Mondodimotori wrote on 2024-09-08, 15:03:
Under spoiler the video and five of the pictures I've took. If more are needed, I'll provide, the PSU is still outside the sistem and readily available. I've made sure to disconnect it and not touch anything inside. Just looking.
OK, I just took a look at the pictures (and also viewed the video, but I'm on my laptop and couldn't hear the whine all that well over the laptop's fans).
The whine doesn't appear to be anything too abnormal, though. I've had a few ATX PSUs over the years that produce similar whining noise, despite being recapped. All of these are still in service and working fine. So whining from the 5VSB isn't necessarily a bad thing. Nevertheless, I can't say if you should ignore it yet or not. In particular, your Energmax PSU appears to be very similar to an Enermax Liberty ELT400AWT I got some years back for free with its wires cut and a label on the PSU that said "broken". It tuned out the PSU just had a few bad electrolytic capacitors inside, including the 5VSB rail... hence likely why someone labeled it as "broken". All I had to do was replace the bad capacitors and voila, the PSU was working fine once again. What's funny is I didn't even have the right capacitor sizes at the time, so I stuck whatever (reused/old) caps I had on hand that seemed to be "close enough...ish". Nevertheless, this was good enough to get the PSU working. In fact, now I remember I made a post about it on badcaps forums (link here, as I did with many of my other PSUs.)
That being said, it's not a given that the caps in your Enermax PSU haven't drifted (or are starting to drift) out of spec. Generaly, CEC-holdings PCE-TUR and PCE-TUL caps are pretty good about bulging when they go bad, so there's a chance they are indeed OK if they aren't bulged. In any case, you should have a closer look at them when you get a chance. To do that, you'll need more than a multimeter - you need one of those "component/transistor testers" found online (ebay, Ali, Amazon, etc.) They have a function to test capacitance and ESR, which are two of the most important parameters that can tell you if an electrolytic capacitor has failed or not.
Thus, the three simplest options I can suggest to you are as follows:
1) Get a multimeter and some 12V car/auto incandescent or halogen bulbs (preferably rated around 10-20 Watts). Using some wires and/or small jumper cables, connect the 12V bulbs to the 5VSB rail, adding one at a time. Use the multimeter to measure the voltage on the 5VSB rail to see how badly it drops. If the caps on the 5VSB circuit are bad, typically the 5VSB circuit will start to whine even more / louder and the voltage usually will drop a lot (<4.75V.) But if nothing changes and the 5VSB seems stable with good voltage close to 5V, then the whining may be normal for your PSU.
2) Get new caps for the 5VSB circuit specifically and just recap that. As mentioned by CharlieFoxtrot, the 5VSB circuit is the one that is always On when the PSU is plugged in and also the source of the whining noise you hear. If new caps on the 5VSB don't stop the issue, then it's normal for the 5VSB circuit to be doing this. Unfortunately, this is one of the very few cases where I didn't do a complete part breakdown of what caps are installed into this PSU, so I can't tell you exactly the values of the caps for the 5VSB. But going from my post on badcaps, there seem to be only two caps on the 5VSB output (which is actually pretty common in most designs). They are not visible in the pictures you provided, but likely should be "in the back", close to the ventilation holes in the upper-right corner of the PSU when using this picture of yours as a reference.
For the first cap, I used a 1500 uF cap and for the 2nd a 1000 uF cap. If memory serves me right, the 2nd cap actually had a smaller capacitance value, but from experience working on PSUs before, it's usually no problem (if anything, it's GOOD) to actually increase the capacitance a little, if space allows. So for the 5VSB, you can grab something like a United Chemicon KY 10 2200 uF cap (pretty common) or Rubycon ZLH 6.3V 2200 uF or .... something along those lines (just preferably a Japanese capacitor brand and "medium" low impedance / ESR) - that's for the 1st cap. And really the same goes for the 2nd cap: just something in the range of 680-1000 uF and 6.3V or 10V (or even 16V)... again, either Chemicon KY or KYB or KZE... or Nichicon HE or HW... or Rubycon ZLH or ZLQ or ZL or ZLJ... or Panasonic FR or FM or FS. All of these are known reliable brands/series of caps suitable for 5VSB (or any other circuit that uses "flyback" design) use.
3) Just use the PSU and hope for the best. beaming face with smiling eyes Either that or not use it.
My personal hunch is that the PSU 5VSB is probably fine with the current caps... at least for now. But without testing anything, we can't know for 100%.
With that said, if getting the equipment and/or caps needed to go with option 1 or 2 above are not too incovinient and you really care about your old hardware, then you should get the PSU checked out and/or recapped.
Mondodimotori wrote on 2024-09-08, 15:03:
So... It also may be the case to get protective glasses and a fire extinguisher ready...?
If you ever get into the hobby (if we can call it that) of repairing power supplies, then safety/protective glasses should surely be on your list of items to have. Fire extinguisher - nah, that's over the top.
I remember when I tested my 1st recapped LCD monitor (internal PSU) and was so scared to power it on. I had safety glasses on, two bottles of water (in case of fire... and yes, electrical fire + water = not good... but only if things are plugged in and powered grinning face ), and one hand on the power cord, ready to yank it out at the slightest event (hence the water bottles.) Turns out... nothing wrong happened. My recapped supply worked just as it should... and still does to this day. grinning face
Mondodimotori wrote on 2024-09-08, 19:56:
EDIT: Here's an idea that went thorugh my mind recently: What about industrial power supply? Let's ignore the price for now (yikes), and let's look at the datasheet for this one: https://www.bicker.de/files/downloads/datenbl … t/bea-635_e.pdf
Just a waste of time and money to look into these, or there could be something here?
