PSU Repair: A Case Study

[Aris_Mp]

At least in my case most of my small devices have linear power bricks and of course I am referring to low capacity ones. Also for low watts you don't need a large transformer and huge caps. On the other hand even for low capacity SMPS you still need filtering stages and these ones contain both caps and inductors (and these increase cost of course). Now there are cases that some cheapo low capacity SMPS don't have the required filters and in this case things go south (e.g. in a 4$ power brick).

In such low wattage a difference 20-25% in efficiency isn't crucial however the EMI and ripple noise is and linear regulators are much much better than switching ones in this case. I still remember a switching power brick that managed to pass noise even through my AC source and I had to shut it down before I take any measurements.

 

[jimhood82]

Daniel, while I certainly appreciate the technical details you included, and the crudish diagrams made simple for us plebians to follow, I spent far too much of my time wondering how the heck you identified the individual power stages in there.

I doubt many of your readers are well versed in electronics this deeply, and I wonder if maybe that is something that Tom's could pick up? Namely, a series of articles that discuss not just a high level overview of "how switching power supplies work", but a more detailed guide - even project guide. Something where people build first, a small (say 5 or 12vdc) power supply, and work into something more complex - so that we could gain a solid understanding of how they work.

I realize that would be incredibly technical, and may even be out of the scope of Tom's, but I know I, for one, would be greatly interested in learning from such a series.

 

[Daniel Sauvageau]

Also for low watts you don't need a large transformer and huge caps. On the other hand even for low capacity SMPS you still need filtering stages and these ones contain both caps and inductors (and these increase cost of course).

They may require more components but the components are orders of magnitude smaller and lighter than those that would be required for a linear supply. Instead of requiring a $0.50 input and $0.20 output electrolytic capacitors, you only need $0.30 input cap and $0.01 surface mount output cap. In a flyback converter, the output filter inductor is optional if the output caps have high enough capacity with low enough ESR to bring noise down to an acceptable level.

As for the efficiency not being critical on low power adapters, it is if you want to meet the class-IV efficiency rating currently required for EnergyStar compliance: for external power supplies below 250W, it requires standby power draw at no load on external power supplies to be less than 0.5W and be more than 70% efficient for a 10W adapter. Good luck meeting those requirements with a laminated iron core transformer which likely has more than 1W worth of core losses from eddy currents, probably another 1W worth of losses from wiring resistance and a linear regulator.

Next year, EnergyStar will bump efficiency requirement to class-V. What this means for 10W adapters is that the no-load maximum power draw drops to 0.3W and the efficiency requirement increases to 73%.

In some countries like France, efficient adapters are not just a good marketing idea: they are required by law for any product that ships over 200k units.

Daniel, while I certainly appreciate the technical details you included, and the crudish diagrams made simple for us plebians to follow, I spent far too much of my time wondering how the heck you identified the individual power stages in there.

In power supplies and PCBs in general, the separation between sections is fairly obvious: there are huge (relative to general board density) gaps, traces or copper pours running between sections.

As for the PSU design walk-through or something of the sort, I thought of doing one a while ago but I have so many other projects that I forgot about it. I would need some high voltage probes and high voltage differential probes if I want to put screen caps to go along... and a handful of power supplies that showcase different implementations of individual power stages for illustration and waveform capture purposes. For this though, I would also need some proper high voltage probes, both single-ended and differential.

Any equipment manufacturers interested in sponsoring some equipment? THG.DanielS on gmail.

 

[Somasonic]

This was a great article! It was informative, as well as interesting. Personally, I've only had one PSU fail on me so far. My brother-in-law's self-built computer had a really low-end Thermaltake PSU. 800W Bronze, and oh man, did that thing go out in a blaze of glory. Very loud popping, and smoke, and funny smells. When I ripped it out of the computer, there were burns inside the case. Miraculously, the other components remained unharmed, so I slapped in a new PSU (750W Gold Seasonic) and fired it up.

Still working good, but I get black soot on my hands every time I open up that case. It's a black case, so it's hard to clean it off properly.

lol I had a thermaltake 800w do the same thing, loud pop, smoke and then the tinkle tinkle of bits of PSU falling through my case. Good times

 

[Behemot]

"Antec's manufacturer (Channelwell in this case) got the live and neutral wires backwards, which means that in the “off” position, the neutral line gets opened and everything on the primary side becomes live instead of neutral."

That's really bad--I doubt it is even legal to sell a power supply wired like that. I've never bought a CWT (Channelwell) power supply, and based on this I wouldn't buy one, except perhaps for a high end model where you might gamble that the company would exercise a bit more care.

Actually, it is not. Accoding to newer regulations it does not even matter where the phase is in your wall, in most european countries. And you probably do not even know than chinese power cords also turn phase and neutral around randomly. As PSU internals are not user serviceable, you are not expected to ever get in touch with the uninsulated part of the AC switch so it does not matter whether it stays live or not.

 

[Behemot]

Is low ESR caps wasted on normal 60hz power filtering(diode rectifier after a low voltage AC transformer)?

They usually have longer lifetime if nothing else.

I liked this article, It is important that the replacement capacitors are 105C type, as the normal 85C type capacitors would not last very long. I often repair power supplies and I use a ESR meter to test in circuit all of the electrolytic capacitors in the power supply which only takes a few minutes.

Using right type and quality capacitor may be much more important than temperature. Though secondary caps should always be 105°C.

