Javascript Menu by


 Repairing Switched Mode Power Supplies by Erich VK5HSE

WARNING: This project may involve high voltages or temperatures  that may cause injury or even be fatal and should only be undertaken if you have the experience or training to do so and only if it is legal in your locality. This also applies to the use of this equipment by yourself or others. Before undetaking this project please refer to our Disclaimer. You have been warned!

Switched mode power supplies work by rectifying alternating mains current and then driving a high frequency transformer with pulses of this rectified current.

The high frequency transformer produces a high frequency output which is then rectified and smoothed by filter capacitors.

The advantages of this approach are that a small, lightweight transformer can be used instead of a big iron cored mains frequency transformer, and very little standby current is required when no load is being drawn.

Two of the major downsides are that the high frequency pulses of current are rich in harmonics which can produce broadband RF noise, and that the ripple in the DC output of the high frequency transformer output can cause heating in the smoothing capacitors used to condition the DC output before it makes it way to the electronics being powered.

In addition to the ripple currents, switch mode power supplies are often very compactly constructed, and sometimes lack airflow, causing even more heating of the components.

Failure of the smoothing capacitors is a common mode of failure in the switch mode power supplies which are becoming almost ubiquitous in consumer electronics, computers and radios.

This short tutorial documents the process of identifying and replacing smoothing capacitors which have failed in service in a switch mode power supply. This is by no means a comprehensive guide to repairing power supplies, but it will give you enough information to resurrect a lot of the failed switched mode power supplies you come across.

This is the device being repaired, a DVD HDD recorder which was looking very lonely one night at an Adelaide Hills Amateur Radio Society meeting.



Usually, you won't be lucky enough to have a label indicating the fault on the device which has suddenly failed. Usually, it will simply show no signs of life when turned on.

Open opening the lid, the power supply board is quite obvious in the upper left hand portion. It has the mains lead going to it, quite a few heatsinks for the voltage regulators, some ferrite cored transformers, various diodes, and lots of electrolytic capacitors. The rest of the boards are full of smaller surface mount devices, crystals and smaller wires look are more reminiscent of computer motherboards. Don't do this on a bedspread like I did unless you want to risk damage from electrostatic discharge. I like to live dangerously.


Switched mode power supplies can harbour lethal DC voltages in excess of 300 volts. Don't poke around a recently powered up power supply unless you know how to safely discharge the electrolytics, and even then, it is safer and easier to leave it a day and come back to it. And whatever you do, don't work on the device with the power lead plugged in, as this is just asking to be electrocuted. For this reason I don't recommend the “insulated screwdriver touching the top of the suspect capacitor and holding your ear to the handle of of the screw driver to listen for clicks suggestive of capacitor failure with the power supply running” technique either.

Note that this manufacturer has not skimped on a power switch on the front panel which forms a switch loop that actually disconnects power from the power supply board. Some manufacturers do not even have this and the switched mode power supply runs all the time, thereby increasing the heat stress on components. This was seen in a poorly ventilated TEAC digital set top box which had died with failed smoothing capacitors. It drew 7 watts even when “asleep”!



After removal of the board from the chassis, it can be inspected visually for obvious problems. The fuse in the black rectangle on the lower left of the power supply board was intact. On detailed inspection, a 2200uF 10V low ESR capacitor has an obvious bulge.



Other things to look for include brown spots on circuit boards where components have overheated. This can be seen under the red zener diode with the white stripe in the left of the above photo, right next to a 220uF 16V electrolytic. Having gone to the effort of dismantling the unit, it pays to make sure nothing else warrants replacement.

A particularly helpful test instrument for checking normal looking electrolytics is an ESR meter, where ESR stands for equivalent series resistance. This ESR meter was built from a kit.

A capacitor, at is simplest, is a pair of plates separated by a dielectric. If the composition of the plates or their surfaces change, it can increase the resistance to current passing through the capacitor, which in turn can lead to increased heat dissipation in the capacitor, which increases resistance even further, and so it goes, slowly cooking the electrolyte in the capacitor. The ripple currents in switched mode power supplies can be significant, so for this reason, low ESR capacitors are used. Unfortunately, even the low ESR capacitors rated to 105 degrees in this unit can fail in service when near heatsinks giving off heat, and in poorly ventilated enclosures.

Of use to us is the fact that failing or failed electrolytic capacitors will increase in resistance, and by checking the ESR of the capacitors, we can sometimes identify normal looking but faulty electrolytic capacitors.



On testing, the bulging 2200uF 10V electrolytic is found to have an ESR of 2 ohms. This is much more than the 0.1 or less ohms we would expect of a new electrolytic of the same value and voltage rating.

Checking the other capacitors, three 220uF 16V electrolytics were also found to have higher than expected ESRs, and were also marked for replacement. One of these was next to the brown spot under the zener diode shown above.

In the absence of an obviously defective electrolytic, and in the absence of an ESR meter, one can start troubleshooting of switched mode power supplies by replacing all the electrolytics near heat sources, and also checking diodes and resistors with a multimeter, but this is a bit of a shotgun approach. Sometimes failed semiconductors will show very obvious signs of failure.

Based on the working hypothesis that the failure was due to the electrolytics, new capacitors were purchased.



The suspect capacitors were then desoldered with a soldering iron and desoldering braid. I have read that copper braid from coaxial cable can be used a desoldering braid substitute, but I have not tried it. The picture shows the 2200uF capacitor being desoldered, and the other three capacitors have been identified and marked for removal.



The new capacitors are then installed



The board was then put back in the chassis.


The unit was then fired up. Ideally, this should be done with the lid back in place to reduce the risk of electrocution, and a less expensive television.



Of interest, the fan was groaning a bit on powering up the unit, which may explain how the power supply died in the first place. A lack of air flow might have caused the overheating and capacitor failure. The fan will be replaced.

Of course, if all of the above fails to work, one must go back to the beginning and look for other obvious or not so obvious component failures, or perhaps cold solder joints. If it all becomes too hard or potentially too expensive, salvage the good bits and throw out the rest.

Good luck with your dead electronics!