I have few units of 24V 3A output smps power bricks. I'm planning to mod one of this to generate 12V at least at 5A for my LED strips.

My idea is to tap two wires from the secondary winding of the 24V unit chopper transformer. This should give me high frequency AC which I can feed into another self made transformer, wound on ferrite toroid. The secondary of this transformer should have half number of turns but at least doubled wire diameter (compared to the primary) to generate 12V at 5 Amps.

After rectify and smoothed the 12V, regulation loop can be closed by sending the correct DC portion to the controller of the 24V smps brick so that it remain stable.

This scheme, if viable, will give me the freedom to wind required turns for any voltage I need for all my other projects.

My question is, how do I determine the wire diameter and number of turns for the primary of the toroidal transformer which I intend to built?

  • \$\begingroup\$ problem is that the transformer may be designed for a specific winding configuration (not just the turns ratio, but the geometry too), the thing may act oddly, but that's not certain. Some smps designs have some funky transformer waveforms so you may get odd results if you go just on the turns ratio (the turns ratio has little effect on the voltage in a flyback based design). There's a bit of an art to smps transformer design, it's probably easier to get a new power supply (you may be able to tweak the voltage +/- 10% by tweaking the feedback potentiometers - assuming there are any) \$\endgroup\$
    – Sam
    Jun 5, 2016 at 8:58
  • \$\begingroup\$ But toroids are the perfect core for transformers. Maybe the winding geometry is what I should worry about? Insulation is not a concern at all because this transformer works on low voltage. Can you elaborate what you mean by 'funky', please? As long as the DC loop is closed it should be fine isn't? \$\endgroup\$ Jun 5, 2016 at 10:51
  • \$\begingroup\$ It's not just the shape, but the material as well. SMPS transformers routinely hit 50'000Hz with some modern ones hitting 1MHz, an iron core will melt, so the core has to be ferrite, nothing else will survive, there are also hundreds of different grades of ferrite, not all are suited to smps applications (there probably are ferrite toroids though), anything from epcos with either N49, N87, 3F3, 3F4, HC35 grade should work (N49 is one of the best power ferrite grades). \$\endgroup\$
    – Sam
    Jun 6, 2016 at 23:44
  • \$\begingroup\$ Some high power or high efficiency transformers run in resonant (or quasi-resonant) configurations, so insted of being fed a square wave from the switches, a complex combination of inductors and capacitors generates a sine-ish looking waveform across the transformer (helps to keep EMI down and can reduce switching losses when done right), many resonant designs depend on a specific transformer design to work and any modifications may cause the circuit to stop working or behave erratically. If you just want to experiment, go for it, but as a working solution, getting a new psu is much easier \$\endgroup\$
    – Sam
    Jun 6, 2016 at 23:49
  • \$\begingroup\$ The plan that I like to execute, will be more likely to be successful on chopper fed smps or sine fed quasi resonant type? My smps bricks are all chopper fed. I think chopper fed type will be more stable with such mod. Guys, pls advice. \$\endgroup\$ Jun 7, 2016 at 3:17

1 Answer 1


12 LED style SMPS supplies are abundant and cheap. You will save time and money to just buy one.

If you are looking to modify your 24 Volt supply for fun and education, that's a different story.

  • \$\begingroup\$ Its for both experimentation and application as well... \$\endgroup\$ Jun 5, 2016 at 2:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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