8
\$\begingroup\$

I have an AC/DC power supply circuit fitted to many machines, which are all running well within the rated power capability of the supplies. However these keep failing, every one burning out a (large) electrolytic capacitor (after say 1 years' service). Oddly these boards vary slightly, some have one cap, some 2, but invariably this is the part(s) that fails.

The boards with 1 cap are rated at 22µF 450v, and with 2 rated at 47µF 450v.

So, what can i do to take the stress off this component? run several in parallel?

UPDATE: Having had a good look over the PSU's they are indeed weird! the motors in the machines (which are cardboard shredders) are rated at 220v DC! so the main "chunky" part of the PSU board is just a rectifier/filter cap (which is the bit burning out) with no transformer, and the rest is a smaller, separate step-down DC supply for things like lights, switching, protection etc.

\$\endgroup\$
3
  • \$\begingroup\$ what is the power rating of the supply(current)? \$\endgroup\$
    – yogece
    Commented Oct 21, 2013 at 14:09
  • 4
    \$\begingroup\$ You must dig into and understand the real reason why the capacitors keep burning. Is this related to heat? over voltage? Too much current constantly being pulled from the capacitors? Low quality capacitors? Without understanding you will not solve the actual issue, just patch it and have it happen again. Take an oscilloscope and look at a running system. \$\endgroup\$ Commented Oct 21, 2013 at 14:57
  • \$\begingroup\$ Yes you are right, its very subjective. The actual reason seems to be heat build-up, but what is causing this is the question. I have replaced the cheap-o Chinese wing-chun whatever caps in the past with good Panasonic ones, but they still warm up quite quickly. Yes i need to look at it running really... \$\endgroup\$ Commented Oct 21, 2013 at 15:00

4 Answers 4

13
\$\begingroup\$

The typical universal input range switching power supply uses a circuit topology much like shown here:

enter image description here

The larger valued high voltage capacitor (highlighted in yellow) has the job of smoothing the rectified DC voltage from the AC power line.

There are two main failure modes for this capacitor. One is high voltage spikes at the input of the supply that make it in through the common mode choke. Spikes in excess of the capacitor voltage rating can cause damage to the insulating dielectric layer of the capacitor leading to internal shorts. High voltage problems should best be solved by finding the source of such spikes in the power system and taking steps to clamp spikes where they are generated. It can also help to improve the input filter of the power supply, add transient absorbtion devices such as transzorbs or MOVs to the input section and increase the voltage rating of the large capacitor.

Another failure mode is the internal heating that can occur when current changes in the capacitor reacting with the series resistance (ESR) of the capacitor. This generates heat that can dry out the internal electrolytic materials in the capacitor which causes a decrease in the capacitance. It can also increase the series resistance thus causing additional heating to occur. In the off-line type power supply this capacitor is working at twice the line frequency and the current pulses in the capacitor (known as ripple current) occur as the capacitor is charged on each half cycle and discharged as the rectified AC voltage goes to zero whilst the capacitor is asked to continue to supply current to the output sections of the supply. Several factors can be considered to increase the reliability of this capacitor. First and foremost is to calculate or measure the ripple current during supply operation and make sure it is well within the ripple current range for the selected capacitor. Another parameter to consider is the temperature rating of the capacitor to ensure that it higher than the operating temperature of the power supply. Finally it can be useful to to find capacitors that have a lower ESR value which will lower internal heating for a given amplitude of ripple current.

\$\endgroup\$
2
  • 3
    \$\begingroup\$ You forgot about the high temperature because of heating from the huge heatsink near the capacitor - bad design or planned obsolescence. \$\endgroup\$
    – johnfound
    Commented Oct 21, 2013 at 15:38
  • 1
    \$\begingroup\$ I, too, would bet that the issue is total ripple current in the OP's supplies. In your particular example, you need to consider both the 120-Hz ripple from the line and the high-frequency ripple from the switching regulator. \$\endgroup\$
    – Dave Tweed
    Commented Oct 21, 2013 at 16:38
6
\$\begingroup\$

Using two capacitors in parallel will be equivalent to using two resistors in series, the capacitance will add up, and it may adversely effect the operation of the circuit (it may not, but thinking worst case here.)

A better way to relieve stress from these components, without any knowledge of the circuit itself would be to use a capacitor with a higher voltage rating. This will translate to a larger, more expensive component but it will stress less. Another way would be to get a "quality" capacitor, from a brand name supplier. There are audiophile capacitors which are known for their quality, but I am not familiar with the exact brand names.

\$\endgroup\$
2
  • \$\begingroup\$ Are audiophile capacitors known for performance in power supplies where they may be constantly charged and discharged, with a large current flowing? \$\endgroup\$
    – Kaz
    Commented Oct 21, 2013 at 19:54
  • \$\begingroup\$ @Kaz - no, they're known for being overpriced bull. \$\endgroup\$
    – John U
    Commented Oct 22, 2013 at 9:43
4
\$\begingroup\$

If it'd be possible (given the size constrains that you have), I'd de-rate your capacitor (use a higher voltage rating than required) and also put a smaller ceramic capacitor in parallel. These are more tolerant to short high-voltage spikes and will help reduce the stress on the electrolytic.

Sparkfun recently ran a video related to this, I'd recommend you to check it out.

\$\endgroup\$
0
\$\begingroup\$

IF they are burning up quickly, there is surely something wrong with the circuit. caps do die eventually, but they should last years if everything is designed right. That said, a higher voltage High quality cap will be less likely to blow. depending on the purpose, you could even put one with higher capacitance to avoid it discharging and recharging fully quite so much. That is only a good idea with caps meant to smooth out the power supply, or something else where the actual capacitance is not very important.

\$\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.