High voltage linear regulators.

[sureshot]

I'm currently looking at high voltage linear regulators, with the intensions of going up to 50 - 60 Volts maximum output. I've come across two candidates that might be worth using. The circuit is the general linear regulator layout with the addition of two series pass transistors to boost current. I've messed about with the LM317 with 2 x and 4 x series pass transistors with good results. But as I'm looking to go up to 60 Volts is there any specific i need to take in to consideration  ? Below are the two data sheets of the regulators im thinking of trying. Its the one with the higher maximum voltage that I'm most interested in. What i think could be an issue is the higher voltage on the emitter base chain of the transistors and resistors, as 1.25 - 60 Volts is a large differential to cover.
Thanks for reading any help appropriated.

PS. PDF files are to large to post.
Its the LM317HV and the TL783 Voltage regulators. And the later of the two I'm most interested in.
Thanks again

[Zero999]

The LM317 has no upper limit to the maximum output voltage. It's a floating regulator, so all that's important is the difference between the input and output voltage, which mustn't exceed 40V for the standard part, or 40V for the HV variant. The main issue with higher input, than the LM317s maximum rating, when the over-temperature/current protection kicks-in, is the full supply voltage will appear across the regulator.

Another transistor can be added in series with the LM317, to drop most of the voltage and ensure the maximum voltage rating is not exceeded.



Note that the above circuit drops about 40V across the LM317, which is too much. The zener should be changed to a <30V part to drop more voltage across the transistor.

[sureshot]

Thanks for the help with this, so is it not possible to current boost the TL783  ? I know the current will be far less at higher voltages. Something like 3 Amps at 15 Volts, and half that 1.5 Amps at say 30 Volts. And maybe 1 Amp or close to it at 60 Volts. Would that be possible ? Thanks again for the help.

[sureshot]

I couldn't post data sheets, it said file to large. But here are a couple of the circuits from the data sheets. Although one has very little current capability, and the other looks similar to an LM317 circuit. But I don't need nowhere close to 125 Volts, but wandering if series pass transistors can be added to the more LM317 type looking circuit.

[Wolfgang]

Hi,

both the LM317 and the TL783 are not extremely robust at high voltages. When I get you correctly, your output current requirement is a few Amps, this means you have to use external pass elements anyway just because of dissipation, especially when the output get shorted. Transistor SOAR will be an important limiting factor then.

After some bad and smelly experiences with the National Semiconductor LM317 high voltage example circuit (same as you have below)
I tried a TL431 based one with a well-cooled linear MOSFET pass element.

IMHO, it has the following advantages:
- input voltage can be up to 400V
- output voltage from 2.5V to Input voltagev - 20V
- exact current limiting
- output current could be up to a feq amps, depending on power FET used, voltage difference and (fan) cooling
- low ripple

Some info can be found at:

https://electronicprojectsforfun.wordpress.com/power-supplies/high-voltage-lab-power-supplies/

or here :

https://electronicprojectsforfun.wordpress.com/power-supplies/

or here:

https://www.changpuak.ch/electronics/VB408.php - this is where I got the idea from, thanks, Alex.

[floobydust]

Microchip/Supertex LR8 is 450VDC max and similar to TO-92 LM317; output is to drive a pass Darlington/mosfet.
TL783 minimum load is 15mA so a bit of heat wasted there.

[Wolfgang]

Agreed. But is has more ripple, a minimum load requirement and is not as robust as the TL431 circuit.

[Zero999]

Thanks for the help with this, so is it not possible to current boost the TL783  ? I know the current will be far less at higher voltages. Something like 3 Amps at 15 Volts, and half that 1.5 Amps at say 30 Volts. And maybe 1 Amp or close to it at 60 Volts. Would that be possible ? Thanks again for the help.

It's possible to current boost higher voltage regulators, just by adding another transistor.

I couldn't post data sheets, it said file to large. But here are a couple of the circuits from the data sheets. Although one has very little current capability, and the other looks similar to an LM317 circuit. But I don't need nowhere close to 125 Volts, but wandering if series pass transistors can be added to the more LM317 type looking circuit.

Another transistor can be added to the circuit I posted previously, to boost the current too.


