HT supply circuit

[MotorBikeLes]

I have been following the current thread of a MOSFET power supply, which, like its originator I want to use for testing valves. I have a Telequipment D52 mains transformer with various secondaries, including 6.3v for heaters and a 146v normally used for HT.
My idea was to use that for a 200v supply with a bridge plus smoothing, but also with a doubler as an alternative 400v where a higher screen grid voltage is required.
I am NOT proficient at circuit design, so would the experts here please criticise as appropriate. I have not shown any switching on the output side, but will arrange a centre off change over to allow some discharge time between high and low. I have left the output 0v common.
Is my input side switching OK, where I have allowed the transformer's 0v line to be common as well?

[daviddeakin]

That will work, though it is perhaps overcomplicated? This is my suggestion. You may wish to add equalising resistors across each of the caps, and maybe another cap between the + and - rails, but you get the idea of allowing the two rectifiers to share components.

[jim_jobe]

If you're going to use a MOSFET variable voltage unit with this Les then make sure your MOSFET is adequately heatsunk - if there is such a word! The IRF840 has the following maximum ratings:
Vdss:500V Id: 8A Power dissipation:125W @ 25degreeC. Junction temp.: 150 degreesC.

That 125W max dissipation decreases by 1W for every degree rise in temp above 25 degrees.
The power dissipation of the MOSFET would rise to 40W if the unit was turned down to just above 0V output at 100mA.

Jim

[PE9ZZ_JO22KI]

Why not use the same trick as computer power supplies? Use a bridge rectifier and two reservoir caps in series. Live goes to one end of the bridge and neutral to the other end. Now put a switch between N and the junction of the two caps. Closing the switch doubles the voltage. Oh, that's almost exactly what David drew above (and even simpler).

Tjerk, 9ZZ

[MotorBikeLes]

@DavidD, I will study that, but I assume it will be ok and saves some bits and bobs. I like it.
@J_J, It was the reminder in your thread that the dissipation goes up as regulated voltage is reduced that made me think of the circuit. 200v input for lower required voltages, 400 if more than 200v needed.
I used to have an old Farnell PSU that had 2N3055 or similar series regulator, (intended for variable fixed output, if you follow) If you needed 0-15v, you paralleled the transformers secondaries, for 0-30 you had them in series. It was a "set it as required, then fix the covers", rather than an "adjust as required" setup.
@TJ9ZZ. That is exactly what I don't want. I have the 146v winding, and I use that for either 0-200 or 0-400 this way. The computer circuit is intended for high or low inputs, and a constant output. I can see how it would work, but it seems somehow "out of place here".
So, nobody has spotted and serious errors in my circuit? That was what I hoped, but I do know my circuit design limitations, and they are very real and low level.
EDIT. Hang on, are not all these cicuits effectively the same? I am now confusing myself. Must go up a corner and think!
Les.

[jim_jobe]

I have a 2N3055 commercial unit as well Les. It has a tapped secondary on the transformer with 6 different tappings which you switch in as required. Very useful!

[Herald1360]

If that FET is a TO220, there's no way it can dissipate 40W on any practical aircooled heatsink. A TO3 device might just cope. Several in parallel with suitable current sharing resistors would be OK on a heatsink better than 2C/W.

[Skywave]

Originally Posted by Herald1360

If that FET is a TO220, there's no way it can dissipate 40W on any practical air-cooled heatsink.

The case is TO-220AB. Data for the IRF840 is shown below.

http://www.vishay.com/docs/91070/91070.pdf

From that data:
"The TO-220AB package is universally preferred for all commercial-industrial applications at power dissipation levels to approximately 50 W".

Al.

[jim_jobe]

This heatsink is rated at 2.6 degrees/Watt but, as a matter of principle, I agree that paralleling MOSFETS is a sensible thing to do where power dissipation is high.

[Al (astral highway)]

Would it be a good idea to merge the two inter-related current threads on high, variable voltage power supply to make one 'go-to' resource ? Just a thought

[Herald1360]

So, 2.6C/W HS-A +1C/W J-C +0.5C/W C-HS is 4.1C/W under ideal conditions. At 40W that gives 164C rise, plus operating ambient of 25C best gives Tj of 189C. Not nice.

A 1C/W HS would get this down to Tj of 125C which is just about acceptable for long term reliability, but it would be rather large unless fan cooling were invoked.

Two in parallel with good sharing on the 2.6C/W HS would bring Tj down to about 135C for each device which would work, but three with Tj at 125C or four at 120C would be better.

Given the cost of devices vs decent heatsinks, the more devices the merrier- certainly for a DC regulator.

Historically at work we reckoned that for TO3, about 40W each was a safe limit even on aircooled sinks; on a watercooled copper plate you could push 250W Ptot devices to about 90W each, provided you could stay clear of second breakdown region.

[jim_jobe]

You've convinced me Chris! Mind you I've decided to limit the input voltage to 250V now so I only need to dissipate 25W, worse case scenario (my wife reckons I'm already dissipated enough!)
Another wheeze would be to have two switchable output voltage ranges, say 0-125V and 120-250V and arrange to switch in a larger current sensor resistor on the low voltage range but that would be more complicated.
Many thanks
Jim