1
\$\begingroup\$

The power supply of my home lighting keeps failing. So far I have had ten identical power supplies fail. Upon inspection the thing that is always damaged is the electrolytic capacitor after the bridge rectifier (the power supply has no step-down transformer). I have always fixed the faulty capacitors by replacing them with new ones.

Aside from the power supply just being badly designed, I can only think of two reasons why they are failing: from getting too hot, or from surge voltages.

I can't really fix the temperature problem with how things are; the only thing I can do is buy electrolytic capacitors that are rated to survive in that environment; 10000 hours @ 100°C and from surge voltages that exceed the capacitor's voltage rating. I'm pretty sure they are using very cheap capacitors - that's why they are failing so fast.

To solve the surge problem I am thinking of placing a MOV on the AC input of the circuit so that the capacitor and the bridge rectifier won't be exposed to surges. The original circuit does not have any protection against surges. I was wondering if my adding MOVs to the circuit can cause problems, because I am essentially modifying the circuit. I don't think there will be, but I am just making sure.

\$\endgroup\$
16
  • 3
    \$\begingroup\$ MOVs can only withstand a limited number of cycles. \$\endgroup\$
    – DKNguyen
    Commented Jan 22, 2022 at 21:19
  • \$\begingroup\$ @DKNguyen that is interesting because datasheets (the brands i have looked at) does not mention life-cycles. By limited about how many are we talking here? \$\endgroup\$
    – DrakeJest
    Commented Jan 22, 2022 at 21:36
  • 1
    \$\begingroup\$ @Kartman To be fair, with gross overload most things explode. \$\endgroup\$
    – DKNguyen
    Commented Jan 22, 2022 at 23:57
  • 3
    \$\begingroup\$ @DKNguyen - have you seen a MOV explode? It’s not pretty. You’ll find that many manufacturers will put a heat-shrink or similar shroud around them to contain it. Nevertheless the OP is mistaken in that adding a MOV will cure his problem - it won’t. The impedance of mains surges can be very low, so the average MOV can’t soak it up and would blow the fuse, explode or both. \$\endgroup\$
    – Kartman
    Commented Jan 23, 2022 at 2:52
  • 2
    \$\begingroup\$ @DKNguyen MOVs degrade with each overvoltage event they clamp, and their withstand voltage may come down and leakage current can go up, so thermal runaway is a common problem, and both Bourns and Littelfuse MOV application notes do mention that MOVs can fail short. \$\endgroup\$
    – Justme
    Commented Jan 23, 2022 at 9:37

3 Answers 3

4
\$\begingroup\$

From Littelfuse, here is an estimate of MOV lifetime vs. energy absorbed.

enter image description here

In my experience they always fail shorted until they burn enough to physically come apart.

\$\endgroup\$
2
  • \$\begingroup\$ I think its also for the best it fails short, as that would indicate the circuit is exposed extreme voltages. And as mentioned by Autistic, MOVs used in input protection are usually paired with fuses . \$\endgroup\$
    – DrakeJest
    Commented Apr 14, 2023 at 21:14
  • 3
    \$\begingroup\$ @DrakeJest I disagree (that it's for the best), but we can't pick and choose how solid-state physics works. Semiconductors tend to fail shorted unless physically damaged. A fuse (or equivalent such as fusible resistor) would be mandatory for safety approvals, but not all stuff on the market meets even the minimum standards. \$\endgroup\$ Commented Apr 14, 2023 at 21:22
2
\$\begingroup\$

So far I have 10 of the same power supply failed. Upon inspection the thing that is always get damaged is the Electrolytic Capacitor after the bridge rectifier (the power supply has no step down transformer). I have always fixed the faulty capacitor by replacing them with a new one.

We'll need to see a legible picture of the front and back of the PCB with the capacitor on it, as well as the ratings of the capacitor you have used as a replacement.

There are billions of devices out there that have bulk DC capacitors attached to line rectifiers. It's not very hard to get them to last long enough for the capacitor to fail from "old age".

If you buy the capacitor from reliable sources, e.g. DigiKey, Mouser, etc. but not Aliexpress or eBay, then if the capacitor is still failing then you've not used a high enough rating. Rectified 240V mains at high line is about 370VDC peak. You need 450V capacitor, and its ripple rating must not be exceeded. You will have to measure the ripple current through a replacement capacitor, and consult the datasheet to make sure you stay well within the specs. Typically some derating would be applied, so e.g. you'd want to stay below 1/2-1/3rd of the rated ripple current for longevity.

The hours rating on a capacitor only applies when the capacitor is used within its ratings.

You've assumed that the capacitors fail due to overvoltage, and that may well be the case, but it'd be prudent to eliminate more mundane causes of trouble:

  1. Exceeding ripple current rating.
  2. Exceeding DC working voltage rating in high line condition.
  3. Forgetting that the full bridge rectifier output voltage is 1.4 times the RMS sine line voltage, e.g. a 350V voltage rating is insufficient, and a 400V rating would leave almost no margin.

In cases of under-rated capacitors, the replacement may be physically bigger or use a different package, so the replacement may need to be mounted outside of the PCB, connected by short lengths of hookup wire, and a bigger case would need to be used to house the PCB and the new capacitor.

\$\endgroup\$
2
  • 2
    \$\begingroup\$ Hi ! Wow its been a year already since i asked this question, Im proud to say that non of the "upgraded" power supply failed after about 1 year and 1 month. I did get high tier capacitors 450V rated rubycons. As well i placed 10nF caps on the input AC (forgot the reason). No MOVs were placed. \$\endgroup\$
    – DrakeJest
    Commented Apr 14, 2023 at 21:10
  • 2
    \$\begingroup\$ Pretty good then! I'm glad that quality capacitors have solved the problem. 10nF across the AC input shunts high slew rate spikes without letting them pollute the rest of the circuit, so that's a good thing. It also prevents the trace inductances from ringing due to those spikes. \$\endgroup\$ Commented Apr 14, 2023 at 22:06
2
\$\begingroup\$

Good to ask. MOVs may not solve your problem with power supply failure. Do be careful, because some MOVs fail into a low resistance state and can get very hot; they can even explode. A metal box is good. MOV after fuse is good. MOV on separate fuse is good. MOV close to delicate SMD parts is bad.

\$\endgroup\$

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.