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

[TechGeek]

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

Transformer core might saturate?????

[momaka]

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

Transformer core might saturate?????

https://en.wikipedia.org/wiki/Saturation_(magnetic)

Basically, each time one of the transistors on the primary switches ON and OFF, you get a pulse of current. This pulse of current travels through windings on the primary side of the transformer and creates a magnetic field inside the metal "walls" of the transformer (i.e. the core). The windings on the secondary side "pick up" the change in magnetic field in the core and produce a current pulse on the secondary side of the PSU.

Now, as the PSU is loaded more, the PWM IC keeps the voltage on the output the same by sending longer PWM pulses to the transistors on the primary side. This makes the transistors stay in the ON state for a longer period of time, which means the current pulses going through the primary winding on the transformer also become longer and the magnetic field in the transformer core increases proportionately... but only up to a point. Once it reaches saturation, the impedance on the primary side can decrease greatly. And when that happens, the primary transistors will be trying to drive a short circuit pretty much. The Mr. Kaboom follows

Anyways, for your typical half-bridge SMPS with a 33-size EE/EI transformer cores, you're looking at about 300 Watts maximum DC output before things start to get "out of control" with the transformer core saturating.

[PeteS in CA]

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

As long as a core's magnetic properties are active, the transformer primary acts as an inductor rather than a simple wire (= short circuit). The core loses its magnetic properties under two conditions: the magnetic field is too great (saturation); the core gets too hot (~200C for power transformer ferrites, the Curie temperature).

When either happens the primary becomes a short circuit (maybe a couple of tenths of an ohm) and the transistors or MOSFETs swiftly become molten silicon, with the likelihood of a bright flash, a bad odor, smoke, and plastic shrapnel.

[momaka]

GenMax 400 Watt ATX Switching Power Supply

Are low quality power supplies really gone for good now? Or is this thread just being forgotten?

Well, I think we can safely assume the latter, so here goes another low quality power supply: the GenMax [TM] ATX Switching Power Supply.


This gem from 2007 boasts 400 Watts on its label and apparently can also come in 350-Watt, 450-Watt, and 500-Watt variants. Sounds very promising! Let’s see the label.


Oh yes! We get 22 Amps for the 3.3V rail, 28 Amps for the 5V rail, and 14 Amps for the 12V rail. No need to worry ourselves with combined power ratings, though. I’m sure this PSU is solid (it really is! Not leaking or vaporizing… yet. ) And look! – label states this PSU is ATX v2.03 compliant too. Great!!!
…
Okay, okay. Leaving the silly sarcasm aside, let’s go down to the guts. As always, here is a top view. Viewer discretion is advised!


Not sure where to start here as there isn’t much to talk about. This is just another low quality PSU based on the much dated half-bridge design. PWM controller is a 7500B chip and my guess would be most protections (if any) are handled by the 339 quad comparator IC. The main transformer looks like the typical EE/EI 33 size found in many low-end PSUs. Likewise, the 5VSB and BJT drive transformers are also quite small. Both the input and output wires are 20 AWG, with the small difference that the input wires are rated for 600 Volts and have a UL number. As for output connectors: all this PSU has is a 20+4 pin motherboard connector, a 4-pin CPU power connector, 4 peripheral plugs, 1 floppy, and a single SATA power connector. That is just enough to make this PSU usable (i.e. dangerous for your hardware). Those of you with sharp eyes might also have spotted the funny grounding scheme of this PSU, with one of the PCB screws also holding down the grounding wire. And if things weren’t cheesy enough for you (or corny if you like, or… well, not beefy for sure), then there is always the pathetic heatsinks to look at. Here is a better shot of that:


In the above shot we also see the text “WW-86250 REV:2309-A1 PCB” printed on the PCB. Can anyone tell who this PSU is made by? Because I have no clue.
(Side thought: I wonder if the manufacturer really made 2309 revisions of this PSU. Makes sense, though. Otherwise, how else would they get the PSU so gutless and still “working” - probably some major trial-and-error on their part. )

Speaking of gutless, look at the secondary side filtering (or rather, the lack of it):

We get the awesome “one cap per rail” treatment, no PI coils, and… this is my favorite : dummy load resistors in spots marked for additional output filter caps!

