The black ones on the right are "Fuhiyyu" crap capacitors. They must be replaced regardless of the optical condition.
The two orange ones are bulging so replacement is inevitable.
I cannot find the blue one though.

majestyk wrote on 2021-11-03, 17:14:

The black ones on the right are "Fuhiyyu" crap capacitors. They must be replaced regardless of the optical condition.
The two orange ones are bulging so replacement is inevitable.
I cannot find the blue one though.

All five caps are branded Fuhjyyu. The fine looking blue cap is a Fuhjyyu TM (Low ESR) and the bulging ones are Fuhjyyu TN (General Purpose)

Should I maybe just leave the TM (Low ESR) cap alone seeing as it seems to be fine, and just replace the GP ones?

Also, do I go with 6.3/10/16V as per the originals or just go with 25V for all, seeing as the lower V rated ones have bloated more?

The blueish Fuhiyyus are still Fuhiyyus so replacement is a must.
25V types won´t fit.
Yes they are standard low ESR, but not ultra low ESR. You can use Rubycon ZLH or some Panasonic low ESR.
Try to find a series with a long lifetime @ 105° centigrade. About 3000-5000h would be perfect. You will see that it´s already a challenge to find a fitting form factor with the current original voltages.

Rubycon/Panasonic not an option for me. What I can get are 2200uF 16V Samwha WL Ultra Low ESR caps rated for 105C and 5000hrs. They'll have to do as replacements for all five.

There have been reports that in some cases using filter capacitors with too low ESRs can cause instabilities or a tendency to oscillations. I have never experienced such a case myself and I have recapped a lot of PSUs.
After replacing the caps here you maybe should observe the working PSU under load for a while if there are any semiconductors overheating and if the voltages have correct and stable values.

majestyk wrote on 2021-11-03, 17:39:

The blueish Fuhiyyus are still Fuhiyyus so replacement is a must.

Especially considering the fact that in FSP PSUs the caps get well cooked.
If it doesn't bulge, it gasses dry.

Also look at the caps between the two big coolers.
It gets very hot there, too.
I don't trust FSP's customary "solution" to often "protect" the caps there with shrink tubing "heat insulation" much.

How much load will your build draw on 12V?
I ask because if the load is less than at least 2 to 4 amperes, personally I would not consider using a FSP PSU for a build.
Reason is their very bad crossloading behaviour; with too low load on 12V, the 12V become almost 15V.
And I seriously doubt that this is good for the drives.

retardware wrote on 2021-11-03, 18:01:

Especially considering the fact that in FSP PSUs the caps get well cooked. If it doesn't bulge, it gasses dry. […]
Show full quote

majestyk wrote on 2021-11-03, 17:39:

The blueish Fuhiyyus are still Fuhiyyus so replacement is a must.

Especially considering the fact that in FSP PSUs the caps get well cooked.
If it doesn't bulge, it gasses dry.

Also look at the caps between the two big coolers.
It gets very hot there, too.
I don't trust FSP's customary "solution" to often "protect" the caps there with shrink tubing "heat insulation" much.

How much load will your build draw on 12V?
I ask because if the load is less than at least 2 to 4 amperes, personally I would not consider using a FSP PSU for a build.
Reason is their very bad crossloading behaviour; with too low load on 12V, the 12V become almost 15V.
And I seriously doubt that this is good for the drives.

Thanks for the headsup, I will check those as well.

This build would have nothing on the 12V rail aside from an HDD, FDD and optical drive - nothing like a Radeon 9700PRO etc. It will be mostly 5V heavy due to AthlonXP CPU.

retardware wrote on 2021-11-03, 18:01:

Especially considering the fact that in FSP PSUs the caps get well cooked. If it doesn't bulge, it gasses dry. […]
Show full quote

majestyk wrote on 2021-11-03, 17:39:

The blueish Fuhiyyus are still Fuhiyyus so replacement is a must.

