gdjacobs wrote:

Simply because of the constant load, yes?

Mostly. Constant load (whenever PSU is switched on) and no fan (whenever PC is turned off), so +5VSB caps have to rely on whatever passive cooling is available inside the PSU for most of their lifetime (unless comp runs 24/7).

A larger can also means lower ESR, which doesn't harm for the output of a flyback converter like the typical +5VSB circuit.

Another example of cap oversizing for +5VSB reliability:

Another nice trick I've spotted on some Delta-made PSUs:

^ That S-shaped aluminium thing right above the white glue blob is a heatsink on top the +5VSB PWM controller. Its lower leg goes thru the PCB and contacts the ground plane for optimum heat sinking.

TELVM wrote:

A larger can also means lower ESR, which doesn't harm for the output of a flyback converter like the typical +5VSB circuit.

In theory lower ESR is nice, but it can play havoc with SMPS stability. For the coarse response characteristics required to initially filter a switch mode buck-boost converter, though, I suspect it will be fine. Caps in more sensitive areas for the feedback circuit would be another story, though.

I'm kind of surprised nobody includes TVS diodes or other forms of protection on the output to guard against cap failure after the flyback. I mean, caps are a perishable item, so this is a predictable problem. Weird!

TELVM wrote:

Another example of cap oversizing for +5VSB reliability: […]
Show full quote

Another example of cap oversizing for +5VSB reliability:

Another nice trick I've spotted on some Delta-made PSUs:

^ That S-shaped aluminium thing right above the white glue blob is a heatsink on top the +5VSB PWM controller. Its lower leg goes thru the PCB and contacts the ground plane for optimum heat sinking.

Delta Electronics. Guilding the lily since 1971.

I'm sure we had another thread on this topic...

After the picture showing the nice Delta unit we need some images showing some real filth!

Let me present the COLORSit Model 340U-SCE... 420W PSU 🤣

This PSU will power my year 2000 Thunderbird build, if it works that is. This unit is a gutless wonder without much substance, its weight is comparable to a pack of cigarettes. It won't have to work very hard though, the system will need about 25 amps (peak) on the 5V rail but not much else and the PSU will not have to suffer through a total load of over 100W for any lenghts of time.

I really like that the model name says outright that it's a 340W unit and that the 420W max load specification is written with red text like a warning... Please do not try to load this PSU with 420W...

There is not much substance here...

All caps are "Fulltec", it's probabably just is another way of saying Fuhjyyu/FuckYou.

Perhaps someone who actually knows something about PSUs can spot something interesting? I can at least not see any bulging caps or the like so I will soon do a smoke test!

Edit

3.3V and 5V looks good but the 12V rail is hovering around 12.8V...

My test rig consists of a Slot-1 motherboard with video card, memory and CPU and also a DVD drive and a HDD to load the 12V rail so it's pretty representative for the actual system even if the test system pulls less amps from the 5V rail.

I will let it run for a while so the FuckYous get a chance to reform...

/Edit

Home made full wave rectifiers with quad diodes are always a bad sign.

gdjacobs wrote:

Home made full wave rectifiers with quad diodes are always a bad sign.

We will see if it survives! 😁

The all important +5VSB standby rail seems to be well regulated... for now. The other voltages also stays within spec even if the 12V rail still reads a bit high but it at least better than when I first powered on the unit for (what I think is) the first time in many years.

The voltages as read by my Abit BF6, my multimeter agrees when it comes to the major rails so I think we can trust the number for the +5VSB aswell.

On a related note what is an affordable/cheap and quiet mdoern PSU that is safe to use with P1/PII/PIII systems? Old PSUs in my retro PC cases are driving me nuts with their 80mm loud fans, so I need an alternative I can buy a couple of, for a fairly low price..

Skyscraper wrote:

... Let me present the COLORSit Model 340U-SCE... 420W PSU 🤣

Not too bad but you really want to recap at least the secondary caps and the smaller caps between heatsinks to reasonably trust that thing.

EDIT - Also replacing the-infamous-four-diode-treatment with a rectifier bridge wouldn't harm.

appiah4 wrote:

... what is an affordable/cheap and quiet mdoern PSU that is safe to use with P1/PII/PIII systems?

https://www.newegg.com/Product/Product.aspx?I … N82E16817151086

TELVM wrote:

https://www.newegg.com/Product/Product.aspx?I … N82E16817151086

Would $10 300W PSUs with 120MM bottom fans be terrible ideas to use with 100W P1/P2 systems? My main issue with old PSUs is noise and these will be guaranteed quieter, and I doubt they would be any less reliable than 20 year old 300W PSUs..