The solder side doesn’t look that good either.
https://www.badcaps.net/forum/attach...1&d=1475344632
As you can see, there are quite a few sloppy solder joints. Surprisingly, I do not recall seeing any loose solder balls around… unlike that crappy Raidmax RX-380K (Sun Pro KY-480ATX) I showed a while back, which looked like someone dropped the PCB in a bucket of hot solder from a height of a 10-story building.

Last but not least, the fan:

No idea who the manufacturer is, but at least the label is honest enough to say that it is a sleeve bearing fan. Rated for 0.22 Amps at 12V.

Anyways here is a more detailed look at the components used in this PSU, starting with the primary side and then moving onto the secondary.

Primary Side…
* 250V glass fuse (can’t see rating)
* NTC thermistor (MF72 5D9)
* no input filtering of any kind other than two 1000 pF blue ceramic caps rated for 2 kV (not Y2-class safety approved)
* four RL207 diodes (rated for 2 Amps) for the bridge rectifier
* two primary caps in series: KYS “ROHS VENT”, 200V, 330 uF, 18 mm dia x 36 mm height (measured capacity showed 166 uF and 170 uF on the ESR-micro v4)
* two generic 13007 NPN BJT transistors (by SXW?) for the h-bridge main supply
* one 13003 NPN BJT for the 2-transistor 5VSB circuit
* two KYS “ROHS VENT”, 50V, 10 uF, 5x11 mm (dia x h) - BJT drive caps
* one KYS “ROHS VENT”, 50V, 10 uF, 5x11 mm (dia x h) - “critical” cap for 5VSB

And another picture of the primary side from a different angle:
https://www.badcaps.net/forum/attach...1&d=1475344632

Secondary Side…
3.3 V rail (mag-amp regulated):
* SBL1640CT schottky rectifier (16 A, 40 V, TO-220AB)
* one KYS “CD288”, 10 V, 1000 uF, 10x16 mm cap, no PI coil
* 15-Ohm, 3 Watt load resistor (dissipates 0.726 Watts @ 3.3V)

5 V rail:
* SBL2040CT schottky rectifier (20 A, 40 V, TO-220AB)
* one KYS “CD288”, 10 V, 1000 uF, 10x16 mm cap, no PI coil
* one empty cap spot, 8 mm dia
* 15-Ohm, 3 Watt load resistor (dissipates 1.666 Watts (!) @ 5V)

12 V rail:
* PEC UR1002FCT fast recovery rectifier (equivalent to UF1002FCT…10 A, 100 V, TO-220 isolated)
* one KYS “CD288”, 16 V, 1000 uF, 10x20 mm cap, no PI coil
* one empty cap spot, 8 mm dia
* 270-Ohm, 1 Watt load resistor (dissipates 0.533 Watts @ 12V)

-12 V rail:
* two FR104 fast recovery diodes (1 A, 400 V)
* one KYS “CD288”, 16 V, 330 uF, 8x14 mm cap, no PI coil
* 150-Ohm, 2 Watt load resistor (dissipates 0.96 Watts (!) @ 12V)

-5 V rail:
* two FR104 fast recovery diodes (1 A, 400 V)
* one KYS “CD288”, 10 V, 330 uF, 8x14 mm cap, no PI coil
* 51-Ohm, 2 Watt load resistor (dissipates ~0.5 Watts @ 5V)

5 VSB rail:
* FR302 fast recovery diode (3 A, 100 V)
* one KYS “CD288”, 10 V, 1000 uF, 10x16 mm cap, no PI coil
* one KYS “ROHS VENT”, 10 V, 220 uF, 6.3x11 mm cap, no PI coil
* 51-Ohm, 2 Watt load resistor (dissipates ~0.5 Watts @ 5V)

And here is a shot of the rectifiers bolted onto the secondary heatsink:
https://www.badcaps.net/forum/attach...1&d=1475344632

So all of that above is supposed to be capable of 400 Watts . Well folks, not today!
(Makes one wonder if and/or how much more skimpy the 350 Watt version is.)

