Simplistic MosFET HV Shunt Regs

[Salas]

What is the temperature the heatsink should stabilise at?

Thank you and congrats on the design and kit.
Andrei

50C max ref when non boxed should be a very manageable heat level.

 

[radu19880625]

So ... I rebuilt the unit and it seems to work fine.

I have a 5K load at 300v with 350Vin.

I can adjust the current just fine(up to 60mA) but turning the R11 trimmer seems to have no effect on Vout.
Maybe I am just not leaving enough headroom for extra current?

Now I got a much bigger heath sink and the unit runs very cold - at room temp or ~5Deg over. Is this ok?


Thanks,
Amdrei

 

[Salas]

You are burning all the available current on the dummy now indeed. Use as many K as needed to simulate your preamp's consumption figure at nominal B+

Lets hope its only that usual wrongly calculated dummy load cause and there is no dead semi or trimmer in the parallel circuit.

 

[radu19880625]

Thanks for the reply.

I changed the semis and checked all the passives and trimmers.

The curious thing is why is it running so cold now? All parameters - Vout and Iout seem fine.
Is there any other voltage worth measuring to be sure?

 

[Mas Penk]

Apologize for stupid question, but I wonder if I can change the value of 0.33uF with others like 0.22 uF or 0.47uF...

 

[Salas]

Thanks for the reply.

I changed the semis and checked all the passives and trimmers.

The curious thing is why is it running so cold now? All parameters - Vout and Iout seem fine.
Is there any other voltage worth measuring to be sure?

If a big part of the CCS current goes to a heavy dummy load then the output MOSFET runs on fumes (trickle current) and don't expect appreciable heat. As stated in the guide, choose a dummy to simulate your real active load you destine the reg for. As an example if you got a line pre with 20mA max consumption (idle+peak) at 300V B+ then set 40mA CCS and 300V out while using a 15K 15W dummy. 20mA go to the dummy as with real preamp, 20mA minimum stay in the reg for own needs and good output impedance. Then M2 should dissipate 6W (300V*0.02A).

In such true to purpose conditions you trim CCS, Vout, and you judge the heat exchange efficiency.

 

[Salas]

Apologize for stupid question, but I wonder if I can change the value of 0.33uF with others like 0.22 uF or 0.47uF...

You better use 0.33uF. In need of no 0.33uF part option use 0.47uF better than 0.22uF. It will work with all of those but 0.33uF is the midpoint kinda optimum.

 

[radu19880625]

So... I tried a 5k, 7k and 10k load... and it seems to be the same - I get no adjustment from R11 - the trimmer is fine, I checked.
R4 works fine.

I have 346Vin and 318 Vout max, at 47mA over 7k load.
The unit still running dead cold.

Any other points I could measure to check? Thanks!

 

[Salas]

Cold when there is reading on TP? How many mV reading? Is the small sink warm? Check VGS readings on the MOSFETS especially the output one. If there isn't 3-4V between gate and source then its not working. Same about the bipolar transistors, they got to show VBE about 0.6V each.

 

[Mas Penk]

Maybe replaced on 2SK117. I built 4 SSHV2 and most of the time the trouble comes from this small fet... especially when we cannot adjust voltage.

 

[radu19880625]

TP reading 0.47 max on the 7k load. I will look at the transistors.

The small heatsink is just slightly warmer to touch after a few minutes.

 

[Salas]

7K on 318V bring 45.43mA draw. Looks like it just tries to hit max Vref regardless of trimmer position all the time for available CCS current. 47mA in this case. If the JFET is just a short or the BJT that keeps it under its VBE level is out then that's the reason.

 

[radu19880625]

Changed the diodes - also measured them and they were ok.
Changed Q6 and Q3
When I measure Pins 1 and 2 of Q4,5 ( with resp to F0) the Led seems to get a bit brighter.
Q3 has no Vgs -- 0 Volts and also 0V at R7.

The voltages over the transistors are:
Vgs:
Q1 - 1.5
Q2 - 1.5
Q3 - 0
Q6 - 8.55

Vbe:
Q4 - 2.12
Q5 - 8.88

 

[Salas]

Q3 bad reading

Q5, Q4 weird. Surely you measure across their E B (outer pins)?

 

[Salas]

Put 10k dummy only btw.

 

[Salas]

Make sure the 4 wires load drive is engaged fully and correctly also.

 

[radu19880625]

All good now - I changed Q4 and Q5 and it works like a charm. Not sure they were both gone but will measure later.
Just want to make sure it's running nice and cool for now

Thank you so much for all the help!

Regards,
Andrei

 

[AndrewT]

So... I tried a 5k, 7k and 10k load... and it seems to be the same - ...................
I have 346Vin and 318 Vout max, at 47mA over 7k load.
The unit still running dead cold. .................

318Vdc output and those resistor load are a lot of power.

10k = 10.1W
7k = 14.4W
5K = 20.2W

If the load simulating resistors are not fitted, then the regulator has to dissipate the power that is NOT going into the resistors.
10W to 20W is mighty HOT !

 

[radu19880625]

I have 15W resistors mounted on a heath sink - they are also cooled by a tiny fan.

Have been testing at:
40mA for 20mins
50mA for 20mins
60mA for 20 mins

All seems cool, the temp of the main IRF840 sits around 27-29 up to 50mA and now went to 36 at 60mA ... But I do not expect to need such a high current draw.

Just want to make sure it will be fine in the case as it's quite tightly packaged and sits with 4 valves around it.

 

[Salas]

Those BJTs that were creating the problem could have been originally inserted the wrong way by any chance?
Also some live probing slip or an erratic contact of even one of the four Kelvin wires to the load can kill semis at such high voltage levels.

More than the usual 20mA spare on top of your preamp max demand is better for output impedance since it looks like you got the sinking from the the temperatures you report.
...Using a largish sink probably. Can post a photo of the build under test?

Go for something like no more than 50C on sink when all is packed. Make sure that the IRF840 has firm but still electrically isolated contact to its sink.