henryVK wrote:

Hm, I tentatively pushed down on the top of the cap but there is no give if that's what you mean.

Yup, this is what I meant. So the cap isn't in the early stage of bulging.
These caps can dry out however, so this is no indication that it's actually good.
Let's see what happens when you use an alternate PSU for diagnosing where the computers'instable part is.

canthearu wrote:

I have 2 PSUs I want to rebuild and use: […]
Show full quote

I have 2 PSUs I want to rebuild and use:

The first:

IMG_0094.jpg

IMG_0095.jpg

This is one of these utterly generic PSUs, however, they look to be mostly intact and not price reduced or under-built for it's wattage.

Replace the dodgy bloated capacitors, it will probably work great for years.

I've gone ahead and removed the big caps and cataloged them.

4 x Fuhjyyu 1000uf caps, TNR series.
2 x Jamicon 2200uf caps, TK series.
2 x Teapo 220vf caps, SEK series.
1 x Capxon 3300vf cap, GL series.

The Jamicon and Fuhjyyu are really crap capacitors, they don't even appear to be LOW ESR, definitely cost cut. I am replacing the Fuhjyyu caps with Rubycon ZLH capacitors. The Jamicons will be replaced with Panasonic FR capacitors.
The Teapo caps I just replaced them because I have lots of rubycon YXF capacitors.
The Capxon cap is a proper LOW ESR capacitor, large, sorta expensive to replace and tested fine on my ESR meter. So I'm going to keep that one. Capxon is not a great brand, but not the worst, and it will likely be fine because it is physically quite large thus they run much cooler than smaller capacitors.

This PSU WILL be great again 😀

I have replaced the 8 capacitors identified in my last post, and the PSU is working nicely now.

What is interesting is that the Rubycon ZLH capacitors are actually significantly smaller than the Fuhjyyu capacitors I replaced. Normally it is the other way around, with the stock capacitors being physcially smaller than the replacement caps, making the job harder.

Crapxon are terrible capacitors and I always replace on sight. Only the main voltage caps 200V/400V/450V are ok.

canthearu wrote:

Yeah, you can often tell by just having a quick look in a PSU, whether a PSU is a complete gutless wonder or not. Things like missing input filters, using 4 x diodes rather than a bridge rectifier, wafer thin heatsinks, more board than component (like the L & C board).

It gets to the point you can almost smell 'em.

wave wrote:

Crapxon are terrible capacitors and I always replace on sight. Only the main voltage caps 200V/400V/450V are ok.

There are much worse IMO. I've seen plenty of devices years and years old working just fine with Capxon capacitors.

In addition, capacitor brand is only a single factor as to why a capacitor would fail. Just as important is the operating environment and device design. A bad device design can cause ANY capacitor to fail.

The capxon capacitor I left in that PSU tested perfectly fine using ESR tester, and due to it's large diameter construction, has a much longer lifetime than the smaller capacitors I replaced. I'm comfortable enough for now that it doesn't need replacing.

canthearu wrote:

Yeah, you can often tell by just having a quick look in a PSU, whether a PSU is a complete gutless wonder or not. Things like missing input filters, using 4 x diodes rather than a bridge rectifier, wafer thin heatsinks, more board than component (like the L & C board).

I mainly agree with you, but I think not using an encased bridge rectifier is okay if well-done.
A look into my 1994 Siemens-Nixdorf 145W AT PSU:

4 big diodes of good quality with thick leads.
Placed in front of the outlet fan, with only the thermistor between diodes and fan. (Active cooling and thermal surveillance.)

No low quality components detected. Notice the size of the coils and the transformers.

There is even more on the PCB underside.

Some more details:

Notice the copper heat spreaders between semiconductors and insulation.

Safety-certified optocouplers.

Thermally conductive brass screw used to connect the thermistor to the cooler.

I have seen bigger output sections with more pi coils, but, well, it's only 145W...

And finally the label...
(SNI = Siemens Nixdorf)

So I guess, 4 diodes instead of bridge rectifier block does not necessarily indicate that the PSU is of inferior quality...

retardware wrote:

So I guess, 4 diodes instead of bridge rectifier block does not necessarily indicate that the PSU is of inferior quality...

