D1 and D2 removed made no difference, they test ok with my meter with .75v forward voltage in one direction, nothing in the other. Removed D3 and it tested bad, .3v one direction, and 1.1v the other direction so i replaced it with one from the spare board.
I then tried reconnecting the two sense wires and put a 10k resistor across the output and we have a dimly lit LED again with a little under 24.7volts on the output, and the entire board is sourcing ~1ma of current. There is some sign of life now.
Is there a minimum output voltage that the board is now struggling to reach?
I then tried reconnecting the two sense wires and put a 10k resistor across the output and we have a dimly lit LED again with a little under 24.7volts on the output, and the entire board is sourcing ~1ma of current. There is some sign of life now.
Is there a minimum output voltage that the board is now struggling to reach?
In low voltage environment the parallel section can't function with enough space to allow for the proper CCS drop, its literally grabbing at straws trying to surface on an order of magnitude higher output voltage than its now input. I.e. Its pulling up to the CCS's bumper. If you want to test CCS and parallel sections simultaneously from a safely low lab PSU feed, solder 1K 1/4W resistors over R9 & R10 (a piece across each). Now you got a low voltage small range output reg that you can see if it responds to both its trimmers, lights LED1, shows Q3 VGS, shows Q4 & Q5 Vbe, gives CCS TP reading, Vout remains stable where set, etc.
LED2 sets the voltage across the resistor and the transistor such that VfLED2 = Vr2+Vbeq1
Since Vbeq1 is nearly constant, then the remaining voltage across R2 sets the CCS current (Vr2=VfLED2-Vbeq1 and Iccs = Vr2/R2)
VfLED1 sets the voltage across Q1 such that VfLED1=Vbeq2+Vcbq1.
Since Vbeq2 is nearly constant, then Vcbq1=VfLED1-Vbeq2
Taking typical values for a green LED and a Vbe, then Vcbq1 should be ~1.9V-0.6V =~1.3Vcb
We know that Vbeq1 is ~0.6Vbe we can estimate Vce (voltage across q1) as ~1.3Vcb+0.6Vbe = ~1.9Vce
This estimate is VfLED1 The voltage across Q1 MUST BE very close to VfLED1. It's the difference in Vbeq1 and Vbeq2 that creates the small difference.
your 0.973V must be something else. (don't omit the leading zero for values smaller than 1)
Last edited:
I do have a wonderfully unsafe Heathkit lab supply that can provide up to 400v @ 100ma so i can test it at any supply level. Am i right in saying the board really needs 100-150v before you can properly test the voltage regulation?
Depends on the IDSS of your Q6 sample, but the range you mentioned is logical. Don't go unsafe, go safe. Solder 1K resistor across R9 and another 1K resistor across R10. So you can provide say 25V DC input to touch, probe, manipulate, think, care free.
25v supply, 1k resistors on R9/R10 i have an output range of 5.8v-11.85v. On the max 11.85v i have,
Q3 VGS 3.578v
Q4 VBE 0.633v
Q5 VBE 0.630v
Alrighty, connected it up to the big PSU and it seems to be working just fine, was able to set 200v output at 40ma. I'll remove the 2nd board now and run through the same tests and see if the same parts had failed.
While i had the big supply out i connected it up to the CCS in the amp and it no worky, no led's or current flowing across the valve which is good news as whatever the issue is there most probably killed the shunt reg boards in the first place.
Thanks again for the help Salas, much appreciated
While i had the big supply out i connected it up to the CCS in the amp and it no worky, no led's or current flowing across the valve which is good news as whatever the issue is there most probably killed the shunt reg boards in the first place.
Thanks again for the help Salas, much appreciated
Nothing much, congratulations. Take cautious steps now to discover what might be causing your plate load CCS misbehaving. You can pre-test your amp's CCS in LV environment too with a proper resistor for its LEDS. Use the RC decoupling on it also.
I found the problem with this channel, i had a sloppy solder joint shorting out my bias adjustment board so the tube was basically open circuit, it wasn't the CCS at fault at all. I have that fixed now and the CCS is working fine. Haven't checked the other channel yet.
