the gutless, bloated, and fried power supply hall of shame

[PeteS in CA]

Re: the gutless, bloated, and fried power supply hall of shame

If that thing is identical to the L-LINK you linked, it's basically a 150W power supply. I doubt the main transformer core is good for 200W, and those input electrolytics might have decent filtering and hold-up for 200W.

[kevin!]

Re: the gutless, bloated, and fried power supply hall of shame

Recently a client brought me a computer that has a corrupted HDD (it had 1000 reallocated sectors in the SMART), and the computer had this 3GO installed. I think the PSU is to blame for the malfunction of the HDD, luckily the client was able to recover all the data from the HDD.
The 5VBS voltage has such a terrible ripple that the multimeter is able to perceive it (the last 2 photos), the 5V line is out of specifications (5.28v) (first photo).

I can't disassemble it because the customer wants to return it to the store where I buy the computer and I can't open it.

Internally it is the same as the L-LINK in this link:
https://www.badcaps.net/forum/showpo...postcount=1477

[PeteS in CA]

Re: the gutless, bloated, and fried power supply hall of shame

Yeah, that 150W P/S is not very safe, given the GP ceramic caps where safety-rated Y-caps should be. The 330uF primary lytics and that gimmicked 12V "rectifier" pretty much make 150W the realistic rating.

[momaka]

Re: the gutless, bloated, and fried power supply hall of shame

Glad you guys got a laugh out of it Yeah, I used to do a lot of stupid sh!t like this when I was a kid.

Didn't we all?

I like to think I outgrew the "mad scientist" phase, but sometimes remnants of it still manage to surface in some of my projects, though not to this degree

Oh see, when I was a kid, I was too "afraid" to mess with most of my working stuff. But now - not anymore. So my "mad scientist" phase has just begun. I'm sure you've noticed that in my long BCN posts that often boast about restoring some junk no one even cares about and with materials I had on hand or salvaged from something else. Don't know why it gives me so much satisfaction.

Great ! The more you learn So it's like an isolation transformer so the IC doesn't drive the transistors directly ?

You nailed it! That is actually the -correct- name for it. I just couldn't remember what it was called.

Haven't had the patience to look up the datasheet (if any) for that IC to see what it does, but it must be a combined PWM driver-supervisor IC then, since it performs both tasks from the secondary

Yup.
Essentially, those Deer CotY ICs ("Chip-of-the-Year") are TL494/KA7500 + some logic for supervision. Same applies for AT2005b, ATX2005, SG6105, and the earlier Deer DR____ variants.

When you see one of these in a PSU, there's a 99.999% chance it's a half-bridge topology.

most SMPSs have the IC on the primary.

No, not really.
It all depends on the topology used.
For half-bridge topology, IC is almost always on the secondary side. Half-bridge topology started in the 70's if I am not mistaken and was widely used up until the mid 90's for PC PSUs. The gutless wonders continued using it even longer (some still do!) because the design is so well understood now.

But anything more recent and decent nowadays typically uses current-mode PWM (i.e. ICs similar to uc384x family and CM680x), and for those, your PWM IC is usually on the primary side.

Then there are self-resonant and other kind of topologies now... but I won't go into those.

Also, what about the STBY output on cheapo's like these? There is an optocoupler there, which I'm assuming is for regulation of the STBY output, but I don't see any driver IC on the primary, so it's like "dumb" regulation.

"Dumb" regulation may not be the technical term for it, but that's a pretty accurate description, as they often don't have any form of protection, other than self-limiting overload.

Typically, those 5VSB circuits with an optocoupler and no IC on the primary are 2-transistor self-oscillating circuits (but beware that some ICs, like the TOP-200 series will look like a MOSFET with 3 pins, yet is actually an IC.)

The next step up I've seen from this is a separate PWM IC on the primary for the 5VSB output, though still no IC for the 5/12v output as well - that is still being driven by that third transformer.

Don't mix the standby (5VSB) and 5V/12V (or 12V only for newer PSUs) sections. They are two separate circuits, independent of each other. You could have a 5VSB with IC and still have an old half-bridge design for the 5V/12V PS. Vice versa too.

[Dannyx]

Re: the gutless, bloated, and fried power supply hall of shame

Glad you guys got a laugh out of it Yeah, I used to do a lot of stupid sh!t like this when I was a kid. I like to think I outgrew the "mad scientist" phase, but sometimes remnants of it still manage to surface in some of my projects, though not to this degree

I can't believe someone went to great lengths to make a Deer modular

From what I recall, it wasn't easy either, so props for that too

I used to think I was a good runner-up for King-of-the-ghetto. But after seeing this, I don't think it's even a competition anymore. You sir, shall take that crown

Those XLR jacks are indeed very unique. Never seen a PSU done like that before. Funny thing: just those 3 jacks probably cost 10x more than the PSU itself!