Mostly a waste of money.
A recapped OEM PSU (Delta, LiteOn, HiPro/Chicony, Bestec, AcBel) can provide more good years of service than you'll ever care to put on it.
Also, without seeing internal pictures of that "industrial" PSU, I can't really say how good it is.
In the modern world that we are in today, there's all kinds of companies. Some do really give you what they say they will, and others just slime and sleeze around, hoping to give you just "enough" of a product that you don't notice its inferior. So case in point with the above industrial PSU: it could be a really nice unit or it could be just a hair better than your typical gutless wonder crap PSU, with the company just hoping that nothing would go bad in the warranty that they give it and/or happily replacing it when it does.
Mondodimotori wrote on 2024-09-09, 13:39:
QUESTION: where would I need to look to find the electrical diagram of this PSU?
You probably won't.
Schematic diagrams for ATX PSUs weren't generally released by manufacturers to anyone. The only ones that can commonly be found are for the older half-bridge "generic / gutless wonder" designs, and these were typically dissected / reverse-engineered ones (i.e. someone had this PSU and drew the schematic by looking all of the parts and drawing it on paper / in CAD. I've done this to the 5VSB circuits of many of the cheaper / no-name ATX PSUs that use a 2-transistor design. It is time-consuming, though.
Mondodimotori wrote on 2024-09-09, 13:39:
For the new one, I dind't perform hard CPU stress test, like prime 95, but only 3D gaming and benchmark stress test, where the rail would drop to a sensor reported 4.66 (not read with mulitimeter) but the benchmarke kept running without issues for more than an hour. Of course, it may also be bad readings from the sensor, wouldn't be the firt time I encouter this (on the 370 mobo, the CPU temp sensor must be fucked because it always report 51°. No matter if immediately at boot or after hours of intense use. Ant it's a Celeron 700 with a decent cooler in it! With the case remaining nice and cool to the touch).
Yeah, motherboard sensors on old boards, particularly the voltage ones, are to be taken with a grain of salt (or more) usually.
If you didn't measure that voltage drop on the 5V rail with a multimeter, then it's not worth worrying about.
Mondodimotori wrote on 2024-09-09, 13:39:
Yeah, unfortunately I went and check even those top of the line 1000+ watt PSUs from reputable brands: All of them don't go over 20 amp on the 5V and 150 watts combined anymore
I keep seeing this here (Vogons) over and over again.
25 Amps or 30 Amps on the 5V rail is not some magic number that is absolutely needed to make a socket A system work. In fact, that's overkill.
In most cases, an honest 15-20 Amp rating ought to be just enough. Well, 15A might be a little low for the late + high-end Athlon XP builds (OK for a low-power Duron or low-end Athlon XP under 60W TDP). Dual socket A (Athlon MP/XP) is where a strong 5V rail really is needed. That or when a high-end socket A CPU is used in combination with these specific video cards: Radeon 9500/9700/9800 (R300-based), as these suck most of their power for the GPU from the 5V rail is well.
With that said, what is needed is a PSU that "prefers" to supply most of its power from the 5V rail rather than the 12V rail - i.e. 12V rail won't go out-of-spec high or 5V rail go out-of-spec low with a heavy 5V load.
Modern PSUs - be it "new" forward-converter -based designs or "old" half-bridge -based designs - are usually tweaked for heavier draw on the 12V rail. Case in point is this gutless wonder from 2019. Yeah, it's "new", alright - as in built recently. But it's based on an old half-bridge design, and group-regulated while at it (as is normally the case... which means the PSU can either be tweaked either for heavier 5V use, heavier 12V use, or somewhere in between with compromises.) In my tests, the PSU did very poorly with any 5V-"heavy" loading - the 12V rail drifted above 12.6V while the 5V rail sagged hard. So much for those "25 Amp" ratings on the label for the 3.3V and 5V rails. Then again, not like those numbers were ever possible, since the 5V and 3.3v rails each only have a 20 Amp rectifier.
On the other hand, this PSU did OK with a heavy 12V load. IIRC, I tested it with up to about 18 Amps on the 12V rail (over 200 Watts) and the 12V rail was happy while the 5V rail didn't climb up too much. So this PSU would do acceptably well for a modern low-ish power PC.
Thus I stand by my argument that I made in another thread here somewhere, that high-current ratings on the 3.3V and 5V rails of modern PSUs are either bullshit or not achievable in any practical scenarios... or at best, just not great for your PSU long-tern (i.e. in cases where the 5V and 3.3V rail are DC-DC based from the 12V rail via buck-regulator modules, as the modules themselves may be prone to over-heating when being loaded down near their rated specs.)
DudeFace wrote on 2024-09-09, 14:07:
rolling on the floor laughing It's like a name mixup of Arctic Silver and FSP BlueStorm series PSUs. rolling on the floor laughing
Probably build by the same Chinese factory that sells the same PSU to 100's of other "no name" companies, each with their own name/logo/brand slapped on the label.
Not necessarily saying it's a bad PSU, though. How capable it is depends on how much the reseller / company on the label paid to the Chinese factory for its particular configuration.
Of course, even the best builds will be full of cheap electrolytic caps. So past 3-4 years of use (if you are lucky), you may want to pull it out of service and recap it.