Especially considering the fact that in FSP PSUs the caps get well cooked.
If it doesn't bulge, it gasses dry.

Also look at the caps between the two big coolers.
It gets very hot there, too.
I don't trust FSP's customary "solution" to often "protect" the caps there with shrink tubing "heat insulation" much.

How much load will your build draw on 12V?
I ask because if the load is less than at least 2 to 4 amperes, personally I would not consider using a FSP PSU for a build.
Reason is their very bad crossloading behaviour; with too low load on 12V, the 12V become almost 15V.
And I seriously doubt that this is good for the drives.

Are you sure about the bad crossloading causing the 12v to go up to 15v with (all?) FSP PSUs? Or is a certain series affected by this?
That doesn't seem like it could find its way into PSUs that have retailed into the many thousands.
Do you have a source of this info?

I'd also like to learn of this 'fact' that it's (apparently) a FSP thing that caps get cooked. This seems rather unlikely to me since there have been perhaps hundreds of different FSP PSU models with many different kinds of cooling solutions. One FSP PSU I had even had the PSU fan controlled with a knob that I could turn to run faster so overheating was quite easy to prevent in this particular unit.

Btw, if you want a tightly packed PSU with poor cooling, I got a Tagan 380W PSU that definitely ran hotter than any FSP PSU I ever encountered (and I encountered a lot of them).

Honestly I've used a lot of FSP PSUs and never saw this kind of weird crossloading problem..

appiah4 wrote on 2021-11-03, 18:32:

Honestly I've used a lot of FSP PSUs and never saw this kind of weird crossloading problem..

Same thing here.
That's why I asked for a source of this information (and especially which models could be affected as there are so many different FSP PSU models out there, it's impossible to accurately generalize on this point).

Tetrium wrote on 2021-11-03, 18:31:

Are you sure about the bad crossloading causing the 12v to go up to 15v with (all?) FSP PSUs? Or is a certain series affected by this?
That doesn't seem like it could find its way into PSUs that have retailed into the many thousands.
Do you have a source of this info?

This is just my personal findings.
I have had a lot of PSUs on my test bench to test for ATX specification conformity (in particular, ripple and crossloading behaviour).
If you are interested in the results, I wrote a bit about that here (german, use translator if needed).

I usually put every PSU onto the test bench before considering using it.
And regarding the FSP ones, old or (quite) new, I haven't yet found any one which does not fail the ATX specification crossloading limits.
Most go near to 14V, some even close to 15V.
For this reason I usually throw them away without even bothering to test them.

Too high "12V" or ripple normally does not lead to immediate failures, so users rarely associate HDDs gone defective with the PSU as a possible cause.
But, especially with rare retro hardware, personally I like to be conservative and make sure the voltages are in the spec limits, as replacement is not always easy/cheap.

retardware wrote on 2021-11-03, 19:21:

This is just my personal findings. I have had a lot of PSUs on my test bench to test for ATX specification conformity (in partic […]
Show full quote

Tetrium wrote on 2021-11-03, 18:31:

Are you sure about the bad crossloading causing the 12v to go up to 15v with (all?) FSP PSUs? Or is a certain series affected by this?
That doesn't seem like it could find its way into PSUs that have retailed into the many thousands.
Do you have a source of this info?

This is just my personal findings.
I have had a lot of PSUs on my test bench to test for ATX specification conformity (in particular, ripple and crossloading behaviour).
If you are interested in the results, I wrote a bit about that here (german, use translator if needed).

I usually put every PSU onto the test bench before considering using it.
And regarding the FSP ones, old or (quite) new, I haven't yet found any one which does not fail the ATX specification crossloading limits.
Most go near to 14V, some even close to 15V.
For this reason I usually throw them away without even bothering to test them.

Too high "12V" or ripple normally does not lead to immediate failures, so users rarely associate HDDs gone defective with the PSU as a possible cause.
But, especially with rare retro hardware, personally I like to be conservative and make sure the voltages are in the spec limits, as replacement is not always easy/cheap.

