Using the experience of the E14 Community to troubleshoot a power supply

That is a classic capacitive dropper power supply which is non-isolated (so be careful while probing). As long as the casing offers double-insulation and humans aren't touching anything metal, it seems to be safe enough.

I suspect that it may be likely that something bad happened for R1 to go open, perhaps C1 had failed short in some way. R1 and R2 are likely to be same value - they are likely there as inrush current protection as the secondary capacitor can cause high current flow at initial plug-in.

If not, then it seems likely to me that U3 zener has gone bad over time - the zener forms a shunt-regulator on the output which burns energy and there are a good many Chinese wireless power switches with similar designs where the zener has overheated and cooked itself. This would then let the secondary side voltage "fly up" to sometimes in excess of 20V, killing everything downstream and consuming enough power to burn something. If it's marginal, there could be a chance that C2 has experienced overvoltages and is consuming excessive current because it is damaged. Usually it will bulge if this is the case.

Another common failure mode is with the C1 capacitor is probably a polyester film or X2 style capacitor - the film inside may have burned or vaporised through accumulated voltage transients over the years. If the capacitance value is significantly less than the nameplate rating, then the current it will let through will be too little to run the doorbell, thus the output voltage rail would be below the rating. I've had a number of capacitive dropper supplies fail in this manner, but it would not explain why R1 would have failed open

I would agree that the best way to test is to "break" things into sections, but such a supply would be unhappy without load. I would say check that the load works with a power supply attached - if it doesn't, then whatever catastrophic event happened may have caused collateral damage. If the load side is fine, then use a current limited supply set to some conservative value just above the ordinary consumption and push it above 12V. The zener stack should clamp the voltage at some value a little above the nominal operating value - if that works, then it's time to look at the capacitive dropper. I would test the full bridge rectifier diodes and if they work fine, I would test the capacitor (or replace it entirely).

- Gough

You should be able to inject a DC voltage through a current limiting resistor to any point in the circuit to test which components still work.

24V dc through a 10K resistor should work.

Thanks for the response folks. Couple that knowledge with what I read from this hackaday post https://hackaday.com/2017/04/04/the-shocking-truth-about-transformerless-power-supplies/ I'm going to be more cautious in my next approach.

The way the AC plug is inserted in the wall determining what is neutral side caused me consternation. I was using an extension cord that had the potential to create the neutral reverse scenario. I was using a VOM for measurements. I'm glad I wasn't using an oscilloscope or I may have had sparks. Well I guess I wouldn't with the current state of the circuit. It just wasn't something I was thinking of.

I wasn't aware of the no-load condition. If my understanding is correct what ever current level the circuit is designed for, that power has to be dissipated. If the load is not doing the dissipation then the remaining components will. The power has to go somewhere.

A resistance measurement at the filter capacitor on the 12V output doesn't show a short. I'm going to assume it's reasonably save to connect an alternate power source at that point. I don't have a power supply that I can control but I do have I have a 12VDC supply that has some overcurrent protection. I've seen the output voltage of the supply drop when I inadvertently applied a short. I will make sure the AC is unplugged during the testand let you know how it goes.

Thanks again for chiming in.

For home ac outlets, always consider both pins hot.

Neither is safe to touch, even if it's the genuine neutral.

Jan Cumps  Absolutely awesome advice.  While neutral may be clamped to a zero volts reference point, it is still a current carrying conductor - and current will take every possible path to earth (ground).  The question isn't if current will flow, only how much.

You've described the device and done a good job of reverse engineering it, but you haven't mentioned what your goal is. This makes a huge difference as to how best to proceed.

• If you just want to hear your doorbell again, then the best course of action would be to bin this and replace it with something safer.
• If you want to understand the device and its failure, then the reading you've done and the tips from everyone here are excellent. Probably still best to bin it and replace it.
• If you want to fix the power supply for the challenge of it, you could start by replacing some components (R1 for a start) and testing it carefully. Even if you get it working it might still be safer in the bin!

Or if the secondary side electronics work, then you could build either a traditional power supply yourself (e.g. a linear transformer, bridge rectifier, smoothing cap, LM7812 linear regulator) or cannibalise a spare 12V regulated DC switchmode supply from something else (e.g. an old external drive power supply) to make it work again safely ...

Incidentally, after my rechargeable shaver power supply failed, I lopped the leads off, terminated it in banana plugs and now I charge my shaver from my lab power supply. A fitting shave for an electronics guy .

- Gough

Your absolutely correct, what is the goal.

I seem to recall the door bell cost around $40 Canadian. What annoyed me was that it is was less than two years old.  It could have been unserviceable for weeks or months. How often in rural area are people ringing your door bell?

When shite breaks my first impulse is to assess what it takes to repair it verses the cost of purchasing a new one. When I popped the back of the unit and discovered the burn't Bakelite and the absent transformer I thought what gives. It is less than two years old. I shouldn't have to purchase another SOB!

I spent about an hour trying to resurrect the skills needed to draw a schematic of the board. 20 year they have been dormant doesn't make for easy recall. Numerous Google searches and five sheets of paper later, I had a schematic drawing. With a visit to the workbench I was hoping to find the on/off switch that was off and not on, causing the unit to stop work. I was cautious after reading the warnings.

I suggested to my spouse that we can live without a door bell. The non verbal response to the statement included a hand on a hip, a tilt of the head and an evil eye. If I don't get it working I'm know money will be spent for that chance that someone will ring.

I appreciate everyone's responses so I took the time to respond toFred27 before posting this update.

The door bell circuit worked after connecting an external PS. I did short intervals of power on and off to limit the amount of smoke I generated. When I got confident it was working on left the PS on. IT didn't shutdown but the middle zener diode released a sweat puff of acid smoke at sometime.

I shut the remote PS down quickly and waved away the ghosts rising from the burnt component. Looking at the schematic I got to thinking, what is limiting the current through the zener diodes. Let say my remote power supply is above 12VDC, those working zeners are going to try and clamp the voltage. I think I burnt the zener.

I didn't have the components in the wattage to attempt any repair on the AC power feed circuit. I also didn't want to purchase another bell ring. This one works if it had 12V.

I then rummaged through the parts bin to find a 12VDC wall wart. I removed the prongs poking through the case for the AC plug and threaded the adapter wires through the hole. An exacto knife cut the traces to isolate the zeners and all AC circuit. In the end I was left me with a working door bell that sported a wall wart.

I found another location in the house with a wall plug to hold the wall wart and conveniently tucked the bell ringing housing behind the china cabinet. The door bell worked so damn good I had to put tape over the speaker hole to dampen the ring.

I told my spouse I had to move the door bell. She was happy. I didn't tell her it now sported an adaptor to operate. I just said I moved it behind the china cabinet and left it at that.

FM, reason for effort decided. I have a working door bell once again.

Brill