R1 is a 0.6Mohm resistor under the Cx1 x2 type capacitor, 0.22uF 270V

Hmm, that's weird. On mine there's no resistor under the yellow cap. Maybe different revisions? (mines rev 2)
I have finished putting the PSU back together, albeit without the secondary transistors, and the bulb flashes once on power up (normal). Not continuously on.
What's weird is that the 5VSB is nonwhere to be seen.... From my understanding the 5VSB here is generated directly from the primary side, and is isolated from main rail generation (ie. 3.3v/5v etc...).
Actually scratch that, I'm still missing an TO 92 transistor which blew up near the PWM ic. Can't find the markings on what it is tho...
DAVE86, could you do me a favour once again and tell me the model of the one I marked in the attached photo?

IC4 is a TL431, AS431, etc. integrated voltage reference in the feedback circuit.

Oh, so that's not an transistor....
Could you check if it is related to 5vsb rail, please?

kotel wrote on 2025-02-22, 20:28:

Could you check if it is related to 5vsb rail, please?

Scrach that, it is on the 5vsb rail.

I have finally replaced all of the components damaged by the 5vsb carnage and its completely dead. No 5VSB and PS_ON. Just dead. Dim bulb tester only flash once when power is applied which would meant the primary side caps are charging, albeit something is still screwed up.
Any ideas?

EDIT: The 5vsb mosfet burned and cracked. The giveaway was one of the pins was shorted in circuit, so I pulled it out, inspected and there was a big crack going all around the case. Weird how it still got detected as an mosfet on my lcr t7.

Injecting 5vsb makes the PWM IC generate PS_ON, but shorting PS_ON to GND does nothing. Primary side should work correctly (sparks flew while I probed one of the caps after disconnecting mains, either because of me being an dumbass and shorting two different pads or black MM probe being connected to gnd).

2-transistor 5VSB circuits can get pretty tricky to troubleshoot, if not some of the most tricky circuits. Reason being: when one component goes bad, the entire circuit suffers cascade damage. Sometimes components will fail in a way that's really hard to detect. The end-game, more or less, is if the 5VSB transformer primary winding either opens or develops shorted turns. If it opens, this should be easy to find (no continuity between the pin that goes to the positive high voltage bus on the mains caps and the MOSFET's Drain pin.) But if the 5VSB traffo develops a shorted winding, there won't be any way to tell, other than opening the transformer. That said, this is a rare failure.
Since your 5VSB MOSFET is damaged, probably also its Source-connected resistor is open... and likely the smaller resistor that goes from the Source resistor to the Base of the small TO-90 transistor that drives the MOSFET off. In fact, right in your 1st post you stated the little transistor was damaged, so that means the MOSFET was damaged from the start. When that happens, the Source resistor opens and then the only path is through the small resistor and into the Base of the small transistor, which then makes it die due to the high voltage.

So again, it's cascaded damage, and just about everything will need to be replaced on the primary side of the 5VSB. Even the optocoupler could be damaged. Dim bulb tester is a must afterwards. That said, if you do remove the 5VSB MOSFET or if the 5VSB MOSFET's Source resistor is open, be careful around the big caps on the primary after you've plugged in the PSU into AC for a test. With no 5VSB circuit, those large caps will only discharge through their balancing resistors that are across them (typically around 300 kOhms), and that may take a minute or two... and still have enough residual voltage to damage other parts or your transistor tester if you try to measure something in circuit.

Anyways... it's probably still fixable, but I'd say this one is a more "advanced" electronics repair lesson, so maybe come back another time at it.

