the gutless, bloated, and fried power supply hall of shame

[Behemot]

Re: the gutless, bloated, and fried power supply hall of shame

So you're in some computer service department?

[Pentium4]

Re: the gutless, bloated, and fried power supply hall of shame

Yeah but mostly production

[cheapie]

Re: the gutless, bloated, and fried power supply hall of shame

Customers recycle their old computers and I have dibs on pretty much anything that I want, and the rest goes to a company that takes what we don't want. I reuse lots of parts and sell used computers as well

I*wish I could do that. It sounds fun.

[ben7]

Re: the gutless, bloated, and fried power supply hall of shame

Got 2 PSUs to teardown ... here is the first one.

~~~

Oh deer!

"VPower" branded DEER

Original owner said it smelled like it was burning ... I think it was burning! Surprised there are no bad caps, however I think the PSU was only being used for about a week or so...

Have fun looking at the pics!

P.S. notice the Rulycon capacitor
P.S.2. The case is almost as thin as tinfoil, it takes almost no effort to bend it!...

[ben7]

Re: the gutless, bloated, and fried power supply hall of shame

Second PSU ...

"@power" brand - not sure who made it, doesn't seem like DEER though

I believe it still works. No bulged capacitors either, and better heatsinks than the DEER. I like the bountiful holes in the back of the PSU, it certainly would be good for a project case! The metal is flimsy though ...

But, it does have the two diodes on a bracket treatment!

[Pentium4]

Re: the gutless, bloated, and fried power supply hall of shame

What are the rest of the caps in the VPower, Viva? Also curious about the fan, maybe a Foxconn fan?? I bet it could do 250W heavy on the 5V if the fan runs at full speed in a cold room and depending on the rectifiers. Rulycon

What are the markings on the main transformer of the @Power? 2 diodes on a bracket, ouch!

[ben7]

Re: the gutless, bloated, and fried power supply hall of shame

What are the rest of the caps in the VPower, Viva? Also curious about the fan, maybe a Foxconn fan?? I bet it could do 250W heavy on the 5V if the fan runs at full speed in a cold room and depending on the rectifiers. Rulycon

What are the markings on the main transformer of the @Power? 2 diodes on a bracket, ouch!

The caps in the Vpower are Viva, a few CapXon and Rulycon.

The caps in the @power are Saturn, Su'scon, and Koshin.

Not sure of the fan brand, didn't look

However, both units have their fan wired right to the 12v.

The Vpower, ugh, look at how there is only a tiny diode for the 5v standby! Plus, the biggest cap on the 5v standby is an un-vented 220uF unit. There is an LM7805 regulator on the heatsink though!

[Pentium4]

Re: the gutless, bloated, and fried power supply hall of shame

That's horrible! Thing must have high ripple on the standby. I think the @Power is a Deer as well, I've only ever seen "Julon" inductors in Deer PSU's, it's got the classic 2003 chip, and Deer loves those Yang Chun Saturn caps

[ben7]

Re: the gutless, bloated, and fried power supply hall of shame

That's horrible! Thing must have high ripple on the standby. I think the @Power is a Deer as well, I've only ever seen "Julon" inductors in Deer PSU's, it's got the classic 2003 chip, and Deer loves those Yang Chun Saturn caps

Interesting thing is, nowhere in or on the PSU does it say DEER!

Both PSUs have the EI-33 size transformers.

[Pentium4]

Re: the gutless, bloated, and fried power supply hall of shame

That is weird...I still think it's a Deer though. Aww how cute, it thinks it can do 430W peak

[Behemot]

Re: the gutless, bloated, and fried power supply hall of shame

Saturn caps We can blow you to Saturn and beyond! Isn't that great slogan?

[goontron]

Re: the gutless, bloated, and fried power supply hall of shame

saturn caps We can blow you to saturn and beyond! Isn't that great slogan?

rofl

[momaka]

Re: the gutless, bloated, and fried power supply hall of shame

The caps in the Vpower are Viva, a few CapXon and Rulycon.

Wow, that is one classic Deer! I bet the production date code indicates late 90's or possibly early 2000's. Does this look familiar? (it should... it's a L&C LC235-ATX. Posted here.)
Only a week of service? Ouch! Probably has Rulycon caps on the 5VSB critical cap spot too. Speaking on which, note the 5VSB design in this Deer - it's self-resonant without feedback as indicated by the lack of an optocoupler next to the 5VSB transformer.

That's horrible! Thing must have high ripple on the standby.

