controling multiple LM317 chained in series, all with single pot?

Thread Starter

Hamlet

Joined Jun 10, 2015
519
I'd like to control three LM317 power supplies, each with an isolated transformer.
I want to use one potentiometer, and I want to put their outputs in series. Obviously,
I can't do that because the outputs are placed in series with each other +-+-+-.

If I were using only to regulators, I'd use a dual pot. Since triple pots are unavailable,
is there some other way I can do this? My aim is to create a 1A, 4v to 90v dc supply.

I'd rather not use two LM317HV with a dual 7.5k pot. That would be too easy.
 

ronv

Joined Nov 12, 2008
3,770
I'd like to control three LM317 power supplies, each with an isolated transformer.
I want to use one potentiometer, and I want to put their outputs in series. Obviously,
I can't do that because the outputs are placed in series with each other +-+-+-.

If I were using only to regulators, I'd use a dual pot. Since triple pots are unavailable,
is there some other way I can do this? My aim is to create a 1A, 4v to 90v dc supply.

I'd rather not use two LM317HV with a dual 7.5k pot. That would be too easy.
I would look for a different way to do it instead of using 317's.
 

ronv

Joined Nov 12, 2008
3,770
I'd like to control three LM317 power supplies, each with an isolated transformer.
I want to use one potentiometer, and I want to put their outputs in series. Obviously,
I can't do that because the outputs are placed in series with each other +-+-+-.

If I were using only to regulators, I'd use a dual pot. Since triple pots are unavailable,
is there some other way I can do this? My aim is to create a 1A, 4v to 90v dc supply.

I'd rather not use two LM317HV with a dual 7.5k pot. That would be too easy.
Think about the power when the voltage is set to 4 volts and the current is 1 amp.
 

Sensacell

Joined Jun 19, 2012
3,447
The problem is that "ground" is offset by the voltage of each output.
I agree that this is not the best way to achieve your goals, but consider these ideas:

1) use digital pots- with opto isolation

2) control the regulators with an adjustable current sink, referenced to the ground of the first supply in the series chain.
This is a simple way to control something that is not at ground potential.
 

Thread Starter

Hamlet

Joined Jun 10, 2015
519
Think about the power when the voltage is set to 4 volts and the current is 1 amp.
Sure, it would be considerable. But if I can control three, I can do four or more. This would be for sporadic, bench top use. At 4v, I'm only burning 29 watts each stage.
 

Thread Starter

Hamlet

Joined Jun 10, 2015
519
I've been puzzling on this idea of series connected linear voltage regulators...
What I am thinking is to create a bench supply that can supply about 5A from
say 5v thru 120v+, all smoothly controlled with a single pot. Using four individual
voltage regulators with 24v transformers would limit my dissipation. (I don't want
a switcher and the attendant noise, I don't care about the heat, and I have plenty
of transformers and heatsinks and fans to play with.)

Obviously, I can't use a single pot on the adjust pins on
serially connected outputs on an LM338 in the traditionally way, as the grounds
need to be isolated...

Might I use an optocoupler between each adjust pin & ground, perhaps a 4N28?
Is there a better way?
 

AnalogKid

Joined Aug 1, 2013
11,042
Using four individual voltage regulators with 24v transformers would limit my dissipation.
No, it won't, not the way the problem is stated. If you switch the individual supplies into the output stack as higher and higher voltage is needed, then yes. But if all four supplies are contributing all the time, the total power dissipation in the circuit is the same as if there were a single regulator circuit with a single pass transistor.

And that is the answer. Not all power problems lend themselves to a 317 or 338. This is a perfect example of what Maslow (below) was talking about. If you want to spread out the power dissipation among four pass devices to keep the temperatures down, that is doable without all of the stacked regulator complications. Plus, the overall regulation will be much better.

ak
 

Thread Starter

Hamlet

Joined Jun 10, 2015
519
I appreciate this. One regulator with a large pass transistor bank is a good idea. Because the voltage
range is large, I am limited to a narrow range of available regulators and pass transistors. I already have a bunch
of LM317, LM338, LM317HV in my parts bins, and dozens of 2N3055, TIP35. I get that the dissipation is the same.
Hopefully series stacked regulation won't be too poor...

No, it won't, not the way the problem is stated. If you switch the individual supplies into the output stack as higher and higher voltage is needed, then yes. But if all four supplies are contributing all the time, the total power dissipation in the circuit is the same as if there were a single regulator circuit with a single pass transistor.

And that is the answer. Not all power problems lend themselves to a 317 or 338. This is a perfect example of what Maslow (below) was talking about. If you want to spread out the power dissipation among four pass devices to keep the temperatures down, that is doable without all of the stacked regulator complications. Plus, the overall regulation will be much better.

ak
 

k7elp60

Joined Nov 4, 2008
562
How about this idea. The bottom 317 use fixed resistors and a pot to adjust from about 2V to about 30V. Top LM317 use one fixed value resistor and a single pole rotary switch to switch adjust terminal to ground or fixed resistors in some reasonable steps.. Do the same with
the middle LM317. Use the bottom 317 to adjust to the exact voltage. Connect the three isolated inputs in series like putting 3 batteries in
series. Connect the + of each input to the appropriate 317 input. Connect the - output of the top 317 to the + output of the second 317, and the - output of second output to the + of the bottom 317. The output will be the + out of top 317 and the - out will be the - of bottom 317.
If you need a schematic I can draw one for you.
 

hp1729

Joined Nov 23, 2015
2,304
Where did you find that?!
I was going to make my own, with gears, or nose-to-nose with a thumb wheel between them...
I'd like to control three LM317 power supplies, each with an isolated transformer.
I want to use one potentiometer, and I want to put their outputs in series. Obviously,
I can't do that because the outputs are placed in series with each other +-+-+-.

If I were using only to regulators, I'd use a dual pot. Since triple pots are unavailable,
is there some other way I can do this? My aim is to create a 1A, 4v to 90v dc supply.

I'd rather not use two LM317HV with a dual 7.5k pot. That would be too easy.
Schematic please. "...each with an isolated transformer" but "connected in series"?
 

Thread Starter

Hamlet

Joined Jun 10, 2015
519
4 jointed type In Taiwan's website.
6 jointed type in Ebay.
Thank you Scott!
I'm still thinking of trying something with an optocoupler... I have a bunch of these,
and maybe I can learn something new with them. I have built a single LM338 power supply,
a LM317 power supply with 2N2955 bypass transistors, and a 14 amp ua723 power
supply. Now I want to try for more voltage.
 

ScottWang

Joined Aug 23, 2012
7,400
Thank you Scott!
I'm still thinking of trying something with an optocoupler... I have a bunch of these,
and maybe I can learn something new with them. I have built a single LM338 power supply,
a LM317 power supply with 2N2955 bypass transistors, and a 14 amp ua723 power
supply. Now I want to try for more voltage.
What's this power supply used for?
Do you have load really need this power or just want to learn the skills?
 

Thread Starter

Hamlet

Joined Jun 10, 2015
519
What's this power supply used for?
Do you have load really need this power or just want to learn the skills?
I want another benchtop supply, one capable of more than 30v, and learn a few more skills.

I also want to test devices to destruction. Are ratings conservative, or over stated? How many volts can certain components really take?

I'm not trying to develop a product or do anything with a particular god-damned "goal" or any effort to make money. I'm sick of that. Just pure research and learning, especially analog, coding with solder.
 
Top