Variable HT regulator circuit.

[Radio Wrangler]

The Mosfet is a source-follower. Its gate is set by a pot across 250v of zener diodes...

So the max output voltage is going to be a few volts less then 250v, not 300. Or have I missed something?


David

[Al (astral highway)]

Well spotted David... so it should have six (282V) of the listed Zeners in the string shown, or seven

Originally Posted by astral highway

but that Mosfet can source a lot of current..would it be an idea to add some simple crowbar s/c protection?

R6 TR2 does a current limit.

[jim_jobe]

Originally Posted by Radio Wrangler

So the max output voltage is going to be a few volts less then 250v, not 300. Or have I missed something?

Yes, my error: that's because I changed my mind about the max voltage but overlooked the top RH corner on my sheet!

[Bazz4CQJ]

Originally Posted by astral highway

My sketch is a bit rough but I didn't tag the original circuit.

Thanks for that Al. I'll certainly add it to my files, but it does look as if with just a few more components, the VV circuit offers a little more versatility in terms of dialing in the HT voltage you want. My particular interest at the moment is that I am rebuilding my HRO and decided in the process to ditch the old dog-kennel psu in favour of a homebrew unit which is a little more compact, but does have silicon diodes doing the rectifer job. The regulated supply will also be useful for my occasional efforts plotting Ia-Vg curves for standard valves for use with my Avo VCM.

B

I don't think the gate vs source voltage matters much as it is only a few volts and the output voltage (the one that matters) is measured. I suppose you could put a few diodes in the earthy end of the pot and zener to compensate.

[daviddeakin]

In practice, most high-voltage zeners produce a slightly higher voltage than they're supposed to. This will more than offset the gate-source voltage.

[julie_m]

I can see that the opto-isolator shorts the gate of the MOSFET to the source while the capacitor is charging, and so keeps it turned off; but shouldn't there be a resistor from HT+ to the gate to turn it on when the capacitor has charged and the LED goes out?

[Al (astral highway)]

Do you mean in the circuit I posted? The op-amp (comparator) only latches on when it detects 2/3 of the supply rail voltage (5V here), hence also switching on the opto-transistor and so the MOSFET. And it's ok for it to stay on until power down.

[daviddeakin]

Originally Posted by ajs_derby

but shouldn't there be a resistor from HT+ to the gate to turn it on when the capacitor has charged and the LED goes out?

I was thinking that too. There's no way for the MOSFET to switch on, except maybe leakage current. Unless the optoisolator is meant to be one of those photo-voltaic devices? But they are very rare.

Also, what is the circuit actually for? It seems to be a slightly elaborate way of delaying the switch on of the HT, not a regulator?

[Al (astral highway)]

Hi David,

If you read back down the thread you 'll see that I put the circuit up precisely in response to a suggestion that there could be a way of delaying HT switch on ... It's really clear!

[Bazz4CQJ]

yep, see my post #58 where I pose the question/suggestion of using the VV circuit to avoid slamming high, off-load HT voltage on to a valve receiver (or amp) during warm up, from a psu with silicon rectifiers. Al's delay circuit was a helpful response to that .

B

[Al (astral highway)]

Hi again folks... if my drawing of the 8 DIP device in my sketch didn't quite make sense, this will help. The device in the circuit shown should be labelled a Dual Photovoltaic Mosfet Driver. It is indeed specifically designed for sourcing a floating turn-on voltage for MOSFETS.

FDA217 is one such device.

It wasn't clear from my sketch and but late in the night, I overlooked the point. Hope the attached link - this one from IXYS -is useful.

The isolation voltage is well over 3kV, useful to know.

Cheers

[Bazz4CQJ]

Thanks Al

B

[Oldcodger]

I'm not certain if special transformers were used, but STC, or was it GEC, used this idea on 4KV submarine cable power feeds. One step up transformer with three windings, one input, one output and third as a saturation winding. Used feedback to control the saturation to stabilise the output. I've often wondered if the principles could be applied to lower ( hundreds of volts) circuits?

[Radio Wrangler]

The thing with the saturation winding is a 'magnetic amplifier' sometimes called a transductor.

The core is designed so that the saturation winding establishes a magnetic flux at right angles to the normal flux direction, so that the transformer doesn't try to output a huge voltage via the saturation winding. If the current in the control (saturation) winding is revved-up, it starts to drive the part of the core its flux crosses up the slopes of the B/H curve and starts to block the normal transformer action flux.

Special steels were developed for mag amp cores.

This technique has been used at every voltage you can make a transformer for, and almost every frequency. There are ferrite versions in some switch-mode power supplies running at hundreds of kilohertz, and I remember an article on doing it to an RF coil to extend the tuning range of a voltage controlled oscillator in the MHz region.

It's been around a bit, but oddly it still isn't a generally known technique.

David

[Bazz4CQJ]

I've just assembled the regulator as per Jim's circuit in post 46 and what I’m finding is that the output voltage will not exceed the volts determined by the zeners (~250V). My input voltage is around 325V and I have a Vsg ranging from ~3.1 to 3.3V across the voltage range. Looking at the circuit in post 13, our German colleague has a zener stack of 310V to enable an output of 300V. I’ve never used this kind of MOSFET before; can anyone offer me some insight? Do I have a problem or should Jim's circuit show an output of 250 rather than ~300 max?

Thanks

B

Further work; looks like Vmax out is always a few volts lower than Vz?

what I’m finding is that the output voltage will not exceed the volts determined by the zeners

That's normal, it won't it will be the turn in voltage of the MOSFET (10V ish) less than the zener volts.

[jim_jobe]

Sorry Bazz, the circuit I posted was just a working sketch. You're quite right; the max output voltage is about the sum of all the zener voltages. If you want a max output of, say 310V, then add another 47V and a 12V zener in series with the five I specified. Make sure that you have adequate cooling with a decently rated heatsink or heatsinks (see the other thread "HT Supply Circuit" , in particular Herald1360's comments on this, particularly with reference to paralleling more MOSFETS. I'm going to use three MOSFETS in parallel when I build this using three 2.6 degreesC/W heatsinks.) The cooling problem mainly arises if you're trying to take near-maximum current out at a low voltage.
Jim

[jim_jobe]

It might be as well to have a fuse in the circuit too before the drain(s) but after the 220uF capacitor. 1A should do the trick - just in case one of the MOSFETs commits suicide! (If you put it before the capacitor it'll blow when you switch on because of the charging surge of the capacitor).