repairing power supply

my friend gave me a power supply that gives +5 and +12 to portable hard drives. He said can u fix it? I had a look.

Before even opening it my guess was a bad low esr cap causing supply to die.

I opened it and there was a bulged low esr cap and two burnt out resistors on the secondary side.

On the primary side there is no controller IC just some simple 2 transistor oscillator driving the power transistor. One of these little transistors had it's face blown off :).

with such a simple design what did suprise me is there is an opto coupler across the isolation barrier bit. i don't understand how such a simple design could possibly regulate itself well.

i conclude that the crappy low esr cap got excessively leaky and drew too much current and there was nothing to protect it against over-current.

without any knowledge of what the blown components type/values are this is one for the bin.


any thoughts?
 
I've seen many computer power supplies die in the same way... resistors toasted, blown capacitors, and exploded transistors.

Not sure what causes it, but it often seems to happen without any overload. My guess is that a component failure causes the circuit to stop oscillating, sending DC into the transformer which in turn causes the transistor(s) to blow up from the excessive current, or something like that.

I've never really investigated it, since when that happens I just bin them cheap things. As for the quality industrial SMPS, well, I haven't had one die yet.

Sure enough, it is possible to create a regulated SMPS with only a few transistors. I haven't gotten into any detail about it but what it looks like is the voltage from the feedback path alters the operating frequency of the oscillator, thereby regulating the output voltage. Not the best way, but it works. I've even seen some with overload protection and a total of 5 transistors.
 

megajocke

Member
2003-01-11 8:01 pm
Works fine as long as the caps are fine. :) Most consumer products don't give the caps an easy life due to cost-cutting and failure will happen sooner or later.

This one is interesting: :eek:
http://www.electricstuff.co.uk/acadapter.html


The failure of your supply could also be because of the capacitor going open-circuit. When the adapter is unloaded the flyback voltage will increase and toast the switching transistor. This then takes out the small transistors. Source/emitter resistor on primary might be open.
 

guitar_joe

Member
2007-05-04 2:54 am
And again......

The same guy bought in a computer power supply saying it would work for a few seconds then shut down.

I suspected bad low esr caps.

The +12v supply and +5v supply caps were leaking electrolyte and bulged.

Replacing these capacitors got the 6+ year old supply going again.
 
TheMG said:
Don't you just love the abundance of poorly designed PSUs and cheap capacitors in consumer electronics?


The beauty of the ATX supply is it will sense over-current / voltage and shut down before transformer meltdown / switching device meltdown.

The downside i guess is when you open one to investigate what went wrong there are no signs usually. eg exploded components :)
 
I had to change my 1500uf caps to 3300uf to run an Athlon XP CPU stable on my motherboard.

It's an Iwill-KK266 board. Performance overclocking Athlon motherboard, but came out before the Athlon XP. Stout enough CPU power supply on board, but not near enough filtering to run Athlon XP without freezing at stock speed! :eek:

Originally had 5 1500's, Now has 8 3300's, I put 3 caps on the underside of the board :D Also added additional 2x 4700uf caps on thick wires to filter the +5V from the ATX PSU. Overkill? Nah, just extra stability. The computer hardly ever crashes now. :)

IMO the board should have had better caps from the start.

But that's just one example of many where just a few better parts would have greatly improved the design of a circuit.
 

wine&dine

Member
2006-05-18 10:16 am
ak_47_boy said:
RealVoltageRateing = ChineseVoltageRateing * 0.75


Voltage rating is hardly ever the problem - at least not in the context of motherboards and PC SMPS where plenty of evil happens at 3.3V or 5V already.

Real problem sources: Plain bad caps, caps running too hot and getting bad far too early, excessive ripple current causing excessive heating in caps, bad thermal layout subjecting caps to external heating (look around CPUs) ...

Oringinally posted by EWorkshop1708
I had to change my 1500uf caps to 3300uf to run an Athlon XP CPU stable on my motherboard.

I don't doubt it ... but I'd rather use twice as many 1500s instead to lower the individual AC currents.
 
wine&dine said:



I don't doubt it ... but I'd rather use twice as many 1500s instead to lower the individual AC currents.

Why??? :confused:

1. You can't double the amount of caps on a motherboard. Where would you put them?

2. Why lower the individual currents? If the cap can take the ripple current, then it's fine. If it can't, then it's a sucky cap anyway. More smaller caps to get the same uf isn't always better.

3. The 3300's are taller but have the same exact diameter and lead spacing as the 1500's, so they are a direct swap-in. This is why I say they should have used bigger caps in the first place.
 
TheMG said:



That should read:

RealCurrentRating = ChineseCurrentRating * 0.75

As far as PC PSUs are concerned. The great majority of them will fail in no time if continuously used at or near their max rating. Some won't even make it there to begin with.

When I first built an AMD Athlon system in 2001, I blew 3 PSU that year alone on the same computer!

250W PSU - blew and took 3GB of data with it! Sometimes showed 1/2 of video image after warm up
300W PSU - random freezing, low 5V rail (4.5V or less) Random shutoffs
400W PSU - twin fan, random shutoffs, would not stay running after 3 months, had to RMA

When I got a Quad-Fan 520W PSU with a 52A 5V rail, no more problems for years, and still going strong.

Overkill = Necessary
 
and yet again....

my computer's power supply had been misbehaving for quite some time, refusing to start up here and there.... computer locking up etc.

opened it up yesterday and found two bad low esr caps.. replaced them... power supply is good as new.

they were the only two blue coloured electros in there and they were both leaking goo... brands were fulltec low esr.