I did start on the 2nd shunt reg board and found Q6 is dead, and D3 reads a lower forward bias than it should so i have to order some parts to replace them. Current side of the board tested out fine so it should be an easy fix.
The RC decoupling is a bit of an issue as i dont really have the space for a large electrolytic cap in there, or an easily accessible point to solder them onto. I'll have to think about that, might have to settle for a small poly film cap.
I did start on the 2nd shunt reg board and found Q6 is dead, and D3 reads a lower forward bias than it should so i have to order some parts to replace them. Current side of the board tested out fine so it should be an easy fix.
The RC decoupling is a bit of an issue as i dont really have the space for a large electrolytic cap in there, or an easily accessible point to solder them onto. I'll have to think about that, might have to settle for a small poly film cap.
I finally finished building the 420V version with help of your suggestion. Thanks.
The reg works fine but I also run into a problem which I don't think it's related to the SSHV2 it's self, but it might have something do to with it that I'm not too sure.
The problem is when switching off the power amp , I have pop sound from the speakers. it seems there is some kind of DC floating at output because I noticed if I mute the preamp before switching off the power amp, all normal and no pop sound. The preamp works fine when I re-connect its original power supply. I guess there maybe some issue with earth or grounding. But before I start to do anything, what's your suggestion ?
Surely the reg does not rely on output capacitor of its own, to be fast in frequency too, and the rail collapses almost immediately at power off. Can this be a possible reason? If the original PSU does power down at same rate then the reason must be elsewhere.
edit: When powering down the amp you wrote not the preamp. I read your description bit fast, sorry. Is there no output coupling capacitor in the preamp? Can you actually measure there is DC leaking or not?
edit: When powering down the amp you wrote not the preamp. I read your description bit fast, sorry. Is there no output coupling capacitor in the preamp? Can you actually measure there is DC leaking or not?
There is certainly the question about the way you used to "mute the preamp", but it seems the most obvious change in the power supply is the absence of the capacitors on the output. The "usual" PSU capacitors keep the voltage for some time after switch-off and are discharge gradually, the shunt regulator keeps the output voltage for some minimal time also (because of the filter capacitors before it), but then cut the load rather abruptly.
Also this reg has unusually wide bandwidth so some transient could be better picked up also. Is there RC suppression across the power amp's on/off switch?
The preamp has coupling caps at output. I did measure the output but no DC I can read. I'm not sure if the power amp has RC supression or not, I will open it up and check.
I did some test. The shunt regs are in separate chassis and they are grounded through F0, S0 wires at the preamp chassis, then there is an earth wire running from there back to reg's chassis . I now changed to ground the regs directly to their local chassis, that improved a little bit only. I'm wondering where is the suggested ground point for the regs in system?
I did some test. The shunt regs are in separate chassis and they are grounded through F0, S0 wires at the preamp chassis, then there is an earth wire running from there back to reg's chassis . I now changed to ground the regs directly to their local chassis, that improved a little bit only. I'm wondering where is the suggested ground point for the regs in system?
Now its more descriptive and more suspicious about transient interference. Long sense lines and a power transient they could be peaking when the amp shuts down. What happens if you go two wire mode? If the pop stops, then its transient interference pick up indeed. Which can be either aided at its source which could be suppressing the switch or at the "antenna" by using shielded coax for the sense lines if not already. If there are plate load current sources in your preamp or very little swinging current stages they would not care at all or very little if its in two wire mode anyway. There is no standard ground point for the regs. Usually where the power in 0 of the preamp leads it.
Actually 8 wires , 4 wires for each reg are 4 twisted pairs and they are shielded using copper braid. I will do test tomorrow as you suggested to run it at two wires mode and see.
Also I forgot to tell that there will be no pop if I just run the amp for a quickly few minutes then shut down. This looks more like the coupling caps are leaking but it does not convince me since the original regs are fine.
Use thin flexible coax like RG174. Those will bring the count down by four sense wires since you will use their shields for S0s. Also far better shielded. And ride them on the force wires. If it proves it is about the sense loop area of course.
Remember you should jumper S0 with F0 & S+ with F+ locally at each output connector and then use thick enough force type wires only. Good luck.