Actually, no. The 7805 regulator cleans up the 5VSB quite nicely. Probably even better than most good PSUs of today. The raw output of the 5VSB transformer is around 9V or so (IIRC) and the 7805 regulator then brings that down to 5V. Horrible for the efficiency? - absolutely! But you get clean power. And it also protects the motherboard from damage when the 5vsb's "critical" cap goes bad. Unfortunately, the PWM controller gets its power from an auxiliary winding on the 5VSB transformer too, but there is nothing to protect the PWM controller other than a simple resistor to limit the current. And those burn marks on the VPower are exactly from that. Just change the critical cap with a good one, and all should be well.
...
But then again, not like that PSU is worth fixing up. It has the "diodes-on-a-bracket" rectifier for the 12V rail and needs lots of other parts to make it complete.

I think the @Power is a Deer as well, I've only ever seen "Julon" inductors in Deer PSU's, it's got the classic 2003 chip, and Deer loves those Yang Chun Saturn caps

Yep , it's a L&C unit (a.k.a. Deer... a.k.a. Allied) - same platform as this PSU, actually (it's the B300-ATX in this post). The company that actually makes them is called Solytech. 370forlife knows a lot more about this than I do, though.

[ben7]

Re: the gutless, bloated, and fried power supply hall of shame

I just scrapped the Vpower one.

Yeah, that does look similar mariushm! The PCB on mine says rev 2.02
Yeah I noticed the lack of feedback on the standby circuit.

Surprisingly the Vpower has decent soldering, not hard to melt at all. I found a 220uF 10v rulycon hiding too. Oh yeah, and two "Jun Fu" caps hiding as well!

[Wester547]

Re: the gutless, bloated, and fried power supply hall of shame

Actually, no. The 7805 regulator cleans up the 5VSB quite nicely. Probably even better than most good PSUs of today. The raw output of the 5VSB transformer is around 9V or so (IIRC) and the 7805 regulator then brings that down to 5V. Horrible for the efficiency? - absolutely! But you get clean power.

This is sort of off-topic, but it has me wondering... I'm sure you're familiar with how many older Hipro and Delta/Newton power supplies almost always have a 7912 regulator for the -12V rail and a 7905 regulator for the even older ones with -5V rails, often soldered to a heatsink (older Sirtec PSUs also have 7905 regulators on a heatsink for the -5V rail). Do those three-terminal regulators just regulate the rail or are they actually there in order to generate and convert the voltages to negative, thusly lowering the efficiency on those rails (though they're hardly ever used if at all for anything) but for a much cleaner output? I ask because I was of the impression that those power supplies already have the correct windings in the main transformer for the -5V (if present) and -12V rails and that the +5V/+12V toroid (in group regulated designs) also regulates the -5V/-12V rails to some extent.

[Behemot]

Re: the gutless, bloated, and fried power supply hall of shame

Usually there is winding but for higher and unregualted voltage. Regulators make it quite nice with high precision and low ripple. I think many todays PSUs could have much better -12 V rail results with similar price but the rail allows +-10 % and 120 mV so who cares…well, I do

[momaka]

Re: the gutless, bloated, and fried power supply hall of shame

I just scrapped the Vpower one.

You mean you didn't pull any components off of it?? Not that there is much in there, but stilll... shame shame shame

Yeah, that does look similar mariushm!

Ummm.... wrong guy?
Okay, close enough. He's my Romanian neighbor after all

Do those three-terminal regulators just regulate the rail or are they actually there in order to generate and convert the voltages to negative, thusly lowering the efficiency on those rails (though they're hardly ever used if at all for anything) but for a much cleaner output?

Like Behemot said, those power supplies usually have unregulated -5V and -12V rails that are actually higher than -5V and -12V. Then those regulators drop the voltage down and regulate it with really good precision. Because they are linear regulators, however, the voltage they drop through them to regulate the output gets converted to heat. If there's only a low or no load on the output, then the efficiency of the -5V and -12V rails isn't really affected. However, as the load increases, those linear regulators start to dissipate a lot more heat. Hence why they are rated usually for only around 1A output. But none of today's systems use the negative voltage rails, so why PSU manufacturers don't put a regulator on the -12V rail is beyond me as well.

