Power Supply Design

Question

I just let the smoke out of my testing setup. I designed and built a simple power supply using this guide. I built a capacitive power supply with a bridge rectifier and a final 3.3VDC voltage regulator, and it appears to have not worked. I plugged my power supply into 110 VAC and tested the voltage output at the regulator and got 3.3 VDC as expected, but there was also still a 6.3 VAC component at the output. Now, I'm not a clever man, and I assumed that this AC component wouldn't hurt anything, and boy was I wrong. I blew out my PIC32, a wifi module, and my PICkit3 -- go me.

For the curious, C1 was 4.7 uF, R1 was 4.7 ohm, C2 was 470 uF, and D1 was 6.8 V

What I'm trying to do is build a low cost, physically small power supply to step down mains power to 3.3VDC. Transformers are large and I'd like to stay away from them if possible. Does anyone have any experience with flyback transformer designs -- they seem complex?

Any advise as to what I did wrong and what I can do to fix it? Any other design ideas?

Answer 1

Do you understand that this type of power supply provides no isoation at all from the mains? It is only intended for applications in which the circuit being powered will have no connections whatsoever to anything else.

You're lucky you didn't lose your PC as well, since you were essentially connecting 120 VAC directly to a USB port! Be thankful that the PICkit sacrificed itself to save your system.

Go out right now and invest in a real bench power supply. Used ones can be had for a few dollars, and even a good low-end new one will be less than $50.

As an aside, I don't understand why that application note has you putting the dropping resistance/capacitance and the diode D2 in the neutral side of the mains connection instead of the hot side. I think it would be at least a little bit less risky to do it the other way around. But maybe they wanted to make sure failures like this one would be more memorable.

Answer 2

AndrejaKo gives good advice, but it can even more simple:

This universal AC-DC converter accepts both 110 V AC and 230 V AC, as well as a range of DC voltages (that's "high voltage": 120 V DC or higher). And it has a regulated 3.3 V output; no external regulator necessary. The modules in the referenced document are 1 W or 2 W, and are smaller than a common transformer. Efficiency is also higher than a classic power supply, and you get better specs than what the capacitive supply will give you, like for instance a 2 % line regulation.

Answer 3

Never use a non isolated power supply such as this for any consumer type circuit (anything with any access to the internal circuitry - e.g. with connections in/out), it is potentially lethal.

In this circuit, because the limiting components (R1 and C1) are on the neutral side (not a good idea IMHO), it means the full mains voltage is available at Vout. Anything that is referenced to mains earth that comes into contact with this point will have 115VAC or 230VAC across it, including yourself and your equipment.

If you are not so experienced with electronics yet, then save designing power supplies till later and invest in a decent bench supply or at least a cheap DC brick.

Another thing to be aware of is not every application note (or web source) is trustworthy, there are often mistakes and bad advice, so it's best to be able to understand it all (or at least the important points) yourself and be able to verify the correctness/safety of the source.
For an example of this (from a large and well respected company), take a look at the precursor to AN954 - TB008 is a particularly bad app note which presents a very dangerous circuit which does not meet code, has no warnings, and is apparently meant to be a safe and cheaper replacement for a transformer based power supply to power a small PIC circuit.
Here is an interesting link which discusses why it's a bad idea.