum tubes have great advantages for high voltage applications and more particularly for HV power supply.

Indeed,
- there is a large distance between anode and cathode, which is a guarantee of safety
- the current is dynamically limited by the emissivity of the cathode.
- the anode is able to absorb strong overloads of dissipation without damage
- there is no SOA.
- A vacuum tube never fails short, that's a safety feature because semiconductors mostly fail short.

Bench HV power supply must be reliable for obvious safety reasons.

Personally, I would not entrust my life by relying on a semiconductor where there is more than 400V between wires placed a few mm from each other.

[T3sl4co1l]

Vacuum tubes have great advantages for high voltage applications and more particularly for HV power supply.

Personally, I would not entrust my life by relying on a semiconductor where there is more than 400V between wires placed a few mm from each other.

That's alright, you guys and your old ways will die off, eventually...... but hopefully not too soon, indeed, you may some day be entrusting your life to such devices without realizing it.  SCRs and IGBTs lurk inside defib machines!   Quite reliable indeed.

Tim