I think Bestec (at least their older 250W and 300W units) are the worst when it comes to -12V rail regulation, in terms of circuit design that is. They use a linear self-regulating load. Usually consists of a sub-100 Ohm resistor (usually 75 Ohms), a BJT transistor, and a 12V Zener diode. The way it works is that the transistors takes current from the -12V rail and feeds it to the resistor, which is connected to ground on one side. When the voltage on the -12V rail rises, the transistor feeds more current to the resistor in hopes of loading down the -12V rail and thus bringing the voltage down. When the voltage on the -12V rail decreases, the transistor does the opposite. But this is stupid, because it wastes a lot of power and makes the loading resistor incredibly hot. I took out this circuit on the two Bestec PSUs I have, and there's hardly any difference on the output of the -12V rail.

[Wester547]

Re: the gutless, bloated, and fried power supply hall of shame

Thanks for the explanation. I think you once found a 1000uF, 16V CapXon KM that bloated on its own in the -12V output of a Bestec... perhaps that self-regulating linear circuit is the reason? Also, I know that most of the time you will find a MOSFET (or three-terminal regulator) and a 431 shunt in a circuit that uses linear regulated DC-DC conversion. Is it possible to linear regulate a rail only using a 431 shunt or would a FET (or another three terminal regulator) of sorts also be a requirement?

[momaka]

Re: the gutless, bloated, and fried power supply hall of shame

Thanks for the explanation. I think you once found a 1000uF, 16V CapXon KM that bloated on its own in the -12V output of a Bestec... perhaps that self-regulating linear circuit is the reason?

No, CapXon are just that crappy. The -12V loading resistor was far away from that cap.

Is it possible to linear regulate a rail only using a 431 shunt or would a FET (or another three terminal regulator) of sorts also be a requirement?

For the large loads PSU rails are made to handle, you do need a MOSFET or a BJT. A 431 shunt can regulate its output only to a few milliamps. Cascading it with a FET or BJT allows the FET or BJT to handle all of the loading. Since this is a linear DC-DC regulator circuit, the voltage difference between the input and output of the FET/BJT multiplied by the current drawn from the output of the FET/BJT is the power that the FET/BJT has to dissipate. So a heavy load = lots of heat produced by the FET/BJT. A 431 shunt is too small to dissipate that much heat, even if you did attach it to a heat sink. A TO-220 FET/BJT can usually dissipate several 10s of watts with a proper heat sink.

[Wester547]

Re: the gutless, bloated, and fried power supply hall of shame

No, CapXon are just that crappy. The -12V loading resistor was far away from that cap.

I agree that CapXon is horrible (probably the worst of the commonly known brands...), but I wonder if that self-regulating circuit squandered lots of power through the capacitor (unless it's after the capacitor and not before)? Another interesting talking point was brought up by ben7 in another thread regarding power bricks... a capacitor doesn't have to be next to a hot component in order to be cooked. Heat is still conducted through the PCB substrate, PCB traces, and ambient air. So I wonder if a lot of these inefficient and shoddy circuits do play a huge part in curtailing the life of capacitors in the outputs they're connected to, such as two transistor +5VSB circuits.

It's probably more what you said, though. I've seen CapXon fail everywhere, even in very low stress applications.

For the large loads PSU rails are made to handle, you do need a MOSFET or a BJT. A 431 shunt can regulate its output only to a few milliamps. Cascading it with a FET or BJT allows the FET or BJT to handle all of the loading. Since this is a linear DC-DC regulator circuit, the voltage difference between the input and output of the FET/BJT multiplied by the current drawn from the output of the FET/BJT is the power that the FET/BJT has to dissipate. So a heavy load = lots of heat produced by the FET/BJT. A 431 shunt is too small to dissipate that much heat, even if you did attach it to a heat sink. A TO-220 FET/BJT can usually dissipate several 10s of watts with a proper heat sink.

That's what I figured. For power supplies that don't have -12V and -5V linear regulators to regulate those rails, does the main transformer already have the appropriate voltages in the windings for those outputs or do they just remain higher and unregulated? And does the +5V/+12V toroid still regulate the -5V/-12V rails at all when they use linear regulators?