That's actually very interesting!
There's a lot that I like about your methodology, including that you seem to like to think for yourself and find your own way to test your surroundings and draw your own conclusions from what you're learned. I find this nothing but commendable!
But it also comes with its flaws and that's that it's also easier for errors and oversights to slip through.

I'm still reading it all (am on page 2 atm. Good thing I can read german natively 🤣 but it's not as fast as I can read english or dutch).

One of the things I notice on the "Fortron/Source FSP350-EMDN" is that (I can't see much from the little pic of the fan) the fan seems to indicate that the owner of the PC that contained that PSU, was a smoker. It would also explain the apparent heatdeath.
Also, the 12cm placement is not necessarily a design flaw. Quite often roughly during the switch from mostly 5v to mostly 12v, another switch occured: The switch from basically 8cm fans to the use of 12cm fans. I have noticed that FSP likely made designs for some of their contemporary designs to be able to be fitted with either an 8cm, or a 12cm fan (just use a different cap used to close off the PSU to which often the 12cm fan is mounted). Often you'd be able to see this as there can actually be room for an 8cm fan near where the powerswitch is (usually not mounted in the cap that closes the PSU body, but the place would be in the main body of the PSU).

Also, PSUs should be cleaned somewhat regularly if becoming dusty. Don't even need to open it if you use some canned air. Heck, I've even used a bicycle tire pump for this when I had ehm...misplaced my can of canned air 😜
It gets most of the dust out, just be sure you don't get any fluids in there when dedusting.

Also don't forget that these old PSUs often will have worn parts which will influence how the units will behave during testing. The minimum load requirements may have changed during the early ATX years, you should keep that in mind as well.

I'm still reading 😜

On page 2, here "Fortron, fortgesetzt :shock:"
Just pointing out that the black circly thingy around the PSU cables coming out of the PSU housing seems to be missing. Usually there would be a black thick plastic circle which could be placed in the hole where the cables exit the housing. It would kinda click into place (it's basically 2 hard plastic half-circles hold together with a bit of softer plastic).

Over there the same things are also mentioned: it being old parts and the critisism about it supposedly being a design flaw that made the units run hot (which in my experience is not the case).

I don't know about the safety with the pink 'sticker' and the yellow bit of plastic. It could be interesting to see if other units of this model are made the same way but I will state that poking into a live PSU is always a bad idea anyway 😜

About the pink stuff on the outside, I have seen odd things stuck to the outsides of PSUs actually fairly regularly, including little square bits. I always assumed these were used as some kind of support when mounted in certain cases but I never actually looked for an explanation as it never really seemed very important to me.

1ST1 from your forums explains one thing quite nicely, about the differences in 12v and 5v heavy systems needing different testing and PSUs needing some base load to even be tested properly.
Also one of the FSP PSUs you tested was a Medion OEM PSU which may actually be different in some way compared to standard ATX specs (like how some Dell PSUs were actually wired differently, which is also not standard ATX spec afaik).

One thing you say here I simply don't agree with:
"Anders gesagt, die einseitige Belastung auf 5V wird für den Retrobereich zunehmend typischer, noch mehr als sie seinerzeit bereits war."
translated it roughly means that for retro purposes the sole load on 5v (while omitting a load on 12v) will become more important, even compared with back then.
It is simply not true, PSUs from that era were literally designed with 5v and 12v load in mind. 12v was used for certain drives and for fans for example, so even when CPUs drew from the 5v rails, there was always a load on the 12v rails as well. Btw the FSP PSU you tested is one of the more recent "mostly 12v" ones.

