My discrete voltage regulator.

[sawreyrw]

Russ,

O2 and Q4 are worthless; all the stored energy must be provided by C3. I don't think you need much capacitance across D1, so delete Q3 and all the extra stuff.

Rick

 

[Russ White]

You mean R2 and R5 Jan? Yes I agree with you.

Q2 & Q4 and Q3 form capacitance multipliers, and are an elegant way of achieving low noise operation.

Nice work Russ. Have you abandoned the MOSFET output device?

Thanks Richie,

You guys helped me a lot along the learning curve.

Jan/Richie (R2 and R5) Ah yes, I just simulated that change. Indeed it does make a notable difference. Thanks!!

As for the MOSFET that is still and option for the high current/voltage version unless Darlington output is a viable alternative.

Cheers!
Russ

 

[Russ White]

Russ,

O2 and Q4 are worthless; all the stored energy must be provided by C3. I don't think you need much capacitance across D1, so delete Q3 and all the extra stuff.

Rick

I think you might be mistaken, but I will take it under advisement.

 

[jackinnj]

OK, now calculate the gain and phase margin ! (Will respond after dinner !)

 

[richie00boy]

The capacitance multipliers are not worthless, you just have to be aware of their correct usage.

True they cannot give you any real stored energy, so you do need the usual large smoothing caps to meet current demand. However, they do massively reduce the ripple on the line as long as the input-output differential is maintained.

So in this application if Russ wants ultra low noise and can accomodate them on board easily then why not.

edit:
Just a curiosity whilst you are playing with the simulator Russ, but try adding a resistor (maybe 100 ohms) inbetween the first cap multiplier and the reg itself, and converting the reg into a shunt reg by putting the output transistor emitter to ground and taking the feedback from the collector, and switching the op-amp inputs again. See if this gives you better noise performance.

 

[Russ White]

OK, now calculate the gain and phase margin ! (Will respond after dinner !)

Stupid question (but sincere), why? What is it you are trying to reveal?

As I said the circuit is tested and working, and tomorrow I will be taking it to the university lab which is the only place I have access to a scope (and people who know much more than I do).

 

[Russ White]

edit:
Just a curiosity whilst you are playing with the simulator Russ, but try adding a resistor (maybe 100 ohms) inbetween the first cap multiplier and the reg itself, and converting the reg into a shunt reg by putting the output transistor emitter to ground and taking the feedback from the collector, and switching the op-amp inputs again. See if this gives you better noise performance.

An excellent idea! I will try your proposal as soon as I am able.

 

[lineup]

hi
there are some ways to use
TL431 Adjustable Precision Shunt Regulator
for high voltage regulator

Try this search!
http://www.google.com/search?num=100&q=shunt+regulator+"high+voltage"

For example here is regulator using TL431
giving 500V out
http://www.tubecad.com/july99/page12.html

Perfect for TUBE people, too.

Should not be too difficult to use same circuit idea
with transistors, MOSFET or Bipolar
------------

Also have a look at this topic from 2004:
High voltage regulator, 46 posts
http://www.diyaudio.com/forums/showthread.php?threadid=39986

with a link to this MOSFET 300 V regulator:
http://www.glass-ware.com/tubecircuits/High_Voltage_Regulator.html

An externally hosted image should be here but it was not working when we last tested it.

 

[Russ White]

lots of interesting stuff.....

lineup,

thanks for the good info.

Let me explane something. I am what people sometimes call a hands on learner. I am the sort of guy who sometimes does not get it until I have to actually design it. So call me thick. But this sort of design from scratch process brings me a lot of knowledge. So while I have gleaned a lot from many google searches and even on this forum, I get the most satisfaction when the thing I worked out for myself actually works. Which in this case it does (sheer luck I asure you).

 

[Russ White]

Just a curiosity whilst you are playing with the simulator Russ, but try adding a resistor (maybe 100 ohms) inbetween the first cap multiplier and the reg itself, and converting the reg into a shunt reg by putting the output transistor emitter to ground and taking the feedback from the collector, and switching the op-amp inputs again. See if this gives you better noise performance.

Ok, I simulated the shunt version, and yes it works, but.... (and this is likely my fault) load regulation seems a sight worse. More than double the voltage noise when regulating a 1000hz 20ma load. Also the useful load become much less.

I wonder if I would improve those things by using a CCS.

Cheers!
Russ

 

[richie00boy]

You did take the output from the node where you took feedback from? (I'm sure you did but...)

A CCS will help but IMO if the gains are not there from just a resistor I would abandon it. You could try playing with the resistor value a bit to ensure it's allowing you to pass an absolute minimum of 20mA easily and see what happens.

The load range will be worse than with a series reg due to the load must always draw less current than the shunt.

 

[ashbytub]

Shunt Regulator

HI,
Is there any P.C. Board avivable for this regulator?
Regards,
Ashby