Oh, and that hand-drawn "shock hazard" symbol.... it's probably not too far from telling the truth what would happen to you eventually as you keep using this POS PSU.

Pffftt, yeah that should totally be a contest here on BC with several categories, including raspberries like these

Driving the BJTs, as Dan81 said.

If you notice, the PWM IC is on the secondary side. So the only way it can drive the BJTs is through that transformer. Optocouplers won't work, because BJTs are current-driven devices.

Great ! The more you learn So it's like an isolation transformer so the IC doesn't drive the transistors directly ? Haven't had the patience to look up the datasheet (if any) for that IC to see what it does, but it must be a combined PWM driver-supervisor IC then, since it performs both tasks from the secondary - most SMPSs have the IC on the primary.
Also, what about the STBY output on cheapo's like these? There is an optocoupler there, which I'm assuming is for regulation of the STBY output, but I don't see any driver IC on the primary, so it's like "dumb" regulation. The next step up I've seen from this is a separate PWM IC on the primary for the 5VSB output, though still no IC for the 5/12v output as well - that is still being driven by that third transformer. The next step up from that is of course "full" regulation, or even individual buck converter risers for the 5/3.3v outputs on the secondary, like in my Corsair.

[momaka]

Re: the gutless, bloated, and fried power supply hall of shame

Reaction to that 400W Premier:

WOW!

I botched the heck out of it over the years, adding XLR jacks and removing the 20P connector to make it "modular". Look at that "custom" case work - state of the art, isn't it ?

YES!

I used to think I was a good runner-up for King-of-the-ghetto. But after seeing this, I don't think it's even a competition anymore. You sir, shall take that crown.

Those XLR jacks are indeed very unique. Never seen a PSU done like that before. Funny thing: just those 3 jacks probably cost 10x more than the PSU itself!

Oh, and that hand-drawn "shock hazard" symbol.... it's probably not too far from telling the truth what would happen to you eventually as you keep using this POS PSU.

The eagle-eyed may notice that cap+resistor job sort-of randomly thrown in there - that was a hack of mine to get it to stay on It would no longer stay on by itself (would make a terrible screeching sound and shut off after a few seconds), so after some experimenting on the bench, I convinced it to run with that cap+resistor combo between the 5V rail and GND...I have no idea what the heck happened there - it was years and years ago when I was just getting started with electronics and would do stuff like THAT I think it's losing its regulation and one of the rails is overshooting, so it's tripping the protection.

Probably failed caps on the output (likely the 5V rail.) Cap helped to keep ripple just low enough to run and resistor just enough load to not cross-load (especially if you were using the 12V rail to power the LEDs and didn't have anything on the 5V rail.)

From what I know, in these cheap generic supplies only the 5v rail is actively regulated (only one photocoupler), with the 12v rail relying only on the 5v rail's regulation to stay in check...

Not quite.
Both the 5V and 12V rails are used for regulation, but the 5V rail in most crap units typically has a bit more "leverage" on how the PSU regulates. 3.3V is always independently regulated (or almost always - I do have one really really crappy PSU example where it isn't.... searching forums.... there, this POS!)

Regulation aside, all rails are typically monitored for under-voltage and over-voltage protection... though on some cheapo PSUs, the 12V rail OV isn't or rarely works. And on many crappy half-bridge PSUs, the UV also doesn't work with low line input. Had one where the 12V rail was happily outputting 5-7V when the input was 70~90 VAC, and it wouldn't shut down.

After all these years though, I'm still not certain about one thing: what is the third transformer for ?

Driving the BJTs, as Dan81 said.

If you notice, the PWM IC is on the secondary side. So the only way it can drive the BJTs is through that transformer. Optocouplers won't work, because BJTs are current-driven devices.

Some more clever half-bridge designs also used this middle "driver" transformer for over-power protection. Others (very rarely) had a 4th transformer specifically dedicated for that function, and it probably worked a lot better (I have an ePower 350W PSU like that.)

[Dan81]

Re: the gutless, bloated, and fried power supply hall of shame

That third one is for driving the BJTs.