This is why I grab 750 watt atx PSU from the shop and add ATX to P2 P3 patch cable where the old motherboards don't do ATX. (550 watt min. recommend)
A spray by grey paint on it is close enough to retro metal 😀
Thats because I am not a qualified technician whom costs a lot to recap and repair power supplies, and I don't have insurance of one if it causes a fire after repair. (i'd want to sue than blame myself)
Everything else outside the power supply box is green to go with repair, refurbish and replace because of low voltage doesn't require a permit or qalification..
*All 200 watt and below power supplies stress much more and die sooner*

Tetrium wrote on 2021-11-03, 20:48:

On page 2, here "Fortron, fortgesetzt :shock:"
Just pointing out that the black circly thingy around the PSU cables coming out of the PSU housing seems to be missing. Usually there would be a black thick plastic circle which could be placed in the hole where the cables exit the housing. It would kinda click into place (it's basically 2 hard plastic half-circles hold together with a bit of softer plastic).

On later Fortron PSUs this plastic protector, which prevents the metal case to cut into the cables and short them out, has been left out.
The sharp cut metal has been bent ("gebördelt" in German, don't know the English word for that) so that the metal no longer cuts into the cables.
This reduces the short risk almost to zero, as usually there is not enough cable movement to make the metal chafe through the insulation.

Tetrium wrote on 2021-11-03, 20:48:

I don't know about the safety with the pink 'sticker' and the yellow bit of plastic. It could be interesting to see if other units of this model are made the same way but I will state that poking into a live PSU is always a bad idea anyway 😜

What I was thinking about was this soft insulation strip can be easily damaged (think of PSUs being thrown/stored in a box like many people do with boards.
This would result in the possibility to get shocked only by touching the outside, like when fiddling in the computer that is connected to the grid.

Tetrium wrote on 2021-11-03, 20:48:

1ST1 from your forums explains one thing quite nicely, about the differences in 12v and 5v heavy systems needing different testi […]
Show full quote

1ST1 from your forums explains one thing quite nicely, about the differences in 12v and 5v heavy systems needing different testing and PSUs needing some base load to even be tested properly.
[...]
One thing you say here I simply don't agree with:
"Anders gesagt, die einseitige Belastung auf 5V wird für den Retrobereich zunehmend typischer, noch mehr als sie seinerzeit bereits war."
translated it roughly means that for retro purposes the sole load on 5v (while omitting a load on 12v) will become more important, even compared with back then.
It is simply not true, PSUs from that era were literally designed with 5v and 12v load in mind. 12v was used for certain drives and for fans for example, so even when CPUs drew from the 5v rails, there was always a load on the 12v rails as well. Btw the FSP PSU you tested is one of the more recent "mostly 12v" ones.

The problem is that badly regulated PSUs, no matter whether 5V or 12V-"centric" fail the ATX specs voltages when not being loaded in a particular "optimal" load ratio.
For this reason there were used load resistors back then.

And what I tried to say was that the HDD "dummy load" decreased with time.
5.25" FH 4+A, 3.5" drive usually ~0.5A, max. 1A, and (almost) no load with flash/SSD, Gotek, etc.

My PSU tester has two big banks of load resistors, and the load can be adjusted from 0A to 18A in roughly 1A steps for both 5 and 12V.
So a wide range of load scenarios can be "emulated" from the 1-chip XT and mini ITX Atom upwards.
The voltages should be in spec limits no matter which the load ratios are, at least according to the ATX spec.

The problem that many people apparently do not see is that the majority of retro PCs is very 5V heavy, with almost no load on 12V.
The todays' "loads" like 3.5"HDDs are way too low to prevent bad PSUs from going near 14, 15V.
This makes the crossloading problem worse than it was when HDDs consumed much larger currents that 1ST1 mentioned as "minimum loads for correctly testing PSUs".
Even back then diskless workstations had no such minimum loads.
For this reason IMHO this lack of defined minimum loads that can be relied upon is reflected correctly by the ATX specs which matches reality by saying the voltages have to be in spec tolerances no matter of the given load ratio, with zero minimum load.

For this reason I think assuming a minimum load like in 5.25 and early 3.5" HDD ages might be correct for the thinking 30, 40 years ago, but just does not match today's reality with low-power HDDs or even flash memory.

