We have designed an "always ON" electronic home automation device which goes behind the switchboard (inside the wall.) This is the first commercial product that I am designing, so bear with me for obvious things.

The total maximum consumption of the device is 600-800 mA. It has an ESP32 module as a controller along with four ATmega328 microcontrollers and around six relays connected outside or away from this device.

This device is powered from 2 A, 5 V i.e 10 W AC to DC converter from one of the reputed brands from China - UA10-220S05P2 (10 W, 5 V) datasheet. In order to reduce the size of this converter, we have removed the outer plastic casing of the AC to DC converter (a picture of this is attached.)

enter image description here

enter image description here

After testing this device for around a year and a half continuously without turning AC to DC converter off, we are facing issues with AC to DC converters. When we opened and saw, it is found that either one and/or both, 400 V 10 uF electrolytic capacitors which are used near rectifier has an open vent and oil leakage from them.

We have tested this on 4 to 5 devices. After approximately 1 year to 1.5 years all the device's AC to DC converters stopped working and has the same issue. To confirm that it is problem only with the suspected capacitor, we removed and replaced this capacitor and the device is working again.

Our guess is that the long-term issue is caused by heat build-up inside the product enclosure, degrading AC/DC converter performance. We tried using an aluminium plate around the AC/DC converter as a heat sink, but I'm not sure it works better.

Any ideas/suggestions/feedback on:

  1. What could be the issues that could be cause it? Any known-standard solution/ideas you could share for it.
  2. How to conveniently to test such long term issues? How can these be emulated so that we know the proposed solution is working?
  3. Was removing the plastic casing a bad idea? When we tried aluminum shield (which should be better than plastic case - in terms of heat dissipation,) we found little difference (temperature on and around AC/DC converter inside our product) when compared with the one without case.
  • \$\begingroup\$ Try to find capacitors with a bigger "long" life, especially when there is local "heat". xppower.com/resources/blog/… \$\endgroup\$
    – Antonio51
    May 25, 2022 at 8:00
  • 8
    \$\begingroup\$ You can tell how reputable the supplier is from the fact that they advertise up to 305 AC / 430 DC on the data sheet, yet use 400 V capacitors. For any equipment, you need components rated for the input rating. For 'always on' operation, you need a large standoff between rated and operating voltage. You need a better PSU supplier. For accelerated testing, you raise the temperature, see Arrhneius. They don't specify lifetime. They do specify use in free air. They say they'll do custom, ask them for better voltage/lifetime rating on the caps. \$\endgroup\$
    – Neil_UK
    May 25, 2022 at 8:06
  • 1
    \$\begingroup\$ On first glance I'd say it's a ptoblem with heat and probably cheap capacitors. You're running the system in an enclosed space which could have quite high ambient temperature. For your usecase there is probably no way around getting an open frame (or closed frame) power supply that is rated for industrial temperature ranges. \$\endgroup\$
    – kruemi
    May 25, 2022 at 8:06
  • 2
    \$\begingroup\$ Welcome! Classic issue. Bad electrolytic capacitors have been a plague of the electronics industry for the last 50 years. What's your average load? How hot does it run? \$\endgroup\$
    – winny
    May 25, 2022 at 8:18

3 Answers 3



  1. Power Supply run outside specifications
  2. Power supply components do not meet stated specifications


Commonly accelerated aging can be achieved by running stuff at higher temperatures or thermal cycling. But it takes a lot of experience to create a protocol and translate the result you get to the real world.

Removing the case: Removing the case was a bad idea. Just because it's generaly a bad idea to remove cases. Most probably it did not contribute to the issues you're experiencing. The issues you see are because those parts are made to be cheap not made to last.


Get a power supply that has specified temperature ranges and failure rates documented.

From the top of my head (and in no way meant as advertisement) Mean Well, XP Power, Delta are known brands with a wide range of options available. Those Supplies can fail as well but normally they tell you what the probability of them is to fail at which lifetime.

It's a good idea to overspec the power supply a bit. But not too much because they tend to get less efficient when the load is too low. Also consider the thermal envoirment of your system. It's running in an enclosed space with little to no ventilation together with other heat sources. Monitor the temperature of the system and issue a warning if it's getting too warm (and shut down before anything breaks).


This picture can help you understand why temperature and other stuff can be important.
From this link.
You can see that lowering temperature from 105 °C to 55 °C may extend the lifetime by x30 time. (150000 hours -> ~ 18 years)

Some data as this can be given for capacitors link

In any case, for such "closed" systems, you should choose with some "better" specs".enter image description here


As heat is an issue, worth noting that 10W is a heck of a lot for an "always on" device nowadays, and about as much as my entire home overnight except when the fridge kicks on every 3 hours.

Measure the actual power consumption from the supply in various conditions, and be clear about where that power is going. You may identify some major savings which will then reduce PSU heating and extend its life, as well as saving money and the planet.

For example : you may find you can clean up 10 to 20mA here and there with the relays off - or replace certain mechanical relays switching low voltage with more efficient solid state relays.

If a specific relay is powered most of the time, you may find a latching relay which only needs power when changing state.

Or you may find a converter that is considerably more efficient at light load, and this may reduce mains power when the rest of the device is asleep.

I should add, this does not replace the other answers, but adds a different perspective.


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.