Simplistic MosFET HV Shunt Regs

pic
 

Attachments

  • p3284221k.jpg
    p3284221k.jpg
    85.4 KB · Views: 1,408
diyAudio Chief Moderator
Joined 2002
Paid Member
Go back to 33R. Its 65mA, more than enough. If it still oscillates see for 1k R7 since the Fairchild has much more input capacitance than the 840, and maybe it oscillates. See if it took out some driver MPSA or BC from the ring of two CCS. Keep leads short and twisted to the Aikido. Also use a local 0.1uF decoupler to ground directly from where the B+ is on Aikido.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
A few notes.

When making some HV shunt, there are practical matters to consider for its layout. Gary Pimm's earlier contribution in this thread about making a PCB has invaluable information about nodes practice, and is applicable in P2P construction style too. Gate stopper resistors for the Mosfets must be connected directly to the gate legs in minimal distance (very short leg) and better be carbon composition. This is very important.
Also when connected to the audio circuitry, the leads inductance must be low. This means twisted pair. Some circuits may react to the very low output impedance of the shunt in very high frequencies. Layout and local EMI decoupling by means of a fast 0,1uF PP cap or stacked ceramic with very short leads directly on the B+ receiving end to ground, may prove invaluable on those audio circuits
Avoid trimming Vout and probing on the shunt when there is some oscillation going on. It can easily take out the driving MPSA mainly, but possibly a BC560 on the Vref ring of two CCS. Making R7 IRF840 gate stopper bigger in value can help in the event that the layout or some other interaction brings possible oscillation. Also a 47R base stopper for the MPSA can help its protection.
 
Excellent advice!

Stoppers are CC but only one really close to the pin (IRFP9240).
I don't have 1K CC resistors so metal film will have to do, and I will move it closer to the gate pin of IRF840.
Wires are short from transformer to the psu but inevitably longer to the Aikido (mere physical size of my breadboard) so I will try to improve on that.
And I will add the decoupling resistor (s)...

Will get back to you tomorrow.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Yes check the BCs. If you have an hfe socket on your DVM see if they are OK. If you want on circuit, I don't advise probing them, but you can look for voltage drops across resistors around them. What 4k resistor? If you don't use the original values around the ring of two, they are going to burn out at those voltages. Just change Q5 as a prediction, and see if it was alive or not off circuit IMO.
 
:hphones:

Music is playing...

Connected to the audio circuit I am playing some music on the cheap test rig...
Observations: can't dial it higher than 170V.
Heatsink of the IRFP9240 is warming up, the heatsink of the IRF840 stays coldish...which might be okay. Other than that only the 47K 5W gets a little warm...
No hum but I will let it play with cheap headphones first - - -

The oscillation seems to be gone, though I am having a few random crackling noises while tubes are warming up (I suspect the 6H30 does that...) Gone after about 30sec's.

This being my second HV regulator being an electronic idiot myself (kind of...) I am not unhappy.

Edit. I am having 260V going in and 170 out... which might be a tad much.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
This sounds like good progress. Not having more than 170V brings to mind 2 things. Either your audio circuit uses more current than expected, or the BCs don't pull as much current as expected so to create as much Vref on given R3. I would parallel 470R on the 100R R5. That will get it down to 82R total. Starting from half trimmer I would see if I can get higher than 170V. Is your R3 82K BTW?
Crackling for 30sec points to tube warm up. IRFP sink a bit hot is normal, it will tame when you go higher Vout.
 
One step forward, one back...

R3 is 82K, R5 now 82R. Measured current draw before the regulator = 56.5mA (2 x 6CG7 and 2 X 6H30).

After quickly taking the DMM out and reconnecting the wire leading into the reg I wanted to measure voltages... and bang - gone was the IRF840. DMM still wired to measure current effectively shorting out the mosfet.
OK, not that big of a deal, but is it likely that I took out other components?

Thanks for bearing with me anyway :D !
 
diyAudio Chief Moderator
Joined 2002
Paid Member
OK such things happen. 840 etc. are cheap stuff. Don't lose heart.
Change only that one, and see if everything still works on dummy load first. So as you measured, your audio circuits draw enough. I would target 75mA through the CCS for start. I normally use double the audio circuits needs, but in your case I would go step by step making sure that I don't hear any compression and establish the lowest threshold for current supply. If you will get no oscillations given your layout etc. you can listen at 100mA CCS later.
 
How warm does the heatsink get? Compared to a Maida.

Maida heatsinks didn't get warm at all (in my application). Barely noticeable...

In my HV shunt the heatsink on the first Mosfet does warm up but it doesn't seem excessive (depending on the applicaton of course).
The heatsink on the IRF840 remained stone cold during the 15 mins of my testing (which again surprised me since I expected more heat on the second Mosfet...).
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Stixx said:


Mind sharing a few values that will have to be changed from now...?

(going from 75mA to 100mA as suggested)

R1=27R looks good for about 79mA at 260V in, 200V out. 0.5W resistor will be comfortable, since it will have to dissipate 165mW and I like resistors nominal at least 3X their dissipation on circuit for reliability. 22R will give about 96mA CCS, 0.5W nominal or 1W still OK.

*I calculate for 1.8V drop LEDS.