And this is why I recommend modern PSUs based on DC-DC converters that create 5 and 3.3V not from the primary, but instead from the 12V rail. These, and only these, have no crossloading issues in form of voltages varying depending of varying loads. With these, and only with these, 12V is always 12V, and 5V is always 5V.

retardware wrote on 2021-11-03, 21:52:

On later Fortron PSUs this plastic protector, which prevents the metal case to cut into the cables and short them out, has been […]
Show full quote

Tetrium wrote on 2021-11-03, 20:48:

On page 2, here "Fortron, fortgesetzt :shock:"
Just pointing out that the black circly thingy around the PSU cables coming out of the PSU housing seems to be missing. Usually there would be a black thick plastic circle which could be placed in the hole where the cables exit the housing. It would kinda click into place (it's basically 2 hard plastic half-circles hold together with a bit of softer plastic).

On later Fortron PSUs this plastic protector, which prevents the metal case to cut into the cables and short them out, has been left out.
The sharp cut metal has been bent ("gebördelt" in German, don't know the English word for that) so that the metal no longer cuts into the cables.
This reduces the short risk almost to zero, as usually there is not enough cable movement to make the metal chafe through the insulation.

Tetrium wrote on 2021-11-03, 20:48:

I don't know about the safety with the pink 'sticker' and the yellow bit of plastic. It could be interesting to see if other units of this model are made the same way but I will state that poking into a live PSU is always a bad idea anyway 😜

What I was thinking about was this soft insulation strip can be easily damaged (think of PSUs being thrown/stored in a box like many people do with boards.
This would result in the possibility to get shocked only by touching the outside, like when fiddling in the computer that is connected to the grid.

Tetrium wrote on 2021-11-03, 20:48:

1ST1 from your forums explains one thing quite nicely, about the differences in 12v and 5v heavy systems needing different testi […]
Show full quote

1ST1 from your forums explains one thing quite nicely, about the differences in 12v and 5v heavy systems needing different testing and PSUs needing some base load to even be tested properly.
[...]
One thing you say here I simply don't agree with:
"Anders gesagt, die einseitige Belastung auf 5V wird für den Retrobereich zunehmend typischer, noch mehr als sie seinerzeit bereits war."
translated it roughly means that for retro purposes the sole load on 5v (while omitting a load on 12v) will become more important, even compared with back then.
It is simply not true, PSUs from that era were literally designed with 5v and 12v load in mind. 12v was used for certain drives and for fans for example, so even when CPUs drew from the 5v rails, there was always a load on the 12v rails as well. Btw the FSP PSU you tested is one of the more recent "mostly 12v" ones.

The problem is that badly regulated PSUs, no matter whether 5V or 12V-"centric" fail the ATX specs voltages when not being loaded in a particular "optimal" load ratio.
For this reason there were used load resistors back then.

And what I tried to say was that the HDD "dummy load" decreased with time.
5.25" FH 4+A, 3.5" drive usually ~0.5A, max. 1A, and (almost) no load with flash/SSD, Gotek, etc.

My PSU tester has two big banks of load resistors, and the load can be adjusted from 0A to 18A in roughly 1A steps for both 5 and 12V.
So a wide range of load scenarios can be "emulated" from the 1-chip XT and mini ITX Atom upwards.
The voltages should be in spec limits no matter which the load ratios are, at least according to the ATX spec.

The problem that many people apparently do not see is that the majority of retro PCs is very 5V heavy, with almost no load on 12V.
The todays' "loads" like 3.5"HDDs are way too low to prevent bad PSUs from going near 14, 15V.
This makes the crossloading problem worse than it was when HDDs consumed much larger currents that 1ST1 mentioned as "minimum loads for correctly testing PSUs".
Even back then diskless workstations had no such minimum loads.
For this reason IMHO this lack of defined minimum loads that can be relied upon is reflected correctly by the ATX specs which matches reality by saying the voltages have to be in spec tolerances no matter of the given load ratio, with zero minimum load.

