[Lowfer] A power supply head scratcher

C. Turner turner at ussc.com
Sun Apr 18 16:07:36 EDT 2010


I have found SMPS wall warts to be unreliable compared to their "iron" 
relatives.  Having quite a number of them around the house and at work, 
I can say that I don't recall a single one of them lasting as long as 5 
years - some of them have failed within hours (really!) to weeks of 
first being powered up.

Almost inevitably, the culprit is electrolytics.

For best reliability, only the best-quality, 105C, Low ESR and long-life 
capacitors should EVER be used in an SMPS - assuming that the SMPS was 
engineered well in the first place.  Since the designers of the SMPS are 
probably competing on price, it is rare that one cracks a dead SMPS and 
sees capacitors with names you've ever heard of!

Were I the seller of gear that required the buying of a wall-wart to 
power my product, I would hope that I'd make at least some attempt to 
find those manufacturers that had a reasonable reputation.  I know that 
some manufacturers have had to eat a lot of their profits on warrantee 
replacements:  Not only does this cost them money off the top, but it 
can't help but harbor ill will from their customers, the majority of 
whom won't know (or care) that it was just the power supply that 
failed.  I even have seen several occasions where the wall wart supplied 
with the device wasn't even able to power that device!  In one case - an 
Ethernet switch - switch itself contained an SMPS:  With the "soft 
start" of the SMPS wall wart, the SMPS on the switch itself drew more 
current - which caused the wall wart SMPS to trip out on overcurrent, 
which caused it to restart when its voltage came up with the "soft 
start", the SMPS on the switch drew more current - The result was that 
the lights would flash a few times per second.  One fix was to plug 
output of the SMPS wall wart AFTER it had been plugged in, but that 
wouldn't have helped you with a power failure.  The "ultimate" fix was 
to just grab an iron-based wall-wart that was handy and with right voltage.

I can't help but wonder to what extent the reported power savings of 
SMPS's wall warts over iron wall warts is offset by the costs of these 
premature failures - either by the manufacturer, or the consumer who 
will often just toss the old box and get a new one - not to mention the 
"green-ness" of filling the dumps with otherwise-good products - because 
the SMPS wall wart might have saved a few bucks a year in electricity.  
(A similar debate rages over CFL's that never seem to last anywhere as 
long as they are claimed - or as the incandescent lamps that they replace.)

I've almost always been able to fix a failed SMPS by "re-capping" it.  
Why fix instead of replace?  By making sure that the highest-quality 
caps are used (Panasonic and Nichicon ultra-low ESR caps) I've been able 
to make the rebuilt power supply last longer than it did before it first 
failed - or (probably) longer than a new replacement SMPS would have 
lasted.  Anyone who made the mistake of trying it has noticed that the 
average, 85C capacitor will either fail immediately - or soon thereafter.

Another thing I've done with the SMPSs is to simply drill a few holes in 
the cases while they are open to allow some of the heat to escape.  The 
number (and location) of the holes is such that they will not pose a 
shock hazard to anyone - although there *would* likely be a problem if 
someone spilled water on them!

***

So, one or more of those SMPSs you used were probably on their last legs 
anyway.  These things will often do odd things when their output 
capacitors dry up/leak/go to high ESR, such as regulation become flaky 
or instable (either high or low voltage), be unreliable in starting, not 
tolerate even minor "brown outs", and not the least is to put out 
switch-frequency riddled voltage.  One of the most insidious effects of 
their capacitors failing is that their efficiency often drops, causing 
them to get hotter, causing the capacitors to be cooked more thoroughly, 
causing the efficiency to drop, causing them to get hotter, causing the 
capacitors to be cooked more thoroughly, causing...

Preaching to the choir, no doubt, but it's worth harping about...

Clint
KA7OEI/CT



More information about the Lowfer mailing list