You need to buy these and keep them on the shelf: http://www.amazon.com/125pcs-Values-Electrolytic-Capacitors-Assortment/dp/B00CWQVXAG/ref=sr_1_7?ie=UTF8&qid=1434325606&sr=8-7&keywords=capacitor+kit

WHY DO PEOPLE KEEP BUYING SUCH CRAP?! You never get quality caps in these wonderfull "kits".

Get good caps. I have nice stock of capacitors for displays, power supplies and more, some of them very custom including caps only I have on the whole planet (like Chemi-Con KZN 3300/16 D10×40 mm - the only 3300uF 16V japanese caps in 10mm diameter!). If you take at least 10 pcs directly from me (not from eBay where I also offer them), you will also get them much cheaper than from stores like Digikey, Distrelec, Farnell, TME etc.

 

[MidnightDistort]

Interesting article. I would have simply replaced the entire unit. You saw how to fix the failure, but how many units were damaged that you didn't see? A ticking time bomb that will eventually send some spike to your much more valuable hardware than a 10 year old PSU. Wasteful, yes. I get it. I don't like to waste either. And if it's on marginal hardware, fine. But on primary systems I'm not willing to take the risk. I'd rather throw away a 200$ part that has a 0.05$ repair solution, than risk frying 800$+ hardware.

Precisely why cheap PSU's should be taken off the market. I'd rather spend $500 on a PSU that will not explode and take the rest of my PC out then to cheap out and spend $20 on one that will most likely set my house on fire.

This article was interesting but if my PSU starts acting up, it'll be best to replace it. I'll replace broken fans on them and use older ones in cheap PCs but once they start making noise, bulging capacitors or simply is struggling, it's just time to replace them with a more efficient, safer PSU.

 

[Crashman]

"Antec's manufacturer (Channelwell in this case) got the live and neutral wires backwards, which means that in the “off” position, the neutral line gets opened and everything on the primary side becomes live instead of neutral."

That's really bad--I doubt it is even legal to sell a power supply wired like that. I've never bought a CWT (Channelwell) power supply, and based on this I wouldn't buy one, except perhaps for a high end model where you might gamble that the company would exercise a bit more care.

Actually, it is not. Accoding to newer regulations it does not even matter where the phase is in your wall, in most european countries. And you probably do not even know than chinese power cords also turn phase and neutral around randomly. As PSU internals are not user serviceable, you are not expected to ever get in touch with the uninsulated part of the AC switch so it does not matter whether it stays live or not.

I never thought about the likelyhood that our cords HAVE two grounds is because idiots screw up the wiring. That would explain a floating ground internally and an earth ground externally, rather than just an earth ground.

 

[Someone Somewhere]

Not necessarily just screw-ups. Random failures can happen even in the best made kit.

Using ground for return current would also lead to issues if you touched a device using large amounts of current and one that's not, because it would try to send some of the current through you to the other path, kind of like shorting earth and neutral.

 

[Daniel Sauvageau]

Using ground for return current would also lead to issues if you touched a device using large amounts of current

10 meters of #14 wiring has a resistance of about 0.08 ohm, add the appliance's power cord and we might be talking 0.1 ohm total. Even if your appliance draws a 100A inrush surge on a 15A circuit, we are still talking only 10V. People won't feel that under normal circumstances - it is no worse than licking a 9V battery to see if it is still good assuming the wiring is in good working order.

The ground wire is there to prevent the chassis from rising to high voltage in case of an electrical fault (ex.: accidentally shorting the metal case to a live surface during this story - that would have been painful had the chassis not been grounded), leakage or the neutral wire going open somewhere for whatever reason. It also diverts most fault current away from signal cable grounds.

In case of a fault to ground, the current and chassis voltage end up being the same as they would have been had the fault current been returned on the neutral wire since they usually are the same wire gauge, so the only extra safety you get here is two completely separate things having to fail before you get electrocuted instead of only one.

 

[Someone Somewhere]

Using ground for return current would also lead to issues if you touched a device using large amounts of current

10 meters of #14 wiring has a resistance of about 0.08 ohm, add the appliance's power cord and we might be talking 0.1 ohm total. Even if your appliance draws a 100A inrush surge on a 15A circuit, we are still talking only 10V. People won't feel that under normal circumstances - it is no worse than licking a 9V battery to see if it is still good assuming the wiring is in good working order.

The ground wire is there to prevent the chassis from rising to high voltage in case of an electrical fault (ex.: accidentally shorting the metal case to a live surface during this story - that would have been painful had the chassis not been grounded), leakage or the neutral wire going open somewhere for whatever reason. It also diverts most fault current away from signal cable grounds.

In case of a fault to ground, the current and chassis voltage end up being the same as they would have been had the fault current been returned on the neutral wire since they usually are the same wire gauge, so the only extra safety you get here is two completely separate things having to fail before you get electrocuted instead of only one.

Enough to trip an RCD, though.

 

[Daniel Sauvageau]

Enough to trip an RCD, though.

Maybe, maybe not.

Wet skin resistance at 10V is typically over 10k while RCD/GFCIs usually require more than 5mA to trip. It is close to the possible trip range but it should not be quite there yet without additional contributing factors.

 

[Daniel Sauvageau]

Do all power supplies have the same structure?

For the most part, yes. There are a few different topologies for the main switcher and output rectifiers but they all have some form of auxiliary supply for the 5VSB output which often gets used to provide auxiliary outputs and a handful of optocouplers for feedback. Most of the fundamental principles remain the same regardless of the specific topology being used.

Where power adapter bricks are concerned, repairs may get heavily complicated if not rendered impossible by potting.