I'll also echo again what I said in my first post on the previous page about getting this PSU fixed "quick and dirty" 😉 : get a 5V 2A adapter and a 12-15V 1 Amp adapter. Use the 5V to backfeed the 5VSB line and the 12-15V to backfeed into the secondary side auxiliary rail... and that should allow you to test if the main PS works (short PS_ON to ground and see.) If yes, just find a place somewhere to "shove in" those adapters in the PSU, or leave them outside, attached to the back and wire everything with long wires... and call it done. 😁

momaka wrote on 2025-02-23, 19:31:

2-transistor 5VSB circuits can get pretty tricky to troubleshoot, if not some of the most tricky circuits. Reason being: when on […]
Show full quote

2-transistor 5VSB circuits can get pretty tricky to troubleshoot, if not some of the most tricky circuits. Reason being: when one component goes bad, the entire circuit suffers cascade damage. Sometimes components will fail in a way that's really hard to detect. The end-game, more or less, is if the 5VSB transformer primary winding either opens or develops shorted turns. If it opens, this should be easy to find (no continuity between the pin that goes to the positive high voltage bus on the mains caps and the MOSFET's Drain pin.) But if the 5VSB traffo develops a shorted winding, there won't be any way to tell, other than opening the transformer. That said, this is a rare failure.
Since your 5VSB MOSFET is damaged, probably also its Source-connected resistor is open... and likely the smaller resistor that goes from the Source resistor to the Base of the small TO-90 transistor that drives the MOSFET off. In fact, right in your 1st post you stated the little transistor was damaged, so that means the MOSFET was damaged from the start. When that happens, the Source resistor opens and then the only path is through the small resistor and into the Base of the small transistor, which then makes it die due to the high voltage.

So again, it's cascaded damage, and just about everything will need to be replaced on the primary side of the 5VSB. Even the optocoupler could be damaged. Dim bulb tester is a must afterwards. That said, if you do remove the 5VSB MOSFET or if the 5VSB MOSFET's Source resistor is open, be careful around the big caps on the primary after you've plugged in the PSU into AC for a test. With no 5VSB circuit, those large caps will only discharge through their balancing resistors that are across them (typically around 300 kOhms), and that may take a minute or two... and still have enough residual voltage to damage other parts or your transistor tester if you try to measure something in circuit.

Anyways... it's probably still fixable, but I'd say this one is a more "advanced" electronics repair lesson, so maybe come back another time at it.

I'll also echo again what I said in my first post on the previous page about getting this PSU fixed "quick and dirty" 😉 : get a 5V 2A adapter and a 12-15V 1 Amp adapter. Use the 5V to backfeed the 5VSB line and the 12-15V to backfeed into the secondary side auxiliary rail... and that should allow you to test if the main PS works (short PS_ON to ground and see.) If yes, just find a place somewhere to "shove in" those adapters in the PSU, or leave them outside, attached to the back and wire everything with long wires... and call it done. 😁

Hmm, so backfeeding 5vsb on its own won't make the PWM chip work? Weird... While I'm backfeeding it I see the PS_ON come up to 4.89v, but when I try to short it, nothing happens...
I have double checked and pin 1 of PWM IC should be around 5v and is directly connected to 5VSB by a few jumpers here and there.

^ It will make the PWM chip work, but the two little TO-92 transistors that drive the middle transformer are powered from another rail that's produced by the 5VSB transformer: the secondary side auxiliary rail (or sec. aux., for short), which is around 10-16V typically.
So when you only backfeed 5V into 5VSB, your PWM chip will be ready to do it's job, but those smaller transistors won't. That's why you need a 2nd power adapter (12-15V) to inject / backfeed into the sec. aux. rail so that the these two transistors can also get powered.
Do that, and you should see the main PS come up (if nothing else is bad, of course. 😁 )

May I know where's the sec aux rail? I cannot find it.

momaka wrote on 2025-02-23, 20:48:

^ It will make the PWM chip work, but the two little TO-92 transistors that drive the middle transformer are powered from another rail that's produced by the 5VSB transformer: the secondary side auxiliary rail (or sec. aux., for short), which is around 10-16V typically.
So when you only backfeed 5V into 5VSB, your PWM chip will be ready to do it's job, but those smaller transistors won't. That's why you need a 2nd power adapter (12-15V) to inject / backfeed into the sec. aux. rail so that the these two transistors can also get powered.
Do that, and you should see the main PS come up (if nothing else is bad, of course. 😁 )

I habe double retraced the connections from 5vsb transformator and none of them seem to be tied to the two small TO-92 transistors near PWM IC, therefore I have no clue where to inject that voltage. They are also the closest to the 5vsb transformator too.

kotel wrote on 2025-02-24, 04:40:

May I know where's the sec aux rail? I cannot find it.
....
please don't smoke me for the trace repair

Hehe, no worries, that trace repair looks quite alright. 😀
I might have a grudge with the quality of the pictures, though. 🤣
Trying my best to work it out right now, but only really the bottom (solder-side picture is clear enough to trace anything.