Also, that might be a candidate for the most funny mod ever. I can't believe someone went to great lengths to make a Deer modular

[Dannyx]

Re: the gutless, bloated, and fried power supply hall of shame

Here's one I got I had this 400w "Premier" in my PC when I was just a snot-nosed kid years ago. It came stock with the case of course. When I upgraded, I kept it around to power different stuff. I botched the heck out of it over the years, adding XLR jacks and removing the 20P connector to make it "modular". Look at that "custom" case work - state of the art, isn't it ? (I wrote my full name on the side there for some reason, so I had to censor that out...used to do that a lot when I was a kid - personal possession complex or something...I dunno ) It ended up running a PC in my closet for a while (hence the case mod...can't remember what the F** I did to it), then it died a couple of times, I revived it, botched it some more and ended up powering some RGB LED strips I have in my balconies. That thing went through hell and back and would refuse to die for good. It kept coming back one way or another

The eagle-eyed may notice that cap+resistor job sort-of randomly thrown in there - that was a hack of mine to get it to stay on It would no longer stay on by itself (would make a terrible screeching sound and shut off after a few seconds), so after some experimenting on the bench, I convinced it to run with that cap+resistor combo between the 5V rail and GND...I have no idea what the heck happened there - it was years and years ago when I was just getting started with electronics and would do stuff like THAT I think it's losing its regulation and one of the rails is overshooting, so it's tripping the protection. From what I know, in these cheap generic supplies only the 5v rail is actively regulated (only one photocoupler), with the 12v rail relying only on the 5v rail's regulation to stay in check as well and the 3.3v rail being mag-amp generated from the 5v rail.

After all these years though, I'm still not certain about one thing: what is the third transformer for ? One is for 5/12v, the other for 5VSB, so what about the third one ? It was only until rather recently that I learned about mag-amps how the 3.3v output is generated, because until now I always thought the 3.3v rail (or one of the other rails) has its own dedicated transformer, which would add up: one for 12/5v, one for 5VSB and the other for 3.3v or whatever...turns out it's not so, so what is it for ? Feedback ?

[PeteS in CA]

Re: the gutless, bloated, and fried power supply hall of shame

Literally, that PSU is a junk xD.
You are right, 330 microfarads do not give a stable line to the switch. coincidentally, a few days ago I published some photos of an Emac manufactured by Zippy, it has 2 capacitors of 680 microfarad 200v of the Rubycon brand, and that PSU really delivers 250W, what a difference from a manufacturer who knows how to manufacture PSU, compared to a manufacturer that makes garbage.
https://www.badcaps.net/forum/showpo...postcount=3043

That source will have come out of some stock stored years ago, as you say. it seems incredible that sources with designs as old as that are sold.

470uF would probably be OK for 250W, for ripple and for hold-up time, but better than necessary is a good thing.

[kevin!]

Re: the gutless, bloated, and fried power supply hall of shame

What a gut-free POS! The best thing that could be said for it is that they didn't substitute GP ceramic caps for Y-caps. They didn't install anything for Y-caps. And while the MJE13007 clones and the size of the output transformer preclude drawing 500W from the thing, those 330uF input caps pretty well limit it to somewhere in the 150W-200W range (unless you like lots of 100Hz ripple on the outputs before the PSU launches). And that pathetic POS was made in early 2013! It looks like someone recycled a warehouse-full of dusty 200W PSUs, blew off the dust, and slapped 500W labels on them.

Literally, that PSU is a junk xD.
You are right, 330 microfarads do not give a stable line to the switch. coincidentally, a few days ago I published some photos of an Emac manufactured by Zippy, it has 2 capacitors of 680 microfarad 200v of the Rubycon brand, and that PSU really delivers 250W, what a difference from a manufacturer who knows how to manufacture PSU, compared to a manufacturer that makes garbage.
https://www.badcaps.net/forum/showpo...postcount=3043

That source will have come out of some stock stored years ago, as you say. it seems incredible that sources with designs as old as that are sold.

[PeteS in CA]

Re: the gutless, bloated, and fried power supply hall of shame

What a gut-free POS! The best thing that could be said for it is that they didn't substitute GP ceramic caps for Y-caps. They didn't install anything for Y-caps. And while the MJE13007 clones and the size of the output transformer preclude drawing 500W from the thing, those 330uF input caps pretty well limit it to somewhere in the 150W-200W range (unless you like lots of 100Hz ripple on the outputs before the PSU launches). And that pathetic POS was made in early 2013! It looks like someone recycled a warehouse-full of dusty 200W PSUs, blew off the dust, and slapped 500W labels on them.

[kevin!]

Re: the gutless, bloated, and fried power supply hall of shame

That thing's OEM'd by Spire. I gutted a 420W Spire that supposedly had PFC (someone took it) and it was the same gutless crap. Smashed the old board and just installed a very beefy Key Mouse unit, that I'd put around 400W of power (ERL39 transformer and it has nice internals to be honest) and replaced the fan as it was broken.

Thank you, I didn't know that the thing was manufactured by Spire.
I gave the PSU back to the client, since when I change a piece I deliver it to the computer, normally the customers reject the broken parts and give them to me, but this time this client took it.
A pity, because I could have reused the PSU case.
I have a 420w Astec, and one day I will adapt it to a damaged PSU casing to install the Astec.
Bye

[Dan81]

Re: the gutless, bloated, and fried power supply hall of shame

That thing's OEM'd by Spire. I gutted a 420W Spire that supposedly had PFC (someone took it) and it was the same gutless crap. Smashed the old board and just installed a very beefy Key Mouse unit, that I'd put around 400W of power (ERL39 transformer and it has nice internals to be honest) and replaced the fan as it was broken.