I'm looking into buying something a couple of these for $8 each. They won't be used for anything more power hungry than a Celeron 1300 + Radeon 8500.

Well, you can buy one and split it open and we'll try to contain our laughter (or our tears).

Personally, I wouldn't trust such a cheap power supply further than I can throw it (which would be quite far, considering how light it probably is).

gdjacobs wrote:

Well, you can buy one and split it open and we'll try to contain our laughter (or our tears).

Personally, I wouldn't trust such a cheap power supply further than I can throw it (which would be quite far, considering how light it probably is).

I'll probably buy just one for now. for science. and for the good of us all. except the ones who are dead.

(I may actually buy a couple more later because today I opened up an old FSP 300W I had and it seems there is a strange sticky looking, honey-like residue covering half of it where there should just be opaque yellow gluing material. It's not sticky, it seems to be pretty frozen actually. I'm thinking something leaked from one or more of the capacitors.. It works but I don't trust it at all, it will get binned..)

The caps don't contain that much electrolyte. It's likely substandard adhesive caulking which can be cleaned with IPA.

gdjacobs wrote:

The caps don't contain that much electrolyte. It's likely substandard adhesive caulking which can be cleaned with IPA.

I suppose that using it isn't unsafe in this case? I'm not very comfortable with sticking an $8 PSU in a Celeron 1300/Radeon 8500 system..

appiah4 wrote:

Would $10 300W PSUs with 120MM bottom fans be terrible ideas to use with 100W P1/P2 systems? ...

Definitely yes, beware of any $10 PSU, guaranteed gutless-wonder fire-hazard. They may appear to work OK at the beggining but in the mid/long term they're a risk for your system and home.

Don't Be Surprised When Your Cheap PSU Blows Up

Low-cost ATX Power Supply Fail - Crash Test

Here's what I'll do. Tonight I'll open up the PSU I'm thinking of replacing and post pictures of what it has (or lacks) inside. For shits and giggles. If it can't possibly be worse than a more silent conptemporary shitty PSU, I'll order the $8 300W model and open it up for Vogons to have a good laugh. 😀

appiah4 wrote:

Here's what I'll do. Tonight I'll open up the PSU I'm thinking of replacing and post pictures of what it has (or lacks) inside. For shits and giggles. If it can't possibly be worse than a more silent conptemporary shitty PSU, I'll order the $8 300W model and open it up for Vogons to have a good laugh. 😀

Ok, so I did this. Here's the unit in question.

It came from an OEM computer with a Celeron 700 by default, which was later upgraded to a Celeron 1300 which now runs on an FSP 300W PSU. I moved this PSU to a Pentium 166MMX system with a Voodoo 2.

And now, here are its innards, for your evaluation of how likely it is to set my house on fire 😀

And here are the final two photos:

Hmm... poor input filtering, output filtering, anemic capacitors. I'm not too good at identifying transistors by sight, but they're probably way overstretched as well. Good luck on this thing actually having any active thermal or overcurrent protection.

gdjacobs wrote:

Hmm... poor input filtering, output filtering, anemic capacitors. I'm not too good at identifying transistors by sight, but they're probably way overstretched as well. Good luck on this thing actually having any active thermal or overcurrent protection.

I expected as much, the whole thing weighs about 500g 😀 I suppose a $8 PSU can't be much worse, at least it would be silent 😁 I'll order one and do the same for it and compare how far cheap PSUs came in 15 years.

appiah4 wrote:

... And now, here are its innards, for your evaluation of how likely it is to set my house on fire 😀

That PoS sux so much that I'll leave aside the comp killing aspects and concentrate on the home razing features:

. .

appiah4 wrote:

... I suppose a $8 PSU can't be much worse, at least it would be silent 😁 I'll order one and do the same for it and compare how far cheap PSUs came in 15 years.

Yes please!

TELVM wrote:

That PoS sux so much that I'll leave aside the comp killing aspects and concentrate on the home razing features:

For maximum fire risk, I wanna see some inadequate bonding methods! Coupled with those lovely filter caps, you have the opportunity of being electrocuted by a computer chassis floating at mains voltage.