Currently I'm working on my first electrical project - A dual voltage power supply. I'm very new to electrical design and have no prior knowledge of it. But I do learn fast so please bear with me. I was designing a power supply using an LM317/337 current boosted circuit, but I've been told that the design is outdated and I should look into other solutions.

lm317/337 dual voltage current boosted circuit R5 and R6 also have a potentiometer following them to limit the minimum voltage to 24 volts with the maximum at 49 volts. The transformer in the diagram has a series resistance of 1.0 Ω on the primary side and 0.6 Ω on the secondaries.

The reason I chose this circuit, is because of it's simplicity, how often it's still being used in audio receivers and amplifiers, and I also have the parts available from salvaged audio equipment.

The power supply needs to be able to provide a minimum of 24 V/-24 V (48 V DC) on each rail.
Maximum voltage is +60 V/-60 V (120 V DC).
It's being used to provide power for a 48 V DC motor controller.
It will also be used for titanium anodizing so the voltage needs to be adjustable.
The 48 V DC motor is 300 W.

EDIT: To better clarify my needs - I am trying to build a 48-54 V linear regulated power supply for a CNC motor controller.

I've added a link to the page with the original design which uses a 5v or 24v pwm signal to control a mosfet motor controller. However mine will use the 12v signal from the CNC controller which I've added a link below to the schematics of the controller. 3018 cnc controller schematic

The motor I am using is linked below. It is a 48 V DC Motor which uses 300 W. I would like to leave some headroom incase I upgrade to the 500 W motor. https://www.amazon.ca/RATTMMOTOR-12-48VDC-Air-Cooled-12000RMP-Engraving/dp/B08HGNXZVF/ref=asc_df_B08HGNXZVF/?tag=googleshopc0c-20&linkCode=df0&hvadid=578803204497&hvpos=&hvnetw=g&hvrand=3139667351443308493&hvpone=&hvptwo=&hvqmt=&hvdev=c&hvdvcmdl=&hvlocint=&hvlocphy=9001016&hvtargid=pla-1853557070810&psc=1

  • 1
    \$\begingroup\$ You don't ask a question, but the regulators are not able to handle the specs you need, so outdated or not, you need either different regulators that can handle your specs or you need to modify your specs so those regulators can handle it. Also the transistors are unknown. \$\endgroup\$
    – Justme
    Apr 1, 2023 at 21:04
  • 1
    \$\begingroup\$ Does this answer your question? Help with diagnosing a dual voltage power supply \$\endgroup\$
    – winny
    Apr 1, 2023 at 21:19
  • \$\begingroup\$ @Justme I’m looking for help to design a power supply capable of handling the loads described in my post. Like you and others have said the transistors aren’t capable of handling the load described. It’s also very inefficient and outdated. Ideally I’d like something with a simple design, since I’m still learning. \$\endgroup\$
    – Infernoman
    Apr 1, 2023 at 22:06
  • \$\begingroup\$ @winny that was my original question. And as suggested by greybeard I’m looking into a newer design. I wanted to post what I have currently as a reference. \$\endgroup\$
    – Infernoman
    Apr 1, 2023 at 22:09
  • 1
    \$\begingroup\$ Buy two 48 V isolated power supplies and attach a buck converter for each to do the regulation? \$\endgroup\$
    – winny
    Apr 1, 2023 at 22:17

2 Answers 2


I'm going to address your question. That doesn't mean a quantitative answer, because your question is about a first electrical project and some general details that you are looking to provide power for a DC motor and titanium anodizing.

  • DC motor: A dual voltage supply for a DC motor may allow you to operate the DC motor in both directions where one side of the motor is grounded and you switch the other side between the two voltages. However, widely available dual half-bridge or full-bridge ICs (or controllers that operate discretes for bridged switching) make dual rails far less necessary. I could argue that you are better off focusing on a very good single-rail supply as your first project and spending time developing your knowledge about bridging for DC motor control, which you'll need to learn about, anyway.
  • Titanium anodizing: I'm no expert on titanium, specifically. But anodizing, in general and specifically when considering this is a 'first electrical project', argues strongly for a constant current source. As the anodizing proceeds the impedance of the thickening layer increases. If you use a constant current source it's a lot easier (a lot easier) to control for a desired result. So here you don't even want a constant voltage supply circuit, but instead a constant current circuit -- which doesn't look the same (unless you buy a commercial power supply that supports both, in which case it does look the same on the outside, at least.)

My advice, since this is a 'first electrical project', is to pick some one thing you want to achieve. It's not the time to try and make yourself an all-in-one mix-master. It can be done. But not as a first project. Learning is a process of incremental steps.