That said, if you look on the 5VSB transformer, you will see 3 pins on the secondary side.
Normally, one is ground, one is the output for the 5VSB (goes to the Anode of the 5VSB diode), and one is the output of the sec. aux. rail (goes to the anode of a smaller diode.)
But occasionally, the 5VSB rectifying diode might be installed on the grounding pin of the 5VSB transformer, with the anode connected to ground and cathode connected to the grounding pin... and then the 5VSB output pin goes directly to the 5VSB. The sec. aux. rail will still have its own diode and connected the same way as before (i.e. anode to sec. aux. output pin and cathode to the actual sec. aux. rail.)

Since I don't know which configuration you might have here (and since your 5VSB is pretty dead anyways), go ahead and remove the 5VSB transformer out of the PSU. After this, use your multimeter to see which pin on the board connects to ground (if any) and which pin connects to the 5VSB anode (if any) or cathode (if any.)

From what I can trace from this picture of yours and looking at the pins of the 5VSB transformer, it looks like you might have case #2 above, where the 5VSB diode is on the grounding pin. In any case, this grounding pin is the right-most pin. I can't tell what the other two are, so either you will need to do a little hunting / figuring out for me or try to provide another photo of the top-side of the board that's more clear and with better lighting (consider using flash on your camera, if you don't have much light in the room.) Anyways, if you do find the diode for the sec. aux rail (again, it will have its anode connected to one of the other two pins of the 5VSB transformer), you can insert / backfeed the 12V from your external PSU at the cathode of that diode.

kotel wrote on 2025-02-24, 18:01:

I habe double retraced the connections from 5vsb transformator and none of them seem to be tied to the two small TO-92 transistors near PWM IC, therefore I have no clue where to inject that voltage. They are also the closest to the 5vsb transformator too.

Yes, the driving transistors won't be directly tied to the sec. aux. rail. Instead, their Bases are pulled up via high(is)-value resistors, typically in the range of 1 to 2.7 kOhms.

momaka wrote on 2025-02-26, 00:09:

Hehe, no worries, that trace repair looks quite alright. :) I might have a grudge with the quality of the pictures, though. LOL […]
Show full quote

kotel wrote on 2025-02-24, 04:40:

May I know where's the sec aux rail? I cannot find it.
....
please don't smoke me for the trace repair

Hehe, no worries, that trace repair looks quite alright. 😀
I might have a grudge with the quality of the pictures, though. 🤣
Trying my best to work it out right now, but only really the bottom (solder-side picture is clear enough to trace anything.

Yeah, the picture quality is nothing past amazing :\ Probably due to too much light and the fact I need to use outlook compression system for them to fit max size of 5mb. Zipping them also doesn't help. I could upload them to google drive and then post the link here, although I don't know if that's convenient for you.
I'll try my best to redo the pictures and trace where the other two pins of the mosfet go. Sadly my local hardware store doesn't have any MOSFET's or transistors, so the only way I can get them is either ali or TME.

kotel wrote on 2025-02-26, 08:41:

Yeah, the picture quality is nothing past amazing :\ Probably due to too much light and the fact I need to use outlook compression system for them to fit max size of 5mb. Zipping them also doesn't help. I could upload them to google drive and then post the link here, although I don't know if that's convenient for you.
I'll try my best to redo the pictures and trace where the other two pins of the mosfet go.

