Transformerless 5V power supply

[Patchouli]

Hey kids,

Going by the standard method of doing a transformerless supply, I get my 5V at the zener. On the scope it looks like it should with the positive and negative "camel bumps" from the rectifiers. But overall it's a stable 5V. The problem is, the PIC won't run.

A simple app - push button, LED lights. One button one LED. That's it. No shorts or solder bridges. No complex wiring. The PIC is getting power, the button pin reads 5V open / 0V shorted. The LED lights fine if I touch a 5V wire to it's pin. Nothing is wrong!!! Yet the same PIC popped out and stuck in a breadboard powered by a real regulated supply runs just fine.

Microchip themselves even suggest it's fine to run transformerless 5V supplies:
**broken link removed**

Does anyone have any experience running one this way?

 

[Nigel Goodwin]

Going by the standard method of doing a transformerless supply, I get my 5V at the zener. On the scope it looks like it should with the positive and negative "camel bumps" from the rectifiers. But overall it's a stable 5V. The problem is, the PIC won't run.

If it's got 'camel humps' how can it be 'a stable 5V', if you are measuring it with a meter it will average the voltage out so will always give an apparently stable reading. You've obviously got a scope, and you've checked it with that, and it wasn't stable - why not believe it?.

You need an electrolytic across the zener to smooth out the humps, it's value depends on the load you're drawing from it - I would suggest starting with 100uF-470uF and see what the ripple looks like then.

 

[Patchouli]

Hi Nigel,

Yes to a meter it looks perfectly stable. My best meter updates @ 250mS and it can't tell. On the scope though...

Re filter, yes that goes without saying and there is a cap! Didn't mention that. There's a 1000uF / 0.01uF pair on it. (MCLR is tied to the positive rail through a 4.7k but didn't mention that either.) Everything's correct, it should work.

It looks like the zener isn't regulating but it would read like 60V unloaded if it wasn't. Here, let me take a pic. I paralleled a 4700uF with the 1000uF with jumper wires and measured from the 4700. If it ain't smooth here, it's smooth nowhere.

Nope. Same deal. Almost half a volt p-p.

It's probably me. I re-re-re-re-checked the circuit, it's correct to every example of a transformerless supply I can find. Hmm, 78LS05 to the rescue?

Thanks, for reading.

 

[Nigel Goodwin]

What sort of load do you have on it?. Have you measured the current it's taking?.

You could try removing the load and substituting different resistors as dummy loads to see what effect the load has on the ripple.

 

[Patchouli]

Go Nigel, go Nigel, go Nigel...

the PIC is almost no load, a mA or two. Stuck a 500 ohm load resistor on. It works fine. I hate these kinds of issues. :^) So it gets a power LED, and I'll tune it's cathode resistor to be just enough to make it behave.

You are, once again, THE man.

 

[Patchouli]

follow-up.

If I want to panelize a sheet with this tiny board (18x22mm) would a load be a smart idea? Or at least pads/holes for one?

I have checked and it seems that this is unusual, that the typical transformerless circuit won't need a dummy load to make it behave. (it's going inside something, it can't acutally utilize a visual LED.)

Think this is just veroboard weirdness?

 

[Nigel Goodwin]

follow-up.

If I want to panelize a sheet with this tiny board (18x22mm) would a load be a smart idea? Or at least pads/holes for one?

I have checked and it seems that this is unusual, that the typical transformerless circuit won't need a dummy load to make it behave. (it's going inside something, it can't acutally utilize a visual LED.)

Think this is just veroboard weirdness?

No, veroboard won't make any difference at all.

 

[Oznog]

So are you taking 110V, running it through a FWB, and feeding that to a resistor+zener, and a filter cap?

Be aware that the system's "ground" is now hot. Scope probes usually have the ground connected to the third pin "ground" of the power outlet, which is incompatible with your transformerless board. This can be very dangerous at its worst.

Just how did you take that scope pic? Was it a battery powered scope?

 

[Nigel Goodwin]

So are you taking 110V, running it through a FWB, and feeding that to a resistor+zener, and a filter cap?

Be aware that the system's "ground" is now hot. Scope probes usually have the ground connected to the third pin "ground" of the power outlet, which is incompatible with your transformerless board. This can be very dangerous at its worst.

Just how did you take that scope pic? Was it a battery powered scope?

It's very common professional practice to run scopes with the earth lead disconnected - personally I would consider an earthed scope too dangerous to use. OBVIOUSLY IT MUST BE CLEARLY MARKED AS SO!.

I'm not just talking about small service departments, I'm talking about large multi-national electronics companies as well - I've been in many training rooms, it's rare to ever have an earthed scope - the only one I've ever used had been sent in from elsewhere, and actually caused substantial damage to the TV I was training on! (luckily my many years of experience prevented me from getting the 240V shock!).

Obviously it's a personal decision, some people advocate using an isolation transformer to feed the scope - but I don't really see how that is any different from using the scopes internal isolation transformer. Another option is to feed the target equipment from an isolation transformer, but this may require a substantially large (and expensive) transformer.

It's important (under any circumstances) to know exactly what you are doing - if you aren't totally confident, keep your scope earthed, and never connect it to any mains powered equipment!.