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

[RJARRRPCGP]

Originally Posted by c_hegge

Besides, that PSU's manufacturing date is 2004 (I can't remember the month, though). P4s were well and truly out by that point, and most Socket A boards had 12V connectors by that point too. 5V heavy designs really should have died off even earlier.

Asus still used +5V for their socket 462 motherboards in 2004. (A7N8X family)

[c_hegge]

Originally Posted by RJARRRPCGP

Asus still used +5V for their socket 462 motherboards in 2004. (A7N8X family)

I know, hence the reason I said most and not all. I had an A7V600-X years ago, which had no ATX12V connector. The problem was that the main 20 pin ATX connector would get very hot and it eventually discoloured the 5V pins on the conenctor - evidently, the CPU was pulling too much from the 5V rail. Typical ASUS and their cheaped out peice of junk boards...

[Pentium4]

So this thing came in a case. It's actually better than I thought it would be. Still really overrated. It uses TO-220 13009's, 6A bridge, "560µF" input caps. The transformer looks decent so it has potential. The 5V uses 30A schottky, 3.3V uses 20A schottky, and 12V uses a 20A ultra fast. There is a second spot for another rectifier at least. And....it needs one bad. I put this in a system with a Pentium D 950, and a Radeon HD 4670. With both at full load, the 12V dropped to 11.59V which made me think it was getting pushed to the max. However, at idle, the 12V only went up to 11.82V. It was pulling 265W AC from the wall when testing. All I've done so far was recap it. The fan was well oiled, and was quiet even after revving up. Not sure who the OEM is. Soldering looks good for such a low end unit though.

Need to find some MOV's for it too and it'd be a nice little unit for a Core 2 and a low end video card

Edit: Okay, I forgot to mention that I also replaced the non safety approved Y caps, and replaced those film caps with real X caps

[c_hegge]

It looks alright - for a 300W

You might want to fix up these few long component pins while you're at it too.

[Pentium4]

Yeah, after a rectifier upgrade.

You might want to fix up these few long component pins while you're at it too.

Already done, that was the "before" solder shot

[Pentium4]

Okay, I upgraded the rectifier last night. I was gonna throw on two 30A schottky packs but decided one was enough for this PSU. I replaced the SFR20S20T ( http://bg.e-neon.ru/user_img/catalog...s/sfr20s20.pdf ) with an MBR30H100CT ( http://www.onsemi.com/pub_link/Colla...30H100CT-D.PDF ) and the 12V is so much stronger and more stable......with enough on the 5V.

I stupidly tested it out with the new rectifier on a brand new Haswell system, which is probably 97% 12V based. Intel Core i5 4670, x4 DDR3 sticks, GTX 650, Intel 240GB SSD, DVD drive. Under a 125W AC load, the 12V was oscillating between 11.42V and 11.92V!

When I put it in an older machine with more 5V, AMD quad core, x2 DDR3 sticks, HD 5450, x2 spinning hard drives, DVD drive. When pulling 120W AC doing the same test, the 12V only bounces between 12.08V and 12.16V. What a difference! It seems to be running more cool now thanks to the lower forward voltage loss of the shottky

[momaka]

And this just goes to show you that you don't need a 500W+ power supply for a modern system with a high end CPU.

Where did you buy those shottkies? Mouser, Digikey? Or free samples?
I'd love to go to one of those manufacturers' factories and take a scoop of their reject high-power rectifiers. I'm sure at least half of them would be fine. Would be fun to upgrade my older PSUs that don't have a 4-pin 12V CPU connector.

[Pentium4]

Originally Posted by momaka

And this just goes to show you that you don't need a 500W+ power supply for a modern system with a high end CPU.

That's right and I was wrong, it's actually a Phenom II x6 1035T.

Where did you buy those shottkies? Mouser, Digikey? Or free samples?

I got them from Digi-Key. Pretty good price too, I got 25 of these: http://www.digikey.com/product-detai...GOS-ND/1477174

I'd love to go to one of those manufacturers' factories and take a scoop of their reject high-power rectifiers. I'm sure at least half of them would be fine. Would be fun to upgrade my older PSUs that don't have a 4-pin 12V CPU connector.

Probably. That would be pretty neat. You mean to say that you have a bunch of AT power supplies?

[Behemot]

I knew those fucking OSTs and Teapos are not to be trusted, even in those OEM Seasonics!

[dmill89]

Originally Posted by Behemot

I knew those fucking OSTs and Teapos are not to be trusted, even in those OEM Seasonics!

While they certainly can fail, the only cases where I've seen failed Teapos and OSTs in "good" (Seasonic, Hi-Pro, Delta, etc.) power supplies were in hot areas on high-hours units (generally pulls from corporate desktops that had been running 5 years 24/7).

[Behemot]

I have no idea how long it was turned on but SB rail does run 24/7 in any case and it always is passive cooling only.

[Stefan Payne]

Datecode of those transformers say that unit was made in the middle of 2007.
So it's about seven and a half years old - that's not that bad, is it?!

[Behemot]

Datecode of the capacitors and board itself (2009/25) says 2nd half of 2009. They used 2 years old crapacitors, it was probably working less than 5 years.

[momaka]

OST RLX are generally quite unreliable. RLS (black/gold, often found in PSUs) usually do better.
Also, are you sure that OST RLX was original to the PSU? The fact that it is 2 years newer and the fact that it is RLX makes me suspicious. Reason being is that RLX is uncommon in PSUs (they are close to specs to HM and are usually used in motherboards).

[goodpsusearch]

what do you know about RLP?

[momaka]

Only had one encounter in a Shuttle computer, and almost all of them were bulged - some even on non-stressful rails and away from heat sources. Just based on that, I would say no better than RLX. But I could be wrong of course - one experience is not enough to conclude, IMO.

[c_hegge]

That RLX is indeed original. I think that the PSU was made in 09 and that the transformer is made earlier. The Seasonic built Antec Earthwatts PSUs always used OST RLX in that position. Interestingly, though, I've never seen that particular cap fail before. The only one I've seen fail on the EA-380 and EA-500 is a larger RLS on the 12V output.

[Behemot]

Yep, some parts were made earlier, the OST in question is 2007, the rest was 2009 ad assembled who knows when, I cannot read much from the label/serial number. It was ditched last year when the original computer was wrote off (partially because of problems with turning on), the guy who got the computer than for parts said he was not able to turn it on at all by the time he got it. That's when I got it for repair which was done this month.

[Heihachi_73]

I think that actually is a 2007 board, it looks like a Japanese style date code on the PCB with the year hidden under that glue (2007/09/25). The ICs are also from 2007 (U101 is dated 0725).

[Behemot]

EMKO JS-300B HAZARDOUS AREA yeah baby around it when it's runnig

It actually even has full input filter with genuine Y caps and varistors+thermistor! And the transformer is not too bad. And bridge rectifier RS405. Other than that, 470uF VIVA caps, 2SC4106 switchers (may be good too), SBL1040CT for +3,3 V, FG1004 for +12 V and SB1540 for +5 V. Oh yeah 30 A my ass

Some of the secondary caps are bloated, every single cap biger than D6,3 on the AOpen motherboard it was powering was bad. And it was only with ceramic Athlon 1000 (Thunderbird) and Manli TNT2 M64. HDD has Spin retry count of 18 with big red FAILED, but runs just OK, possibly it was unable to start due to poor power from this unit.If I had nothing to do than I may renovate this, but it's waste of time with all those Fortron units I have here which are 10times better after recap.

Note that small cap with K-shaped vent under the input connector. Yes, it is our beloved Rulycon