For this reason I think assuming a minimum load like in 5.25 and early 3.5" HDD ages might be correct for the thinking 30, 40 years ago, but just does not match today's reality with low-power HDDs or even flash memory.

And this is why I recommend modern PSUs based on DC-DC converters that create 5 and 3.3V not from the primary, but instead from the 12V rail. These, and only these, have no crossloading issues in form of voltages varying depending of varying loads. With these, and only with these, 12V is always 12V, and 5V is always 5V.

Even the FSP Hexa PSUs have at least a plastic ring protecting the wires at the point coming out of the PSU.
The old 2xhalf rings were quite easy to go missing. It's possible it did in your case as well

Even though I don't have a handling manual or similar for that particular PSU at hand I can confidently say that I am fairly sure that PSUs are not meant to be thrown around and handled roughly. Doing so would basically invalidate crisisism to the desing if it gets damaged that way because any PSU can get damaged if, for instance, a tank were to drive over it.
Same error as your reasoning regarding you disregarding the need for a 12v laod due to certain 12v appliances no longer being feasable according to you that is, which is essentially just your personal opinion), this does not invalidate the designs in any way because they weren't designed for that purpose, therefor you cannot say the design is flawed if you test it using a methodology that inherently ignores the reasons these units were designed in the first place. That's like saying that FSP formula 1 cars suck because you found that they are doing a terrible job at riding off-road, which is simply unfair.

PSUs were supposed to be loaded within a specified way anyway.
And I seriously doubt that, as you are implying, FSP has been massively ignoring ATX design specs for many years now and only you are the very first person on this earth that saw through their devilish ways...
Look here,
You tested only a handful of units and your testing methods have several things to be left desired, so it's simply too early to bin all FSP PSUs altogether, based on the limited data you have provided.

Personally I think testing is a great way to learn more about certain products, but your testing does not support your claims about this particular brand (and you should have been able to see this, if only because there's virtually no circumstantial evidence supporting your claim, may people have actually got relatively good experiences and you simply shove this away with your sometimes just flawed reasonings).

You're pulling the trigger too fast and it's a shame because in a way I really like your way of testing and interpreting. And it's ok to make mistakes but only if you learn from them.

I'm curious, was that PSU you mentioned as having had a heat death from a smoker or not? The fan from that PSU really seemed to have the telltale signs of one (it should be brownish and can be moderately sticky). Perhaps the fan has malfunctioned, it can seize up when the turning mechanism gets stuck.

Yes, the voltages should be in spec, depending on what ATX specs the specific PSU was designed for and if the unit is brand new.

Tetrium wrote on 2021-11-03, 23:37:

Even the FSP Hexa PSUs have at least a plastic ring protecting the wires at the point coming out of the PSU.
The old 2xhalf rings were quite easy to go missing. It's possible it did in your case as well

I'll make+post photos of that later to illustrate

Tetrium wrote on 2021-11-03, 23:37:

PSUs were supposed to be loaded within a specified way anyway.
And I seriously doubt that, as you are implying, FSP has been massively ignoring ATX design specs for many years now and only you are the very first person on this earth that saw through their devilish ways...

Most ATX PSUs do not keep ATX specifications, which are only a recommendation, not a law.
Fortron is no exception there.
The big difference is just their good marketing, building some reputation in the end user market, which causes (imho unjustified) trust into their products.
I just do not let trick me into this because I have actually tested these, which probably less than 1 in 10000 users do. Thus for me, Fortron has become category "noname", after I saw how bad they are.

Thus, there is no problem for the "bad manufacturers" because there is no knowledge about the technical implications that arise when using PSUs that do not keep voltages in spec. Users normally do not put a seemingly "working" PSU into closer observation.

For example, @pixelatedscraps recently invested much time to diagnose a non-functioning board, until he put into consideration to use a Seasonic PSU instead of a Fortron PSU, and bingo, all things turned out to work fine.