It's not a matter of picture size or resolution, but more of a poor focus and not great light.
When I annotated your picture in the Tagan PSU thread on BCN, note how much of the image I cropped away.
I almost never use the largest image format of my camera (be it my phone or my digital cam) at the time of taking the picture. That way, I'm not taking huge images and then having to deal with huge image and file sizes that don't really add anything to the quality. For most of my PSU pics, I use 2 MP resolution (1600x1200). Only for the bottom / solder-side shots I might bump that up to 3M or 5M, if there are many tiny SMD components. And once I'm done taking the photos, I always post-process them with software (I use an old version of Photoshop - 7.0... but Paint.NET is a pretty good free replacement) to crop away any unnecessary areas and reduce the image size.

Also, if anything, I think your pictures could use more light. 😀 It's best to take shots near a window where you have natural light (usually during the day). If you don't have that, perhaps you can take the pictures outside. And if that's not an option either, then use the flash function of your camera as a last resort. Just beware that flash does have some side effects on the photos. For example, if you take the photos straight-down, a lot of times the PCB and some shinier components will flash back at the lens, making things hard to see. At the same time, if you angle the camera too much, the components will cast long shadows from the flash, and that also will make it hard to see things on the board. So the optimum is usually almost-straight-down shots, but angled ever so slightly to minimize flash back. Now, for solder-side shots, that's one place where I *always* turn off camera flash. Regardless if you're using a phone or a stand-alone digital camera, the flash is a point-source of light, so it's very hard to take a nice shot with even light everywhere. So for solder-side shots, I always take my device near a window and take the shots there. The difference is night and day in the clarity.

Speaking of clear images... I think I also have an FSP PSU that's very similar to yours here: a Sparkle Power FSP250-60ATV. Post on BCN can be found here (scroll to post #3122 on that page... should be the 2nd post from top.)
https://www.badcaps.net/forum/troubleshooting … 478#post1520478
... so lucky for you, I may be able to use my own images again. 🤣
Well, OK, not sure if everything matches between these two units, so that you would have to verify. Also, I forgot where / in which PC I have it in storage.

*EDIT*
OK, I had a closer look, and it seems my Sparkle Power PSU PCB matches yours and is even the same revision (2). Using a combination of your and my pictures, I think I've figured out where the sec. aux. rail is. See attached image. If backfeeding power into the sec. aux. rail, do so at the cathode of the diode that I drew with *yellow* on the annotated picture. Also, the yellow circle is the filter cap for the aux. sec. rail.

kotel wrote on 2025-02-26, 08:41:

Sadly my local hardware store doesn't have any MOSFET's or transistors, so the only way I can get them is either ali or TME.

Or other scrap electronics? 😉
Check if you have any scrap metal recyclers locally, and if you do, ask them if when they get circuit boards to buy them.
I still haven't done that here, but I have found a flea market close by that has all kinds of scrap / junk for sale, often including scrap circuit boards from TVs, audio systems, PCs, and etc. The PSUs from LCD power supplies often have a pretty good selection of stuff on them. And ATX PSUs I can often get for $0.50 to $1 - cheaper than buying a set of transistors on AE or anywhere else! 🤣

momaka wrote on 2025-02-27, 11:27:

It's not a matter of picture size or resolution, but more of a poor focus and not great light. When I annotated your picture in […]
Show full quote

kotel wrote on 2025-02-26, 08:41:

Yeah, the picture quality is nothing past amazing :\ Probably due to too much light and the fact I need to use outlook compression system for them to fit max size of 5mb. Zipping them also doesn't help. I could upload them to google drive and then post the link here, although I don't know if that's convenient for you.
I'll try my best to redo the pictures and trace where the other two pins of the mosfet go.

It's not a matter of picture size or resolution, but more of a poor focus and not great light.
When I annotated your picture in the Tagan PSU thread on BCN, note how much of the image I cropped away.
I almost never use the largest image format of my camera (be it my phone or my digital cam) at the time of taking the picture. That way, I'm not taking huge images and then having to deal with huge image and file sizes that don't really add anything to the quality. For most of my PSU pics, I use 2 MP resolution (1600x1200). Only for the bottom / solder-side shots I might bump that up to 3M or 5M, if there are many tiny SMD components. And once I'm done taking the photos, I always post-process them with software (I use an old version of Photoshop - 7.0... but Paint.NET is a pretty good free replacement) to crop away any unnecessary areas and reduce the image size.

