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

[Behemot]

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

Ordinary ceramic 222, just blue and slightly bigger? Got pair of them too from one PSU

This is what I call "fake" Y caps

[tleu8472]

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

Attached are pics of a CEO-250P MATX power supply. Thoughts would be welcome. It seems to be lacking vital input filtering components?

[Behemot]

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

Well, it seems to be built slightly better than most wonders claiming to be 350+ W Pi coils seem to be missing which may be a problem as for ripple being in specs. Could be solved by increasing capacitance to absolute limits (~3300 uF for D10), for this power (250 W).

Transformer is OK, rectifying diodes could be as long as they are at least 2 A rated. What is rated capacitance of those big caps on primary? What transistors are there and what rectifiers on secondary?

[ben7]

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

That is a cute little PSU there! Recap, add some EMI filtering, and it will be pretty decent!

[PeteS in CA]

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

Heatsinks look adequate, O/P inductors look reasonable. The I/P rectifiers look the same size as 1N400x series, which are 1A rated parts. Hope they're better than that! wouldn't operate that thing in the same room as a TV or AM radio! If you have decent switch transistors (probably BJTs) and O/P rectifiers, it might be worth kludging in some I/P filtering to reduce interference with other nearby electronic devices.

[lti]

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

It looks like there isn't room on the board for adequate filtering on the outputs.

[Pentium4]

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

I wouldn't trust it in a P4 computer with only 8A on the 12V

[tleu8472]

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

To add EMI filtering, can I just add X capacitors from random (junk) power supplies? I will take a closer look at the ratings of some of the components later.

I had a peek inside my Bestec 250W which is working fine except for some fan noise. I noticed a lot of the caps are not soldered on close to the PCB - the leads stand up off the board (see pic attached). Does this matter?

[Behemot]

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

To add EMI filtering, can I just add X capacitors from random (junk) power supplies? I will take a closer look at the ratings of some of the components later.

Sure, we all do that. Just use those safety-aproved X caps, not ordinary ceramics. The same for Y caps.

I had a peek inside my Bestec 250W which is working fine except for some fan noise. I noticed a lot of the caps are not soldered on close to the PCB - the leads stand up off the board (see pic attached). Does this matter?

Don't think so.

[Behemot]

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

I've blown my first PSU OCZ CoreXStream. No OCP, was trying to determine whether it has at least OPP on the PFC/PWM side. Nooooooooo

First the loader just switched off. Than I tried to turn the PSU on again with mild load only. Nothing. After a while some shitstorm came and everything in the whole room turned of Blew up at some 700 W on output

Pictures and stuff gonna come later.

[Behemot]

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

Seems that guys propagating the oidea of two bridge rectifiers in parallel not working that great were right after all. It seems it is able to increase efficiency, that's rather strange, because the power is not increasing just by adding other one.

In my case, two 30A rectifiers blown up at 41,57 A. Seems that it's short protection was PoS as well because one primary transistor seems to be shorted, one didoe in there is blown like Zeppelin from the sky. My theory is core got saturated because of improper reaction of the driver combo (CM6800TX) thus vasting transistor and diode nearby. It should go into protection mode and survive everything, which obviously did not happenned.

Also the fuse did not react, breakers one level higher upon breakers in the room directly affected sooner. Why they still use them bloody slow fuses?!

Some pics before the catastrophy. Enjoy, nobody else ever reviwed this It is also not published yet so double enjoy Worth saying I like them filtering chokes sitting on those X caps…

[tom66]

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

It's a really, really bad idea to parallel diodes due to thermal runaway; in many cases the current capacity is barely more than that of a single diode. Same for bridge rectifiers. But who the hell needs 30A going into a PSU? At nominal current of ~2Arms at 230V (460W), a bridge rectifier would contribute less than 3W loss.

[Behemot]

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

Secondary shottkies, the word "bridge" was not really accurate…

But yeah, I have seen also supplies with two bridge rectifiers on the primary side

[Wester547]

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

Same for bridge rectifiers. But who the hell needs 30A going into a PSU? At nominal current of ~2Arms at 230V (460W), a bridge rectifier would contribute less than 3W loss.

Whoa... having bridge rectifiers in parallel is a bad idea? I've seen many OEMs do it (AcBel Polytech included) and I never noticed it to cause a problem. Perhaps they need to be mounted to a heatsink at that point (so as to subdue the 'thermal runaway' and voltage drop effects it may have).

[mariushm]

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

Well, in theory it's bad because the heat generated by one bridge rectifier can heat up the other bridge rectifier

Also, there's manufacturing differences between two bridge rectifiers, however small they are.
Depending on the current going through the diodes and the temperature, the voltage drop on the diodes inside varies, hence the output of the bridge rectifiers would also vary a tiny bit, by a tenth of a volt or so, and the leads of the two bridge rectifiers would probably act in a way as balancing resistors warming up in time.

But, it may just be that the extra surface of two bridge rectifiers on a heatsink may dissipate the heat faster compared to having a single one.

At the same time, by moving less current through each bridge rectifier, the forward voltage drop of each element inside the rectifier will drop, not by much, usually it's about 0.2-0.3v difference between 1A and 4A.

So who knows, maybe that 0.5v of less voltage drop on the two rectifiers improves the efficiency of the active pfc ever so slightly, especially at 90-110v ac, enough to get the power supply to be classified as bronze or something like that.... just a hypothesis.

[Wester547]

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

So bridge rectifiers get HOT even when only a fraction of their full load is being drawn? :P That's not a surprise though since they usually aren't soldered to a heatsink (nor do they have a thermal compound) and only have a fan to remove heat from them.

[mariushm]

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

Well duh, of course they heat up, no electronic component is perfect.

0.5-0.9v drop per diode, two diodes enabled all the time in a bridge rectifier, 1v - 2v times 1a = 1-2 watts per amp dissipated as heat

The glass the diodes are inside helps increase the surface through which heat gets dissipated, so a rectifier is usually rated to work at the max current listed at 100c which rarely happens inside a power supply as the fan moves air around helping the rectifier stay cool.

[Wester547]

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

Sort of off-topic, but...

I have noticed in many schottky barrier rectifier datasheets that they are said to be rated at "Single phase half wave, 60Hz, resistive or inductive load."... though I see "half sine-wave" apply to peak currents and "60Hz" apply to cycles, and "resistive or inductive load" apply to the maximum average rectified current, I have also noticed that some sheets have "Single phase half wave, 60Hz, resistive or inductive load." added inside the average rectified current graph... does this mean schottkys are rated at 60Hz/half-wave rectification? (so one could take the rated current of the rectifier as accurate in half-wave PSUs, in that case, despite that they are usually rated at very high temperatures)

[Behemot]

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

What I have seen in most datasheets lately when writing reviews, they are rated „average rectified current“ or low frequency repetitive current (your 50/60 Hz sine or square).

But what I find more interesting, some (especially ON Semiconductor) do state, they can work normally under double the current in repetitive, usually 20 kHz mode (square). I think this practically confirms what we have been many times discussing - the single diode in rectifier being able to handle twice the current (or one diode handling power-per-device alone).

[tom66]

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

It depends on what the device is intended for.

If it's for a flyback, then the rating may well be combined (20A x 2 = rated 40A) and to achieve 40A you must use both diodes. (In this case paralelling diodes is OK because they're on the same die and very closely thermally coupled, so you limit chances of thermal runaway.)

Other times, the devices are for forward converters are the total is per diode, with a pulse rating specified for each.