Simplistic MosFET HV Shunt Regs

i asked the question to Alex and he said:

pos: around 30ma
neg: around 35ma

with around 1ma change in current draw with well matched input jfets

also he doubt i ll see much difference with a different current source than his, but he is curious to know.

edit:so if we aim for Iccs of 35 and 40 ma it should be sufficient as per his website
 
Last edited:
diyAudio Chief Moderator
Joined 2002
Paid Member
Use 47R 1W for R1 in both, and green 2Vf LEDS. Make sure that you got 50mA or more running, by dividing the voltage drop you will measure across each R1 by its Ohmic value. No, not 35&40, we aim for +20mA over consumption to keep Zo spec OK. The SSHV needs that minimum to keep it low.

P.S. He specs 1 Ohm Zo in his non filtered Zener(s) referenced reg. We worth a try. Plus we use less parts and different Vref.
 
ok 50 ma it will be then.

yes my intuiton also tells me there could be more than ''not much difference''. less parts like you said and different design, plus he uses a handful of zeners. dont know though if z out will make an appreciable difference though.

well we ll see. it s got to be working properly first...

i ll let you know when i built the reg.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
I hope it will hit 55mA with 47R and green 2V LEDS. The minus needs 35+20. 7.5W & 6W heat on each shunt Mosfet. Only 0.5W on each CCS Mosfet with 320V DCin. Sink it adequately and keep your output cabling short and twisted. Bypass C2 with Teflon if you can. Then we roll big way.
 
just powered up the reg and everything SEEMS fine. output voltage is steady and the trimpot works well. i dont have anything to test it on though. i guess now i should be getting my ES amp boards in a couple weeks/less than a month, so we shall see.

so far so good.

also,i begin to be seriously looking for a used oscilloscope to test my gear, so i could become a division of Salas Regulators Research inc. :)
 
oh yes i forgot measuring the current. i get 31ma with 60r and 44ma with 40r. so you say the current depends on the led reference and vgs...is it matter of finding the right combination and that s it? they are all 3 of them 1.94v leds.

my phones are sr-5. pretty satifiying, but thinking about getting lambdas. i m running the sr-5 with my beta22 +hammond 1615 transformers.
 
Last edited:
diyAudio Chief Moderator
Joined 2002
Paid Member
thinking about i m not sure i understood correctly what you meant. the reference leds and Vgs sets the current range, then i have adjust r1 to get the current i want is that it...?

Was expecting 2V-2.1Vf, arrange the resistor so to get those 55mA. The Leds and Vgs have a difference and that is divided by the resistor. Its easier to trim the resistor value. Tolerances in Mosfets and Leds are sure between builds. Its better to go with the resistor. Otherwise is like the pianist joke that pulls the piano instead of pushing his seat so to adjust his position better.:D
Nice headphones BTW.
 
ok will adjust.

yes my father was going to throw them away last year so i just took them i think and wasnt expecting much. what a surprise i had! they since became my reference. i ve tried them (for the first time togheter with your regs) yesterday. it reminded how good those phones are and of course, how good your regs are :)
 
...A consideration is the CCS mosfet when doing heavy duty load experiments. Offload voltage of raw supply, violent load swings, are demanding. Something like FQP4P40 would be an interesting substitution test. Don't know the outcome, just an idea.

I'll bite, this looks interesting. The spec sheet of the FQP4P40 says it'll handle a continuous current drain of -3.5A, a single pulsed avalanche energy of 260mJ, an avalanche current of -3.5A, and power dissipation of 85W (!).

The question I have is, which other components would need to be upgraded to handle a higher current?

I'm thinking of a specific example, a 6C33C-B, 200mA @ 200V.

Vout*iccsmA*X (shunt at +25mA above top consumption, say 300mA)

One channel 6C33C-B: (300+25)-200= 125mA @ 200v -> 25W shared between FQP4P40 and the CCS Mosfet. Quite feasible with heatsinks.

Salas says earlier in the thread, "Keep the constant current source running at least on double the mA that your application circuit needs in steady state. Triple is nicer." So should the above calculation be:

One channel 6C33C-B: (600+25)-200= 425mA @ 200v -> 85W shared between FQP4P40 and the CCS Mosfet, which is a little high for practical purposes.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
FQP4P40 for CCS Mosfet was the thought. I.e. replacing 9610.
For SE class A you would shoot for double its bias plus some margin I suppose. I would measure the real top draw, so to be economic.
Double and triple mA is nice when we talk phono and line level it was conceived for. Amps are something people aspire to push it for, I say its uncharted territory. Vaya Con Dios. :)
 
FQP4P40 for CCS Mosfet was the thought. I.e. replacing 9610.
For SE class A you would shoot for double its bias plus some margin I suppose. I would measure the real top draw, so to be economic.
Double and triple mA is nice when we talk phono and line level it was conceived for. Amps are something people aspire to push it for, I say its uncharted territory. Vaya Con Dios. :)

Thanks for the blessing! :) The IRF840 will handle 8A, if I remember correctly, so it might be worth a try. Methinks it's all gonna run too hot for Thailand though... Possibly, in a OTL it'd be okay. A headphone amp would be a more appropriate first step.

I'll have to wait until the 6C33C-B's together. But thanks for the design Salas. It's these sorts of things that make this forum great.
 
Hi salas, I'm in trouble with the regulator, instead of IRFP9240 I received IRF9610, mounted, and the v-out, with about 300Vcc V-IN is just 50volts and it being very hot in few secs. Also the IRF840 being hot, also with no load.

The prob is the wrong Q1?
 
Last edited:
diyAudio Chief Moderator
Joined 2002
Paid Member
Those boards as far as I am aware of they work right away. I would check that my 2SK170 is OK first. Giving only 50V gives suspicion that the Vref is starved. Either the trimmer isn't going low or the 2SK is bust. Then I would see about the MJE is OK. Also use load. Is there anyway to do something wrong with the board? I have seen it takes centre tap or not trafo connection. Maybe a trouble there?
 
I've changed all transistors, all zeners D5 D6 D7 D8 are ok. V in 220+220Volts leds switch on

after the rectifier diode I have 300Volt cc and just 50 out regulator, IRF9610 and IRF840 are very hot in few seconds also with no load...

this is the diagram
 

Attachments

  • Senza nome.gif
    Senza nome.gif
    121.8 KB · Views: 641
Last edited: