• Disclaimer: This Vendor's Forum is a paid-for commercial area. Unlike the rest of diyAudio, the Vendor has complete control of what may or may not be posted in this forum. If you wish to discuss technical matters outside the bounds of what is permitted by the Vendor, please use the non-commercial areas of diyAudio to do so.

OLD THREAD Simplistic Mosfet HV Shunt (The Simpler Simplistic Design by Salas)

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Thanks for update

PCB Group buy, round 2. Please Update the list yourselves when you type:

Note
1Pcb have: 2 SSHV + 2 filament!

1. kstlfido = 2 pcb
2. Sorr = 1 1/2 pcb
3. ksnider1 = 2 pcb
4. fredlock = 1 pcb
5. pinholer = 1 pcb
6. j45yip = 1 pcb
7. Skylar = 1 pcb
8. DIYspaceW = 1pcb
9. neskor = 1 pcb
10. damohpi = 2 pcb
11. mach1 = 2 pcb
12. Quark = 2 pcb
13. wendell c = 1pcb
14.ksnider1
.....
 
PCB Group buy, round 2. Please Update the list yourselves when you type:

Note
1Pcb have: 2 SSHV + 2 filament!

1. kstlfido = 2 pcb
2. Sorr = 1 1/2 pcb
3. ksnider1 = 2 pcb
4. fredlock = 1 pcb
5. pinholer = 1 pcb
6. j45yip = 1 pcb
7. Skylar = 1 pcb
8. DIYspaceW = 1pcb
9. neskor = 1 pcb
10. damohpi = 2 pcb
11. mach1 = 2 pcb
12. Quark = 2 pcb
13. wendell c = 1pcb

.....

I'm on the list on #3. I deleted my self for #14. I was asking when PCB's will be in production. When you start shipping ?
 
PCB Group buy, round 2. Please Update the list yourselves when you type:

Note
1Pcb have: 2 SSHV + 2 filament!

1. kstlfido = 2 pcb
2. Sorr = 1 1/2 pcb
3. ksnider1 = 2 pcb
4. fredlock = 1 pcb
5. pinholer = 1 pcb
6. j45yip = 1 pcb
7. Skylar = 1 pcb
8. DIYspaceW = 1pcb
9. neskor = 1 pcb
10. damohpi = 2 pcb
11. mach1 = 2 pcb
12. Quark = 2 pcb
13. wendell c = 1pcb
14. sonata149 = 1 pcb
.....
 
Hi! Regal.
I have one up and working. I only used it on the bench for now and as soon I find the time, I will put it in my tube preamplifier. That will be the real test, sonicaly speaking I mean.
On the bench it is working very well, easy to assemble. I ajusted the voltage a bit lower, because it rises slowly in time to stabilise after 10 minutes of operation. The rise is about 7 volts now. It depends on the 2SK170, because with an other one earlier the rise was around 13V.
I tried a 2SK246. It worked also but this JFET has a negative temperature coefficient, so the voltage starts high and lowers in time.
I did not see improvement in the voltage variation, so I returned to a 2SK170.
THe 2SK246 makes a better CCS than the K170. Best of all is the K330 which is almost ideal in this application, but difficult to source.
Both these JFETs have P complementary: 2SJ103 and 2SJ105 respectively.
Since Salas is an ace in shunt PS design using CCS, I hope the information above can be usefull to him in any application he desire.
The K330 can be found in Orient, but I do not have acces there. May be somebody can help for this.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
The drift with JFETs is annoying indeed. But if we think about it, not really a problem in my view. Because after it sets its steady, if we are not blowing over and let it be that is. The Norton Ref will always translate I to V more touchily when utilizing a large resistor like here. It has much less noise than high voltage Zeners, subjectively a very nice signature and can utilize a small film filter cap due to its higher impedance. No sluggish HV electrolytics, which are very audible in that position. Now there are JFETs with less drift, but I would prefer the positive drifting ones amongst best, its safer to know you start lower than higher. Its a bit of a slow start too. Fixing the voltage after some minutes is the practice, I was reminding that steadily in the SSHV thread. Especially with valves we would expect the system to warm up for 15 minutes anyway before focusing. When selecting JFETs the faster settling ones are tell tale when stabilizing IDSS on DVM and are mostly the over 10mA ones, at least in those batches I have measured. There is a relation of IDSS, needed current to reach Vout and -Vp that hits their steadiest, but its beyond scope here. The experts know what I mean. Does 2SK330 come at about -0.5-0.6Vp on average? I see enough RDS (ON) and rather low Yfs, it may average over -2Vp and in that case would not be regulating that well under just Vbe tension here. If it wasn't for that, a large -Vp JFET is steadier. But if we lift Vbe with say an LED or two, unfortunately the loop gain suffers due to degeneration of the driver BJT.
 
The drift with JFETs is annoying indeed. But if we think about it, not really a problem in my view. Because after it sets its steady, if we are not blowing over and let it be that is.


In my case it is not a real problem. I set the voltage lower and let it stabilize at the desired operating voltage. Do you think a K170V could improve this performance?



Now there are JFETs with less drift, but I would prefer the positive drifting ones amongst best, its safer to know you start lower than higher.


I agree with the above statement and that is why I returned to the K170.
On the bench it is working very well.
I now have to test the sonic performance in my tube preamplifier. This should not be long and I will report to you.

I used JFETs for CCS in other applications different from yours, but unfortunately I never could put my hands on K330BLs Their ID-VDS curves are very interesting and matches the CR160 series current diodes by Vishay.
The K246BL is close and the K118 is found in GR only. The K246 I tried in your supply is a GR type at 4.2 ma. I beleive it is too low. That is all I have.
Anyway this is all fun, thank you very much.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Do you think a K170V could improve this performance?

When I was measuring and noting down IDSS from a large batch of BL and 12pcs around 15mA V lately, the about 10mA BL and the V looked like settling significantly faster than the 7-9mA BL. I am not sure its a trend, but if you have some strong pieces check out which ones look like taking final IDSS value on the DVM faster than others. See if this translates to less drift time on the circuit of interest. Will be happy to hear about your impressions when the reg will be feeding your pre amp too. Good luck.
 
I don't think drift in its self is too much a problem , but when it has a low periodic frequency then there are issues. Lowish hz (<0.1 or so) can pass right thru an average sized coupling cap and present itself as DC to the next stage, usually not an issue but can play havoc with some circuits.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.