[kevin!]

Re: the gutless, bloated, and fried power supply hall of shame

I bring photos of this terrible PSU "500 watts", and as an extra the fan they tried to fix.
Although the fault was not caused by bad capacitors, but because the 230v switch was moved to 115v, the last capacitor exploded (the switch bridges the second capacitor in the series).
The PSU is terrible, thin heatsinks, very small filter winding, small transformer, few capacitors to filter the secondary line, no Y-capacitors, no filters for PFC xDD.
Most terrible, the wiring voltage switch does not carry any insulating protection, and for more, the cable ends are very close to the psu chassis.
And the SCHOTTKY diodes are rated 20A according to their datasheet, and there are only 2, so that those 500w are a lie xD.
What seems curious to me is that these kingcon capacitors on the secondary rail have the same ventilation design as suncon xD.
When the client brought me the computer to fix it, he told me that the heatsink fan was not working, and when I started to diagnose it, I found the heatsink of the cpu loose, the fan cables cut off, I removed processor from the motherboard to proceed to turn on the computer without burning the CPU, I put the cables in their fan place, plug the fan and it does not turn (what the owner burned when putting the cables wrong) (the computer turned on but did not give video and the dvd reader and the HDD made strange noises) and I also hear sparks inside the PSU due to an electric leak in the exploded capacitor, I change the PSU for a new one, put the CPU in its place, put a new heatsink and it worked again.
The source was creating a terrible ripple in the electric rails, preventing the computer from turning on, and since the owner of the computer did not know that this was the cause, he tried to take out the cpu, in the process he broke the fan.

This PSU is the second of my personal list of the most dangerous PSUs that I found (the first is an AT PSU that has huge holes that contact the primary).
That's why always, you must buy quality PSU and recognized brands.

[Dan81]

Re: the gutless, bloated, and fried power supply hall of shame

WTF? That PSU just doesn't make any sense. Barely one capacitor per rail, and yet they bothered with line filtering? Why?

Come to think of it, this stupid design was introduced by them as early as 2005. I have 2 more units with the same PCB and same silkscreens.

And if you haven't noticed, the PI coils for 3.3, GND and 5v are bypassed altogether on the underside Ironically, the only one that actually works as intended is the 12v one.

Interestingly, the same design was actually fixed in a Allied PSU I have from the same era - same PCB design as above but they finally bothered to fully make the PI coils work, along with adding all caps. (I had to recap it, obviously, because they were your usual Yang Chun crapcaps)

[TechGeek]

Re: the gutless, bloated, and fried power supply hall of shame

Ooo, a PSU with a built in heater! /sarcasm

[PeteS in CA]

Re: the gutless, bloated, and fried power supply hall of shame

What a pathetic POS (Oh, Deer?). It'd probably do 200W-250W, crappily, but it has GP ceramic Y caps. Unsafe at any power. The label rating says 400W, but even the output harness wires aren't good for that. They'd get really warm ... unless something blew up too quickly.

[Agent24]

Re: the gutless, bloated, and fried power supply hall of shame

WTF? That PSU just doesn't make any sense. Barely one capacitor per rail, and yet they bothered with line filtering? Why?

[Dan81]

Re: the gutless, bloated, and fried power supply hall of shame

Check this hot garbage.

"Torrent Computers" LC-8400BTX, "400W". Erm...for how many picoseconds?

Then... 2 12v rails? Yeah, right. I'd want to see them on the PCB.

Cracking it open reveals a sad joke of a PSU. The AC receptacle looks interesting though, in the way it's mounted. Also, the casing is now given a much more updated look from the boring lines present on other Deers! Dunno why they cheaped on two screws, but that's L&C for you.

Also, fake PI coil spots! The only one that ACTUALLY connects something is for the 12v rail. Guess this would be a worthy opponent for the mighty Google "RAM eater" Chrome?

I'll probably just save the casing, fan, heatsinks and sillicon, then just junk the rest. Maybe save some of the wiring, since it seems it came with a SATA connector.

[PeteS in CA]

Re: the gutless, bloated, and fried power supply hall of shame

ae7oo, if the TL494 clone PWM is on the secondary side then no feedback optocoupler or feedback transformer is needed. The pulse width modulation is done by the PWM IC, which drives the switch transistors.

For having been built in the second half of 2000, that 1A 5VSB is pretty wimpy. The 2SC3866 switch transistor is probably for the 5VSB, while the 2SC4106 switch transistors are the main inverter, probably half bridge. Everything about it screams noisy gutless wonder. Trying to draw 250W continuous from it probably would have resulted in a dramatic event in a matter of minutes or less.