Also, if anything, I think your pictures could use more light. 😀 It's best to take shots near a window where you have natural light (usually during the day). If you don't have that, perhaps you can take the pictures outside. And if that's not an option either, then use the flash function of your camera as a last resort. Just beware that flash does have some side effects on the photos. For example, if you take the photos straight-down, a lot of times the PCB and some shinier components will flash back at the lens, making things hard to see. At the same time, if you angle the camera too much, the components will cast long shadows from the flash, and that also will make it hard to see things on the board. So the optimum is usually almost-straight-down shots, but angled ever so slightly to minimize flash back. Now, for solder-side shots, that's one place where I *always* turn off camera flash. Regardless if you're using a phone or a stand-alone digital camera, the flash is a point-source of light, so it's very hard to take a nice shot with even light everywhere. So for solder-side shots, I always take my device near a window and take the shots there. The difference is night and day in the clarity.

I might actually try near a window, but I am working most, if not all of the time in my shop where I have no access to natural light :\ I have seen that this phone cam can become quite picky at what it wants to focus when there's too much light. Also, thanks for the tip on taking and editing pictures 😀

momaka wrote on 2025-02-27, 11:27:

*EDIT*
OK, I had a closer look, and it seems my Sparkle Power PSU PCB matches yours and is even the same revision (2). Using a combination of your and my pictures, I think I've figured out where the sec. aux. rail is. See attached image. If backfeeding power into the sec. aux. rail, do so at the cathode of the diode that I drew with *yellow* on the annotated picture. Also, the yellow circle is the filter cap for the aux. sec. rail.

Thank you! I will attempt to backfeed 10v to it, and if that doesn't satisfy the PSU I'll step it up to 12v and maybe 14v as a last resort.

momaka wrote on 2025-02-27, 11:27:

Or other scrap electronics? ;) Check if you have any scrap metal recyclers locally, and if you do, ask them if when they get cir […]
Show full quote

kotel wrote on 2025-02-26, 08:41:

Sadly my local hardware store doesn't have any MOSFET's or transistors, so the only way I can get them is either ali or TME.

Or other scrap electronics? 😉
Check if you have any scrap metal recyclers locally, and if you do, ask them if when they get circuit boards to buy them.
I still haven't done that here, but I have found a flea market close by that has all kinds of scrap / junk for sale, often including scrap circuit boards from TVs, audio systems, PCs, and etc. The PSUs from LCD power supplies often have a pretty good selection of stuff on them. And ATX PSUs I can often get for $0.50 to $1 - cheaper than buying a set of transistors on AE or anywhere else! 🤣

Well, there is an scrapper near me (around like 5-10km IIRC) although that one has become "screwny" if you get what I mean. And besides I'd need to arrange a meeting at his house at least 2 weeks in advance...
Then there's the next one which has a crap ton of PC's and PC parts, although that one is pretty much on the other side of poland and he want's more money per PC than the mentioned above.
And finally there is one more hw store near me, although that one had mostly RTV accessories but he did have an wall full of capacitors, but low chances he would have the 4n60b 🙁

I have tried injecting 10v into the cathode of that diode and it was pulling max set amperage (690mA, then 1A, then 2A) with its voltage down to 28mV on 2A. I suspect its because the PSU generates a higher voltage on that line, although before I try injecting 14v I would like to consult the profesionals if this behavior is normal.

kotel wrote on 2025-02-28, 17:35:

I have tried injecting 10v into the cathode of that diode and it was pulling max set amperage (690mA, then 1A, then 2A) with its voltage down to 28mV on 2A. I suspect its because the PSU generates a higher voltage on that line, although before I try injecting 14v I would like to consult the profesionals if this behavior is normal.

Nope, that's not normal.

