Simplistic MosFET HV Shunt Regs

[Salas]

The Zeners and/or the FET are suspicious. The FET especially goes pop without drama in over voltage. But check the CCS works after R1 change and possibly D1 or worse Q1. See at 12V as you say. It was left on how high a voltage input for 5sec?

 

[RollE2k]

230V input voltage, and it may have been 10 sec, since i had time to take the multimeter measuring. 100mA load.
Will check the zeners also. By fet you mean the IRF840?

 

[Salas]

No, the JFET. 230V to 13.5V * 0.1A hmm, it gave Q1 some stick for 10 sec. Proceed to step autopsy.

 

[RollE2k]

the Q1 sink didn't become that warm, but the R1 did burn, so it may have lived through it.
So i will try change the Q6 then, after i have checked the CCS part.
D3 seems like a candidate which may have died?

 

[Salas]

Yep.

 

[RollE2k]

Changed all diodes, since they seemed to be broken (i got contact on them both ways with multimeter), then i changed the R1 and the JFET. Tried measure all resistors on board, and it looked right. tested the CCS with 12v between +input and +f which worked fine.
plugged everything in, started with 230v filtered input, and R1 burned again. Seems like something with the shunting?

 

[Salas]

Do the replaced diodes do the same conducting when off circuit? The 840's gate oxide could be hit when sparking, you are not left with much to check. That and the two BJTs.

 

[roger57]

Sorry if this question has already been posted; did some searching but could not find. Please reply with posts if already answered or threaded.

Has anyone set up an SSHV2 with about 450V input (about 30mA at load required), and what components were changed (need to be changed) ?

Thanks,
Gary

 

[Salas]

If the Jfet has enough Idss you can got about 400V out without changing anything, about input over 400V me no, because the CCS Mosfets are 400V rated, though that condition is over run only just before the output develops, but at your risk.

 

[Salas]

RollE2K, does your DMM light up LEDS in diode mode? Has 2-2.5V? See that the LED can still light up to be sure before proceeding.

 

[roger57]

If the Jfet has enough Idss you can got about 400V out without changing anything, about input over 400V me no, because the CCS Mosfets are 400V rated, though that condition is over run only just before the output develops, but at your risk.

So all other things equal, if the DN2540 CCS devices are replaced by ones with similar specs, except a higher Vds (let's say 500V), would it be reasonably "safe"? I see the IRF840 is at 500V, perhaps a bit marginal?

Gary

 

[Salas]

I don't know if there will be oscillations if dropping in straight higher capacitance Mosfets.

 

[roger57]

I don't know if there will be oscillations if dropping in straight higher capacitance Mosfets.

In my last post, I wasn't implying the use of the IRF840. I was mentioning it in the context of the input voltage being at 450V, and it's maximum Vds at 500V.

For the input CCS, I'm looking at this one:

IXTP1R4N60P IXYS | IXTP1R4N60P-ND | DigiKey

DN2540-----IXTP1R4N60P
Ciss :200pF-----140pF
Coss :12pF-----17pF
Crss :1pF-----2.4pF
Rds(on) :25-----9

Are these numbers too marginal to say it should be OK, without simulating?

Gary

 

[Salas]

That is not depletion mode, its fundamentally incompatible. Try with the standard part, have spares, and if it won't blow for enough on/off cycles then you will be OK.

 

[roger57]

That is not depletion mode, its fundamentally incompatible. Try with the standard part, have spares, and if it won't blow for enough on/off cycles then you will be OK.

What do you mean by "...if it won't blow for enough on/off cycles ..." ?

 

[Salas]

I mean that the part will be exposed over-spec not during working conditions because then it sees only Vin-Vout across. That will be quick until the 10uF cap goes 50V during power on, in its way charging to full Vout. So if the standard part can survive that after many times you had switched on, it will probably serve well. Don't forget to put a 0.1uF across the input if the last filter cap isn't near to input. Then again if you don't wanna risk failure, don't. The official spec is 400V input. If someone else has used more than 400Vin with no problem he may let us know. The 2 bjts should be mini sinked above 300Vout.

 

[roger57]

I mean that the part will be exposed over-spec not during working conditions because then it sees only Vin-Vout across. That will be quick until the 10uF cap goes 50V during power on, in its way charging to full Vout. So if the standard part can survive that after many times you had switched on, it will probably serve well. Don't forget to put a 0.1uF across the input if the last filter cap isn't near to input. Then again if you don't wanna risk failure, don't. The official spec is 400V input. If someone else has used more than 400Vin with no problem he may let us know. The 2 bjts should be mini sinked above 300Vout.

OK. No point in blowing up anything. An alternative is to use a voltage dropping resistor from the 450V supply, to the regulator input?

 

[Salas]

What is the Vout goal?

 

[roger57]

What is the Vout goal?

375V at 32mA load current

 

[Salas]

OK drop 55V on an RC cell, and no worries.