As I said, most times it works with voltages off-spec. But for example, some FSC PSUs which limit 12V overvolting, in return they reduce the 5V supply.
I think it is not too hard to understand that some 5V heavy boards will still (barely) work with, say, 4.1V, while others don't.

So users of bad PSUs do not even realize the damage it does to them by making them believe boards or other hardware is broken, and in turn they throw away perfectly good hardware, just because a bad PSU tricked them into believing this.
Isn't this sad?

Tetrium wrote on 2021-11-03, 23:37:

You're pulling the trigger too fast and it's a shame because in a way I really like your way of testing and interpreting. And it's ok to make mistakes but only if you learn from them.

I just follow my personal criteria: the PSUs I use and recommend to use must be in spec, not damage things or cause bad side effects like the ones I described above.

BTW, this is also the reason why OEMs like HP, Dell etc usually use PSUs that keep specs, it is just a financial matter.
RMAs because of bad PSUs cause high cost.
Does anyone wonder why one never finds Fortron PSUs in computers of these companies?

Tetrium wrote on 2021-11-03, 23:37:

I'm curious, was that PSU you mentioned as having had a heat death from a smoker or not? The fan from that PSU really seemed to have the telltale signs of one (it should be brownish and can be moderately sticky). Perhaps the fan has malfunctioned, it can seize up when the turning mechanism gets stuck.

No, it didn't look like smoker (no grime). It looked more like a PC that stood on the floor (the bottom of the room is the dustiest) in a corner of a carpeted room, collecting/filtering dust.

Tetrium wrote on 2021-11-03, 23:37:

Yes, the voltages should be in spec, depending on what ATX specs the specific PSU was designed for and if the unit is brand new.

Output voltages are not dependent of age, they are dependent of circuit design.
Again, many manufacturers, particularly low-end ones, do not seem to care too much about the tight ATX spec, they care only about that the tolerances are in a range that limits RMA costs to a particular threshold.

It is upon every user to decide what (s)he wants and is willing to accept.
Some care, some care less, some don't.

Edit: And many just don't know. This is why some people including me try to make people aware, but there are too few.

retardware wrote on 2021-11-04, 18:41:

I'll make+post photos of that later to illustrate […]
Show full quote

Tetrium wrote on 2021-11-03, 23:37:

Even the FSP Hexa PSUs have at least a plastic ring protecting the wires at the point coming out of the PSU.
The old 2xhalf rings were quite easy to go missing. It's possible it did in your case as well

I'll make+post photos of that later to illustrate

Tetrium wrote on 2021-11-03, 23:37:

PSUs were supposed to be loaded within a specified way anyway.
And I seriously doubt that, as you are implying, FSP has been massively ignoring ATX design specs for many years now and only you are the very first person on this earth that saw through their devilish ways...

Most ATX PSUs do not keep ATX specifications, which are only a recommendation, not a law.
Fortron is no exception there.
The big difference is just their good marketing, building some reputation in the end user market, which causes (imho unjustified) trust into their products.
I just do not let trick me into this because I have actually tested these, which probably less than 1 in 10000 users do. Thus for me, Fortron has become category "noname", after I saw how bad they are.

Thus, there is no problem for the "bad manufacturers" because there is no knowledge about the technical implications that arise when using PSUs that do not keep voltages in spec. Users normally do not put a seemingly "working" PSU into closer observation.

For example, @pixelatedscraps recently invested much time to diagnose a non-functioning board, until he put into consideration to use a Seasonic PSU instead of a Fortron PSU, and bingo, all things turned out to work fine.

As I said, most times it works with voltages off-spec. But for example, some FSC PSUs which limit 12V overvolting, in return they reduce the 5V supply.
I think it is not too hard to understand that some 5V heavy boards will still (barely) work with, say, 4.1V, while others don't.

So users of bad PSUs do not even realize the damage it does to them by making them believe boards or other hardware is broken, and in turn they throw away perfectly good hardware, just because a bad PSU tricked them into believing this.
Isn't this sad?