That said, did you check if the orientation of that diode I annotated in yellow (for the sec. aux. rail) on the picture actually matches the one on your board? Just want to make sure I didn't make a mistake somewhere with that. Also, I'm sure you know this, but also just checking things on your side (one can never check things too many times 😉 ) - the cathode side of the diode is the one with the band marked on the diode. If all of this checks out, then either that diode for the sec. aux. rail is faulty or your TO-92 driver transistors (Q4 and Q5, I think) are bad, or their protection diodes (D11 and D26??) are bad. Remove and test each and report what you find in detail (mV readings, and etc.)

Also, since your 5VSB had quite the blowout, the 5VSB transformer might be damaged / bad. Normally injecting 5V at 5VSB and 10-15V at the sec. aux. rail should *not* affect anything even if the 5VSB transformer is bad, since the 5VSB and sec. aux. rectifying diodes shouldn't allow reverse current... but just in case, remove the 5VSB transformer and see if the high current goes away.

In fact, here's how to proceed:
1) remove 5VSB transformer and inject 5V into 5VSB and 10-15V into sec. aux. rail. Is the current draw on sec. aux. rail still high (over 500 mA)? If yes, proceed to 2) below.
2) remove the two 1-KOhm resistors that are next to driver transistor Q5. Inject 5V and 10-15V into 5VSB and sec. aux. rail again. Is the current still high? If yes, proceed to 3) below.
3) Remove driver transistors Q4 and Q5. Inject the voltages again. Is the current still high? If yes, proceed to 4) below.
4) Remove the protection diodes (D11 and D26) that are close to the spots for transistors Q4 and Q5. Inject voltage again. Is the current still high? If yes, proceed to 5) below.
5) Remove the sec. aux. rail rectifying diode (one annotated with yellow in my diagram, not sure what designator this is on the board). Inject the voltages again. Current still high? If yes, go to 6) below.
6) There's really not other components that can cause a high current draw, so if you've reached this step, double-check all your work again, because you've probably made an error somewhere.

Report back to what step you got and any additional information.

momaka wrote on 2025-02-28, 19:56:

Nope, that's not normal. […]
Show full quote

kotel wrote on 2025-02-28, 17:35:

I have tried injecting 10v into the cathode of that diode and it was pulling max set amperage (690mA, then 1A, then 2A) with its voltage down to 28mV on 2A. I suspect its because the PSU generates a higher voltage on that line, although before I try injecting 14v I would like to consult the profesionals if this behavior is normal.

Nope, that's not normal.

That said, did you check if the orientation of that diode I annotated in yellow (for the sec. aux. rail) on the picture actually matches the one on your board? Just want to make sure I didn't make a mistake somewhere with that. Also, I'm sure you know this, but also just checking things on your side (one can never check things too many times 😉 ) - the cathode side of the diode is the one with the band marked on the diode. If all of this checks out, then either that diode for the sec. aux. rail is faulty or your TO-92 driver transistors (Q4 and Q5, I think) are bad, or their protection diodes (D11 and D26??) are bad. Remove and test each and report what you find in detail (mV readings, and etc.)

Also, since your 5VSB had quite the blowout, the 5VSB transformer might be damaged / bad. Normally injecting 5V at 5VSB and 10-15V at the sec. aux. rail should *not* affect anything even if the 5VSB transformer is bad, since the 5VSB and sec. aux. rectifying diodes shouldn't allow reverse current... but just in case, remove the 5VSB transformer and see if the high current goes away.

Oh.... the cathode is where the band is.... fudge.... This is my third screw-up today.... Dropping my asus board and now it having an tantrum everytime it boots, then an deer clone (almost) killing my asrock board due to 5vsb becoming insane and now this......
Yes, the diode orientation on my board is like you've wrote.
I don't think this screw-up would have caused more damage, but who knows...