That is fair enough. It is a small PSU, with otherwise good components and accurate specifications on the label.

It isn't a 300W or higher PSU with tiny rectifying diodes,

Okay, boys and girls:

I got an ATX PSU for testing purposes which turned out DOA. I tried to jump start it but it just sits there.

Now, this made me realize that I don't quite know how best to... ahem... power up this computer, as the motherboard (MS 5129) doesn't seem to have a connector for a separate power switch. Is it okay to just permanently connect the 4th and 5th upper pins on a ATX PSU and then use the PSUs I/O rocker switch to turn a computer on and off, or is that bad form?

Well, this is how I turn my ATX PSU on and off since I like to make sure it's off (and not soft-off) when I install or remove stuff from mobos. I've yet ecounter any problems with this setup, short maybe of premature wear on the PSU switch. But that can be easily and cheaply replaced, unlike some of the 386/486 stuff I have.

No problem leaving an ATX power supply permanently on, then use the PSU rocker switch to turn your ATX power supply on and off, to power your AT motherboard.

I opened up a fairly "light" psu today. I am interested if this psu is usable for light loads in P3/ early P4 use without having to fear that it might kill off parts of the pc it is in.
There was a handwritten note on it with "Bestec ATX" on it, but I googled it and they look kinda alive but I could not find an exactly identical looking psu from them.
See the below pictures, first 5 in this post the rest in the next one:

The remaining 2 pictures:

If there is something that needs to have another picture taken to be clear then let me know.
Only other observation I did was that the big caps are Fuhjyyu caps and as far as I know they aren't a good quality cap manufacturer

Please make sure this is not the bestec model that kills mobos, having 5vsb going as high as 13V!

wave wrote on 2020-02-23, 00:50:

Please make sure this is not the bestec model that kills mobos, having 5vsb going as high as 13V!

I suspect you mean the ATX-300-12E, and this isn't one of those.

I can't say an orphaned PSU with no label fills me with confidence, but I don't see anything alarming in the pics.

pewpewpew wrote on 2020-02-23, 01:19:

I suspect you mean the ATX-300-12E, and this isn't one of those.

The ATX-250-12E was more notorious.

So it wont kill my boards? And is it really 300w capable? The 240w hp psu that I cleaned and gave a new fan is alot more loaded with (bigger) parts. That is what worries me most.

Judging from the two primary 470uF fuhjyyu capacitors (bad brand), the EI33 type primary transformer and that 16 legged pwm ic on the secondary this is a 250W max capable half bridge topology unit.
The 5V standby circuitry is not inspiring confidence. Very similir to those in the notorius Bestec units. Looks like a small signal transistor is controlling the high voltage BJT on the heatsink. There are several critical capacitors around there that could cause the voltage regulaton to go bad when these capacitors values deteriorate.
I think you should change all the caps, if you want to use this psu for a P3/light P4 era computer.

canthearu wrote on 2019-03-26, 08:58:

Yeah, you can often tell by just having a quick look in a PSU, whether a PSU is a complete gutless wonder or not. Things like missing input filters, using 4 x diodes rather than a bridge rectifier, wafer thin heatsinks, more board than component (like the L & C board).

A bridge rectifier is essentially 4 diodes. Does the packaging matter so much?

RacoonRider wrote on 2020-03-02, 07:02:

A bridge rectifier is essentially 4 diodes. Does the packaging matter so much? https://upload.wikimedia.org/wikipedia/commons/3/ […]
Show full quote

canthearu wrote on 2019-03-26, 08:58:

Yeah, you can often tell by just having a quick look in a PSU, whether a PSU is a complete gutless wonder or not. Things like missing input filters, using 4 x diodes rather than a bridge rectifier, wafer thin heatsinks, more board than component (like the L & C board).

A bridge rectifier is essentially 4 diodes. Does the packaging matter so much?

In theory it doesn't matter.

In practice, bridge rectifiers use typically bigger diodes internally and the form factor allows for better heat dissipation and more current carrying than using general purpose diodes in a bridge configuration.

Using discrete diodes instead of a single component bridge rectifier is a cost cutting measure typically. Doesn't automatically mean the PSU is a dangerous misrated gutless wonder, but it normally isn't a good sign.