I have been needing a 3.3 volt power supply that can use at least 12-24 volts ac dc in. I have looked online and managed to make this schematic


but I have no idea if it works.

I was hoping if you guys know if this might work

(please don't test I just want to know if theoretical work)

Link to project.


  • this circuit should power as esp12-f and a relay and it would be nice to have some head room

  • the second regulator can handle 20v (I think) (from the nodemcu board)

  • I'm not certain that the first (7805) regulator is the best choice due to the high heat production as stated by Spehro Pefhany

  • if you know an alternative solution I'd be happy to know

sorry for any inconveniences I have caused I mainly do the software side of things

  • \$\begingroup\$ For under 500mA, I would first use a MC34063 to drop to 8V and then a 7805 \$\endgroup\$ – Indraneel Mar 24 at 1:04
  • \$\begingroup\$ I'd use one of those $1 adjustable buck modules, but I'm lazy. \$\endgroup\$ – Jasen Mar 24 at 2:39

Nobody uses linear regulators for this type of application anymore. The issue with using linear regulators with large Vin to Vout differentials, as pointed out above, is power dissipation.

A far better approach is to use a buck switcher like the LT1076 (www.analog.com). Although the device costs a bit more than a linear regulator, you can easily get 1 ampere or more from them. There is a good example schematic on the first page of the LT1076 (for example) data sheet.

A good way to think of this problem is to prefer switchers unless: 1) Currents are low (<100mA) AND 2) Vin-Vout is low (<5V)

Sure, a switching design is a little more complex and you'll need to select an appropriate inductor but the benefits of cool and efficient operation generally make it a far better choice when building non-trivial power supplies.

Linear Tech/Analog Devices offer an excellent Spice simulator you can use to evaluate any of their devices. Similar switchers are available from TI and other semiconductor manufacturers.


It will provide 3.3V out, however you have not specified the current, and the efficiency will be miserable.

For example, with 24VAC in the efficiency will be around 10% (and the input voltage will be uncomfortably close to the maximum the 7805 can withstand- a slightly higher than nominal mains voltage will stress the chip beyond the absolute maximum 35V input).

That means if you want 0.1A @ 3.3V (about 1/3 of a W) the losses will be 3W in the 7805 and 200mW in the NCP1117, so 3.2W loss for 0.33W out. That means a fairly large heatsink on the 7805 and maybe a fan.

  • \$\begingroup\$ It's also worth noting that the 1117 is picky about its output capacitor, and you can't be certain it won't oscillate if we don't know what the ESR is. \$\endgroup\$ – Hearth Mar 23 at 23:40
  • \$\begingroup\$ @Hearth Good point, the NCP1117 specifically states tantalum or ceramic "should be" adequate. \$\endgroup\$ – Spehro Pefhany Mar 23 at 23:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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