High Current, Variable Output, Dual, Lab Power Supply Unit with Single Potentiometer Control

Note 1: Liquiphonics assumes no responsibility for the following circuit diagrams.

Also, bear in mind that it has been designed for myself and so may contain parts you may not need.

This is the circuit diagram (or schematic) for the dual supply, variable output voltage, high current, single potentiometer controlled power supply unit I will be using for testing the Liquiphonics’ HiFi design. The diagram also shows the ILP MOS248 MOSFET power amplifier monoblocks I’ll be including.

The heavy lines show the large gauge wire used to connect the amp power lines and earthing.  This thicker wire is to ensure PSU ripple doesn’t get into the signal and to sink any RF interference in the earth wiring. Some connection dots are larger than others, this is a bug in the app I use called Inkscape.  It is excellent though, especially as it is free, so I’m not complaining.

I haven’t reinvented any wheels, all this information has come from the web, either the manufacturers application notes for the LM317 and LM337 positive and negative variable regulators or circuits from other people who have been good enough to share their work on the web (thanks).

Click the image below to open a larger version of the schematic, feel free to download it.

Lab Power Supply Unit (Dual, Variable, Tracking).

Lab Power Supply Unit (Dual, Variable, Tracking).

The VDR absorbs any voltage surges, probably destroying itself in the process.  The fuse should always be in the live power line only; if you put a fuse in both lines and the negative line only blows, the whole circuit will still be live. The Mains Filter is an off-the-shelf unit I bought from RS Components 30 odd years ago when I originally put this power amp together (the one I am now restructuring and adding to to make this Lab PSU). It has a rating of  3A continuous. The DC Filter helps to reduce the audible hum/buzzing from the mains transformers. I have a vague memory that the mains transformers were rated at 7A continuous (so more than enough for the job then). These are followed by 2 x 22,000µF electrolytic capacitors per channel.  The rating of 63V is a little close but a higher one would be even more expensive. Generally speaking, given the same amount of capacitance, more capacitors of lower value is better than fewer capacitors of higher value as the ripple current is spread between them thus reducing their temperature and increasing their working life.

ICs 1, 2, 4 and 5 (LM317 and LM337) have a maximum current of 1.5A which is a little low. 2N3055 transistors have an absolute maximum of 15A collector current so should be good for a lot more than the ICs can give. IC3 and 4 allow the use of only one potentiometer for changing the ±voltage outputs by ensuring the negative voltage automatically tracks the positive voltage.  I am having to use a rather expensive OP454 as it can handle the high voltage rails I’m using – these op amps get their supplies from the outputs of the transformer. Check the maximum Supply Voltage rating (Vcc plus and minus). If you have reduced supply values you may be able to use something cheaper such as a TL072 . To ensure good tracking make sure the potential divider resistors of R5/R11 and R16/R22 are 1% or as close to each other in value as you can.


About Matal

Hand made HiFi, photos, a friends poetry.
This entry was posted in Construction, Power Amplifier, Power Supplies and tagged , , , , , , , , , , , , , , , , , , , , , , , , , . Bookmark the permalink.

7 Responses to High Current, Variable Output, Dual, Lab Power Supply Unit with Single Potentiometer Control

  1. Hi,
    I am trying to build you power differential power supply supply. Everything works ok on the positive side (LM317), adjustment goes from 1.25 to 23.5 V, however on the negative side (LM337) the voltage adjustment goes from 23.5 only to 22V when I rotate the pot. The output voltage of the opamp OPA445AP stays at about 9.5V – > don’t go to ZERO. Is it correct to have the opamp in openloop?

    Many Thanks,



  2. Hi Christian,

    I built this circuit diagram from information I found online, I have attached the original for you to see. Once I had built the Power Supply Unit myself I did not use it but did notice that the negative supply did not completely track the positive supply. But it tracked much better than yours so was usable.

    I am very busy at the moment, repairing a friend’s Quad 405-2 where one channel has stopped working and then I have to go away for a few days. I will look into this problem as soon as I get a chance, looking at my own PSU. I have always thought it must be something to do with the bias resistors (R5, R11, R10 on my schematic and R2, R3 and R4 on the attached jpg). But I agree with you; openloop opamp, that does seem odd. But I assumed that since they guy who designed it put it online he must have had it working. That is not always a wise thing to do I know but it’s a long road if you have to reinvent the wheel to drive down it.

    Keep in touch about any solution or progress you make and I will do the same.

    Best regards,

    PS Would you mind if I put this conversation on the blog so others may learn from it?


  3. Hi,
    I fix it. The problem was a 2N2222 transistor from Fairchild. This transistor (TO92 case) has the E and C reversed on the case (stupid isn’t it…). Regarding the bias for OPAMP, on your schematic you have a voltage divider build with 2x 4K7 resistors. The V- of the OPAMP is connected here. This means – in small signal terms that the opamp sees on V- 4k7||4k7 = 2.35K. Than, for ideal balancing of bias the V+ input of opamp has to have a series resistor of 2.35K or 2x4k7 in parallel. The control works perfectly from 1.25V (LM317 / 337 reference) up to the desired voltage.

    Regarding the open loop, THIS is NOT open loop. The feedback is provided via ADJ of LM337. How it works… The slightest imbalance between the ground and the voltage divider tension is amplified and applied to the ADJ pin so the imbalance (btw V- and V+) goes to zero. -> this is the feedback loop.

    And, yes please post the conversation on your blog.

    Could you send me some QUAD docs (schematic, pcb, etc). Maybe I’ll try to build one…

    Many Thanks,



  4. Hi
    Are you satisfied with this schematic?
    Any problem with OPA454 temperature? A heat sink is necessary?
    Thank you,


    • Matal says:

      Hi Dan,

      I think I mentioned once before, I did not use this supply for long. But I was happy with the schematic as the supply did work well – the only problem was the negative supply not tracking (following) the positive supply as the negative voltage output wasn’t quite the same voltage as the positive output. But that problem should have been easy to solve by adjusting the resistors on the input to the OPA454 – so make say R22 a variable?

      I do not remember any problems with the temperature of the OPA454 and would not expect it to get hot as it only measures the voltage of the positive supply and then controls the voltage of the negative regulator (LM337). If you are concerned with that opamp’s temperature, put a finger on it when the circuit is switched on and if you cannot hold it on for more than a couple of seconds it is too hot. But if that is the case, there is probably something wrong as I would not expect it to get that hot. Bear in mind that this opamp is very expensive and a cheaper one could easily be used (e.g. TL071/081). The only reason I used the OPA454 is because of the high voltages coming out of the regulators (LM317 and LM337); it was the only one I could find that could take that voltage. That voltage was governed by the transformers I had which output +/-43V. If you have lower voltages (and you probably do), then you could use an opamp that has a lower maximum rail voltage (Vcc max) and it will be much cheaper.

      If I’ve said anything you don’t understand or need more clarification – don’t hesitate to ask!

      Best regards,



  5. Thank you for reply.
    To modify R22 could be an option but I’m looking also to R10.
    Probably a good start will be a circuit simulation. I’ll comeback with some results.
    All the best,


  6. Matal says:

    Looking forward to seeing the results and interested to see what the simulation comes up with. Hopefully you won’t mind me publishing the results here so others (including me) can learn.
    Thanks Dan


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s