The problem with dropping excessive voltage across BJTs is passing any appreciable current for a reasonable period of time, exceeds the safe operating area of the device, resulting in secondary breakdown and lots of smoke.

https://en.wikipedia.org/wiki/Safe_operating_area
http://www.onsemi.jp/pub/Collateral/AN875-D.PDF

If you don't want constant current operation, then why not have fold-back current limiting? Once the over current protection is tripped, the current through the transistor can be reduced, to both save power and keep it within its safe operating area.
https://www.google.com/search?client=firefox-b&ei=RARyW4mgNtT2gQa_-7noBg&q=foldback+current+limiting&oq=fold-back+cur&gs_l=psy-ab.1.0.0i10k1l5j0i22i30k1l2j0i22i10i30k1j0i22i30k1j0i22i10i30k1.95172.101169.0.103939.21.17.4.0.0.0.160.2218.0j17.17.0....0...1.1.64.psy-ab..0.21.2422...0j0i131i67k1j0i131k1j0i67k1j0i13i30k1j0i13i10i30k1j0i10i30k1j0i13i10k1.0._myMLqnob08

[Wolfgang]

In order to avoid SOAR problems, linear MOSFETs are a better choice for pass transistors than bipolar units.
They can also be paralleled, if you provide equalizing resistors just like with bipolar transistors.
I dont know any commercial HV power supply design for more than 100mA that relies on bipolar transistors.
Foldback current limiting is OK, but can lead to startup problems with loads that soak up a lot of current during startup.

[sureshot]

Thank you for your replys, links and schematics. I will definitely read up on this stuff before attempting to put a higher voltage adjustable linear power supply together. Thanks for the suggestion on the other regulator, I will go carefully over that also. I've wanted to use a mosfet as a series pass element, but believe it's voltage driven over normal BJT's current driven.
Thanks again for all the help.

[Zero999]

In order to avoid SOAR problems, linear MOSFETs are a better choice for pass transistors than bipolar units.
They can also be paralleled, if you provide equalizing resistors just like with bipolar transistors.
I dont know any commercial HV power supply design for more than 100mA that relies on bipolar transistors.
Foldback current limiting is OK, but can lead to startup problems with loads that soak up a lot of current during startup.

The standard fix for start-up problems is to add a short time delay to the current fold-back.

Thank you for your replys, links and schematics. I will definitely read up on this stuff before attempting to put a higher voltage adjustable linear power supply together. Thanks for the suggestion on the other regulator, I will go carefully over that also. I've wanted to use a mosfet as a series pass element, but believe it's voltage driven over normal BJT's current driven.
Thanks again for all the help.

A MOSFET can be used as a pass element, but the turn on voltage is higher than a BJT, so it will drop more voltage, than a BJT, but that might not matter. It's also possible to use a low drop-out configuration with a P-MOSFET, but it's difficult to stabilise.

[mikerj]

Thanks for the help with this, so is it not possible to current boost the TL783  ? I know the current will be far less at higher voltages. Something like 3 Amps at 15 Volts, and half that 1.5 Amps at say 30 Volts. And maybe 1 Amp or close to it at 60 Volts. Would that be possible ? Thanks again for the help.

How are you intending to power this regulator?  Is it from a fixed voltage a little above your maximum requirement (e.g. 65v) or do you have pre-regulator to minimise the voltage drop across the regulator?  If the first one then high current and low voltage is the worst possible case for power dissipation, which may be a limiting factor.

[iMo]

What about an LM723??
You may build a low-noise (as the people report here), good stability, almost any ampers/volts linear regulator out of it..

[sureshot]

An LM723 is something I've yet to do, but i expect it will be one of the general schematics for that regulator. I know there expensive, but toroidal transformers as i prefer there low profile over chassis mount types. I did try feeding an LM317 circuit with darlington, and smps for the input, but the final output was unstable up or down a volt from the output fluctuating.

[Wolfgang]

Hi,

some thoughts about the LM723 at high voltages:

- The LM723 datasheets have some HV regulators that are floating. The problem: They have a minimum load requirement, and their stability is not as good as the low voltages ones.

- You can make a good and stable floating LM723 high voltage supply, but you need to make a separate little power supply for the LM723 itself (make it 12-15V stabilized) That works really fine.

[iMo]

The 723 requires maybe 3-4mA for itself, 50y back they used a 12V zener for the 723, the high-side transistors could be stacked up for almost any voltages/currents (as they are current driven out of the 723).