Tetrium wrote on 2021-11-03, 23:37:

You're pulling the trigger too fast and it's a shame because in a way I really like your way of testing and interpreting. And it's ok to make mistakes but only if you learn from them.

I just follow my personal criteria: the PSUs I use and recommend to use must be in spec, not damage things or cause bad side effects like the ones I described above.

BTW, this is also the reason why OEMs like HP, Dell etc usually use PSUs that keep specs, it is just a financial matter.
RMAs because of bad PSUs cause high cost.
Does anyone wonder why one never finds Fortron PSUs in computers of these companies?

Tetrium wrote on 2021-11-03, 23:37:

I'm curious, was that PSU you mentioned as having had a heat death from a smoker or not? The fan from that PSU really seemed to have the telltale signs of one (it should be brownish and can be moderately sticky). Perhaps the fan has malfunctioned, it can seize up when the turning mechanism gets stuck.

No, it didn't look like smoker (no grime). It looked more like a PC that stood on the floor (the bottom of the room is the dustiest) in a corner of a carpeted room, collecting/filtering dust.

Tetrium wrote on 2021-11-03, 23:37:

Yes, the voltages should be in spec, depending on what ATX specs the specific PSU was designed for and if the unit is brand new.

Output voltages are not dependent of age, they are dependent of circuit design.
Again, many manufacturers, particularly low-end ones, do not seem to care too much about the tight ATX spec, they care only about that the tolerances are in a range that limits RMA costs to a particular threshold.

It is upon every user to decide what (s)he wants and is willing to accept.
Some care, some care less, some don't.

Edit: And many just don't know. This is why some people including me try to make people aware, but there are too few.

I'm not going to respond to everything, as I don't have the time rn.
But FSP PSUs were used in OEMs. Fujitsu Siemens is one well known brand that used FSP PSUs for their systems. Medion is another one and MyCom (used to be a well known PC builder here in The Netherlands) also used FSP PSUs extensively so I'm kinda surprised you weren't aware of that.

The big difference is just their good marketing

What marketing?
Fortron (particularly back then) did actually not do any significant advertising. They were a strictly OEM PSU manufacturer back then, often building designs for other brands like AOpen and Zalman or supplier of several large OEMs. They did the advertising.
It's only in much later years that Fortron actually started making PSUs that were designed to have more appeal with end users but the old boxes were just grey plain boring looking boxes.

This part

So users of bad PSUs do not even realize the damage it does to them by making them believe boards or other hardware is broken, and in turn they throw away perfectly good hardware, just because a bad PSU tricked them into believing this.
Isn't this sad?

is not telling much of anything. It's just your personal stance on things.

Fine, you don't like this brand. But you are not liking it for the wrong reasons and based on too small a control group combined with too little actual data. It's like you got bitten by a dog, so all dogs must be evil or something. And your presenting your personal findings as an ultimate truth or something which is simply inaccurate. And misleading. It's fine to have an opinion but you should stick to the facts. FSP has probably build millions of PSU by now (the oldest I found were made even before 2000) and hundreds of reviews have been published and if FSP was really as bad as you claim they are, they would most likely have been caught by now. And your claims and findings are a lot less substantial than some of the ones out there. It's kinda too bad that the older ones (from the 5v era, as I like to call it) didn't get the full spectrum of reviews as such a thing sumply didn't exist on the net back then.

Just to clarify, I'm talking about mostly the older designs (the ones that still had good 5v lines). The period after that I felt FSP has started to slowly cheap out a bit on their designs (using thinner wires etc and cheaper caps). But they have never belonged into the "gutless wonders" category and usually their PSUs are at least somewhat decent.
But the old designs from the sA era? Those were pretty solid, particularly when compared to other contemporary designs. Their major flaw back then (especially the 250W and 300W units) was the use of plague caps but this was a thing for many major PSU suppliers, including Antec for instance.