Okay I corrected my mistake and it lives!
Although the rails are a tad bit of... 12v is 11.84v, 5v is 5.34v, 3.3v is within spec (3.4v), -5v is -4.60v and -12v is -10.6v. All are unloaded. That might be due to the small primary caps going dry and/or the dim bulb protection being to powerful for this PSU (although the bulbs only flash once when the primary caps are being charged, while running they are off). The rest of the secondary were pretty much all OST's which are fairly good IMO.

I guess all that's left is to buy an 4N60B MOSFET? Or are there any equivalents?

kotel wrote on 2025-03-01, 20:20:

Okay I corrected my mistake and it lives!

Yay, congrats, one up for the count... well almost there anyways. 😉

kotel wrote on 2025-03-01, 20:20:

Although the rails are a tad bit of... 12v is 11.84v, 5v is 5.34v, 3.3v is within spec (3.4v), -5v is -4.60v and -12v is -10.6v. All are unloaded. That might be due to the small primary caps going dry and/or the dim bulb protection being to powerful for this PSU (although the bulbs only flash once when the primary caps are being charged, while running they are off).

I haven't ran my Sparkle 250W unloaded, but I think these voltages are normal without a load. It's an old and 5V-heavy design, so if you put at least 1-2 Amps of load on the 5V rail, and at least 0.5 Amps on the 12V rail, the voltages should stabilize pretty close to their nominal values.

The dim bulb tester won't cause the voltages to change like this. If it's not powerful enough, all of the voltages (including the 5V rail) will start to sag by equal % relative to nominal. The fact that 5V is high and the rest are low is the PSU telling you, I need more load on the 5V rail, please.

The small primary caps should always be changed or at least checked. For the two that are part of the Base drive for the BJTs (typically small 50V caps with rating of 1 uF, 2.2 uF, 4.7 uF, or 10 uF), if these go bad, the BJTs won't be turning on as quickly, so they will waste more power and may even overheat. And if the 5VSB has a critical cap (it appears that this PSU doesn't, though), that should always be checked or changed to something very reliable.

kotel wrote on 2025-03-01, 20:20:

The rest of the secondary were pretty much all OST's which are fairly good IMO.

For the most part, they are... until one fail silently. :p:
So I advise to pull them out and check them, just in case.

kotel wrote on 2025-03-01, 20:20:

I guess all that's left is to buy an 4N60B MOSFET? Or are there any equivalents?

These 2-transistor 5VSB circuits are not picky at all, so anything relatively equivalent should work.
MOSFET just needs to be rated for at least 500-600V Drain-to-Source voltage, 3-6 Amps continuous Drain current, and similar Gate-to-Source voltage (typically 30V or more). The rest won't matter that much.

That said, if you're expecting to just get away with replacing that MOSFET to get the 5VSB running... good luck! 😁
The small helper transistor will also likely be damaged, as well as the MOSFET's Source resistor... and many many other things. There's also always the chance the 5VSB transformer's primary winding could have shorted. So if you do replace all faulty components and are ready to test, definitely use the dim bulb tester... and nothing more than a 100W bulb. Heck, downsizing to the 60W bulb would be even more reasonable.

But before doing even any of that testing, it might be a good idea to remove power to the main PWM IC so that you don't kill it, should the 5VSB circuit output voltage go rogue. To do that, remove jumper J12. Also remove driver transistors Q4 and Q5, along with their pull-up 1-KOhm resistors. Doing this should be enough to protect your PWM IC.

momaka wrote on 2025-03-01, 20:53:

kotel wrote on 2025-03-01, 20:20:

Okay I corrected my mistake and it lives!

Yay, congrats, one up for the count... well almost there anyways. 😉

Was hoping the tagan would get fixed firstly, but I take whatever I can fix. But still, knowing my skill it's a very big leap in the PSU repair stuff 😀

momaka wrote on 2025-03-01, 20:53:

kotel wrote on 2025-03-01, 20:20:

Although the rails are a tad bit of... 12v is 11.84v, 5v is 5.34v, 3.3v is within spec (3.4v), -5v is -4.60v and -12v is -10.6v. All are unloaded. That might be due to the small primary caps going dry and/or the dim bulb protection being to powerful for this PSU (although the bulbs only flash once when the primary caps are being charged, while running they are off).

