Simplistic Mosfet HV Shunt Regs

[Salas]

Originally Posted by merlin el mago

Salas uses 2N6520, Quanghao uses MJE350 for all Vout changing some resistors values on his pcb, so if you intended to use Salas schematic I will put 2N6520 & if you will use Quanghao pcb you can use MJE350.

MJE350
VCBO 300V
VCEO 300V

2N6520
VCBO 350V
VCEO 350V

Originally Posted by kacernator

So I should go with higher voltage version then. I cant buy 2N6520 in my country, but they do have BF493S (also 350V), so I hope it will work well as substitute.

My schematic has MJE350 for up to 300V. QH has shown some 340V Vo build with the same transistor still holding. That is certainly beyond its datasheet max, thus I can't recommend it for reliability. You may try it for a bit over 300V but its your responsibility. Some newer production wafers for those are more rugged than when the specs were made it seems, they are old designs. Those TO-92 you mentioned can be certainly used since they make the spec. Use mini sinks on TO-92. I would still prefer the larger format MJE5731A 400V that member Hearinspace have used successfully if you can find it. The reg has proven elastic for hosting different driver transistors up to now, but the main thing is those transistors to prove stable themselves when near their max.

[kacernator]

Ok, MJE350 will propably survive 300V Vout, will try. I will have 350V in and 300V out. Should I also change resistors R3 to 82k and R4 to 220k when Vin is 350V?

Thanks.

[vgeorge]

Hi Salas,

on more question from me.
Is it OK to use the 2N6520 even for voltages lower than 200V, or it is recommended to use the MJE350?

[Salas]

Originally Posted by pedefede

finished the positive part of my dual 300V supply for Exstata. Its P2p.

The heatsinks are a bit small (the ones for exstata's PSU), so the tempereture on the heatsinks stabilized after 15 minutes.
For irf9610 to 50 C, and for irf840 to 80C

(A bit high i guess. Adding bigger heatsinks is a problem in my chassis design, so i will add a fan. For test i just put up a 12'' PC-fan, and the tempereture fell to only 40 C

I also measured the V-out during warmup, and there was quite a nice linear correlation between V-out and tempereture of (the heatsink of) irf840. V-out raised roughly 0,5V per degree C. Is that normal??

The Jfet current is susceptible to its thermal environment. There is enough drift until settling always, due to direct translation in voltage on the large value reference resistor. The more heat or suddenly blowing air in there the larger the drift or change. I still prefer the low noise easily filtered with small MKP & subjective results with that simple passive voltage reference than using a high voltage Zener that would be very tight in wandering. Drift is a nuisance, needing to watch it and trim for ballpark settling value in final installation. I have mentioned there will be such a thing from the very first posts. So in this reg concept it is normal to answer your question. Accentuated in your build non the less.

[Salas]

Originally Posted by kacernator

Ok, MJE350 will propably survive 300V Vout, will try. I will have 350V in and 300V out. Should I also change resistors R3 to 82k and R4 to 220k when Vin is 350V?

Thanks.

MJE350 has 300V Vceo spec no doubt. Not so high Vo to make other changes pressing, just use R3 68K 5W for more comfortable dissipation I would think.

[Salas]

Originally Posted by vgeorge

Hi Salas,

on more question from me.
Is it OK to use the 2N6520 even for voltages lower than 200V, or it is recommended to use the MJE350?

Both work as it proved. 2N6520 is lower Cob and higher Ft. Below 200V it will dissipate comfortably too. You can try MJE also and decide on subjective grounds.

[vgeorge]

I use the 350 now, but I have a few 6520 and want to try.
Thanks

[dodo]

Originally Posted by dvb-projekt

That´s the problem. Nobody has them in stock. I try to get a quote from Caddock directly.

I wouldn't use any high wattage resistor in the audio aplication , which has only 1000h liftime , no mettter how others good parameters it has

read the data more carefully

[merlin el mago]

Originally Posted by dodo

I wouldn't use any high wattage resistor in the audio aplication , which has only 1000h liftime , no mettter how others good parameters it has

read the data more carefully

Load life changing value not life time.

[dodo]

Originally Posted by merlin el mago

Load life changing value not life time.

merlin , it's the question how you read the data
As I see it , the producer confirm during the 1000h of work , there will be no failure and max resistance change will be less then 0,5%
after those 1000 h producer can't guaranty anything
just for the compariton
MBE0414 Beyschlag 50ppm - 33 cents/ pice
after 1000h - delta R less 0,4%
after 8000 h - delta R less 0,8%

Those Caddocks are designed for aerospace , where there is a strict procedure what kind of elements you have to change after exact working time ( they calculate it )

I doubt you have some notice how many hours you use your audio system every day