Final note. The reason for suggesting emitter resistors in the external bypass PNP (or NPN) BJTs is because BJTs vary, one to another, and it almost always comes to pass that just one of them picks up the hogs share of the load and the others do little. The emitter resistor adds so-called negative feedback. (This is something that comes in many shapes and flavors, so be aware that you will need to acquire a lot of learning to be able to see it in its various disguises.) By doing so, when one of the BJTs tries to pick up too much of the load the resistor 'resists' this by signalling the other BJTs (via the shared base voltage) to do their share of the work. We can always get into more detail about exactly how that works. But that's for another time. My point here is that there is already enough for you to learn in designing a basic DC power supply. It's a great first design project if you don't let it get out of hand. Keep it simple and focused. Solve just one problem for now. And milk that project for all you can learn from it. There's a lot just in a very simple DC power supply, not to go complicating the learning process by biting off too much, right now.

  • \$\begingroup\$ Thank you for taking the time to explain. I have been doing a lot of testing and reading about transistors and how they work as well as different biasing arrangements and negative feedback. I'm sure I still have a lot to learn but I have a basic grasp of how they work. Originally that was my plan for building a regulator but after learning about feedback, error amplifying etc. that went down the drain. If I had to pick one thing it would be the regulated supply capable of powering a mosfet based motor controller. I only need single direction since its for a small CNC machine. \$\endgroup\$
    – Infernoman
    Apr 1, 2023 at 23:44
  • \$\begingroup\$ Here's the motor controller I've based mine on. Mine will use a 12v pwm gate signal from the cnc controller instead of the 5v/24v pwm. And instead of 36v i'd like to be able to provide 48-54 volts. pedja.supurovic.net/36v-cnc-spindle-motor-driver/?lang=lat \$\endgroup\$
    – Infernoman
    Apr 1, 2023 at 23:47
  • 1
    \$\begingroup\$ @Infernoman That site helps a lot in understanding. I'd recommend that you rephrase your question and provide that link. (Or write a new question -- I haven't figured out how the site managers make their decisions about what has to be combined, re-edited, deleted and re-asked, etc., and I'm not interested in making a study of the subject, either.) You clarified a lot for me with that link. You just want to be able to control higher voltages (which will likely mean higher currents, too.) So this is a beefy first project. But at least it is a narrower one than before! \$\endgroup\$ Apr 1, 2023 at 23:52
  • \$\begingroup\$ @Infernoman You may want to specify what PWM source you intend using, as well. If it is a commercial product, say so and specify which one. If it is a custom unit you built, then you may want to provide detail for that, too. But one huge point here is that this is PWM and that you probably do not need a variable-voltage supply. That's a good thing. All this provides context. Even if your question is only about making a DC supply to use for a later PWM circuit, letting us know this fact is still important. It clears away a lot of questions, fast. \$\endgroup\$ Apr 1, 2023 at 23:54
  • \$\begingroup\$ @Infernoman Note that if you are looking for 54 V at the motor and able to support more than the 23 A into the motor, say up to 30 A?, then you are in the ballpark of a 2 kW power supply. Even with a fancy switcher supply this could easily mean having to dissipate hundreds of watts in the supply itself. And learning to do that well is a study course, with multiple chapters and topics, all of its own. My first design project was a single rail DC power supply, too. But 5 V and 6 A. (No ICs back then. Just discretes.) Which is down-right easy by comparison. \$\endgroup\$ Apr 2, 2023 at 0:02

(I have salvaged parts available valuable info, if late in the process.
getting to the bottom of what's needed is hard, and I found iterations of "designs" (how to do this? would that work?) helpful.)

Powering a motor does not need a stabilised/regulated supply.
While regulation sort of comes free with switch-mode supplies, it is entirely dispensable with mains frequency transformer, rectifier, (optionally a choke,) and load capacitor.
You can even afford a substantial ripple,
a giant one (valley fill rectifier, >50%) if you compensate via pulse width ("open loop control" instead of "closed loop regulation" with stability concerns).

A finger exercise in mains frequency rectifiers:
(Resistive "valley fill load" for lack of a better idea - series inductance didn't work out.
Component values starting with assumptions about the transformer and 300 W load you mentioned, "hand tuned" iteratively.)


simulate this circuit – Schematic created using CircuitLab
(scroll down transient response for power graph.
Notice the rectifier input currents. (never mind overloaded 1N5408s)
The 1mF capacitor is for low ripple with low loads: increase load to, say, 33 Ω.)


Not the answer you're looking for? Browse other questions tagged or ask your own question.