I haven't ran my Sparkle 250W unloaded, but I think these voltages are normal without a load. It's an old and 5V-heavy design, so if you put at least 1-2 Amps of load on the 5V rail, and at least 0.5 Amps on the 12V rail, the voltages should stabilize pretty close to their nominal values.

The dim bulb tester won't cause the voltages to change like this. If it's not powerful enough, all of the voltages (including the 5V rail) will start to sag by equal % relative to nominal. The fact that 5V is high and the rest are low is the PSU telling you, I need more load on the 5V rail, please.

Aha, got it. I'll try to put an 35w 12v bulb on 5v and see if that does anything.

momaka wrote on 2025-03-01, 20:53:

The small primary caps should always be changed or at least checked. For the two that are part of the Base drive for the BJTs (typically small 50V caps with rating of 1 uF, 2.2 uF, 4.7 uF, or 10 uF), if these go bad, the BJTs won't be turning on as quickly, so they will waste more power and may even overheat. And if the 5VSB has a critical cap (it appears that this PSU doesn't, though), that should always be checked or changed to something very reliable.

May I know how to see if they are okay? Since those are generic purpose OST or teapo or whatever FSP used don't give info about normal ESR readings. And now I got told from TRW server that I should just see if the capacitance is in spec. If it is then keep, if it isn't then bin.
This whole time I was using an ESR spreadsheet for the different voltage and capacitance and idk if to still use it. Sure, it isn't 99% accurate but it at least gives me an rough average example of what's within "tolerance".

momaka wrote on 2025-03-01, 20:53:

kotel wrote on 2025-03-01, 20:20:

The rest of the secondary were pretty much all OST's which are fairly good IMO.

For the most part, they are... until one fail silently. :p:
So I advise to pull them out and check them, just in case.

Yep, will do that but pretty much most of them show 0 ohm esr on my lcr t7 and capacitance is within spec on the big secondary ones. The smaller are yet to be tested.

momaka wrote on 2025-03-01, 20:53:

These 2-transistor 5VSB circuits are not picky at all, so anything relatively equivalent should work. MOSFET just needs to be ra […]
Show full quote

kotel wrote on 2025-03-01, 20:20:

I guess all that's left is to buy an 4N60B MOSFET? Or are there any equivalents?

These 2-transistor 5VSB circuits are not picky at all, so anything relatively equivalent should work.
MOSFET just needs to be rated for at least 500-600V Drain-to-Source voltage, 3-6 Amps continuous Drain current, and similar Gate-to-Source voltage (typically 30V or more). The rest won't matter that much.

That said, if you're expecting to just get away with replacing that MOSFET to get the 5VSB running... good luck! 😁
The small helper transistor will also likely be damaged, as well as the MOSFET's Source resistor... and many many other things. There's also always the chance the 5VSB transformer's primary winding could have shorted. So if you do replace all faulty components and are ready to test, definitely use the dim bulb tester... and nothing more than a 100W bulb. Heck, downsizing to the 60W bulb would be even more reasonable.

But before doing even any of that testing, it might be a good idea to remove power to the main PWM IC so that you don't kill it, should the 5VSB circuit output voltage go rogue. To do that, remove jumper J12. Also remove driver transistors Q4 and Q5, along with their pull-up 1-KOhm resistors. Doing this should be enough to protect your PWM IC.

I have found two replacement MOSFET's on TME. One is an P3F60HP2-5600 and the other is an WML04N60C2-CYG. Would any of these fit the required specification?

Yeah, I am quite worried the 5VSB MOSFET won't be the only issue that's left.... Especially since I inject 5VSB at the ATX connector. Although all of the primary side resistors had been replaced with its 4148's and the transistor near the MOSFET looks to be okay. I should inspect it carefully knowing what happened to this MOSFET tho 😅.....

As for the transformator are there any ways I can detect an short without disasembling it or testing it on an live circuit? I should have some left-over from dead LC clones and one from an dead delta 500w PSU.