You reached the end of my post – congrats!

[Wester547]

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

Wow, what a pile of trash. I give it no more than 200W, but that's without the ripple voltage or voltage regulation in spec. Thanks for the thorough breakdown. I can't seem to find anything about UR1002CT, though. The datasheet I can find for UF1002CT describes it to be a 10A ultrafast and not 16A. The datecode I can see on the label says March 2006 - "Month 3" "2006".

EDIT: Okay, I figured there was another year on the end of the label not visible in the picture. March 2007 then.

[momaka]

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

I can't seem to find anything about UR1002CT, though. The datasheet I can find for UF1002CT describes it to be a 10A ultrafast and not 16A.

Yes, it's a 10A part, sorry. Fixed that typo in time this time . Thanks!
(Copy-pasta mistake, as I reused another post to make the body of this post.)

The datecode I can see on the label says March 2006 - "Month 3" "2006".

There's also a box with "2007" text on the label (supposedly right next to the "2006" text), but it got torn off. The 2007 text was crossed off with a pen, so that suggests this PSU was made in 2007.

Thanks for the thorough breakdown.

You're welcome!

[ChaosLegionnaire]

Re: GenMax 400 Watt ATX Switching Power Supply

Makes one wonder if and/or how much more skimpy the 350 Watt version is.

lol skimpy? this 400w version is already skimpy. the 350w version would probably be nude, the 450w version barely clothed and the 500w version still considered inappropriate for certain conservative countries!

so i am going to reiterate my post that was in another topic:

errr why would u be calling this psu a cutie? it aint exactly a chick! even if it was, her body is cheap, her clothes (labels) are cheap, her insides and her head are all empty and her fans? well... she has none to speak of. i'd put a paper bag over her head anyday...

[forza764]

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

Question: Can a bigger 12v/5v output inductor lower the ripple in cheap psus?

[momaka]

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

Question: Can a bigger 12v/5v output inductor lower the ripple in cheap psus?

Possibly... if it has more turns for each rail (i.e. higher inductance).
A bigger toroid inductor should definitely allow for higher power output, though. Or at least somewhat higher (the wires of the big inductor also need to be thicker, otherwise it may be just as capable as a smaller ones).

lol skimpy? this 400w version is already skimpy. the 350w version would probably be nude, the 450w version barely clothed and the 500w version still considered inappropriate for certain conservative countries!

The content of this website is not available in your country.
(like those stupid YouTube blockers.)

[TechGeek]

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

This post contains content from Sony. The post content is unavailable in your country.

[forza764]

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

Possibly... if it has more turns for each rail (i.e. higher inductance).
A bigger toroid inductor should definitely allow for higher power output, though. Or at least somewhat higher (the wires of the big inductor also need to be thicker, otherwise it may be just as capable as a smaller ones).

Thanks for the answer, i have another question, there are some cheap psus with no places for pi coils, can more capacitance compensate for that, or the ripple will be through the roof regardless? for example 2x2200uF 16V instead of 1000uF

[PeteS in CA]

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

Concur with W547, 200W max. And with 330uF I/P caps the hold-up time will be marginal or worse. Two things I did not see are a 7912 for the -12V and a 7905 for the -5V. Are those O/Ps semi-regulated? Or did I some how miss the 3Ts?

Those "Y-caps" look like generic ceramic disc caps, maybe rated for 2KV, not safety agency approved parts. X-caps? We don't got no steenking X-caps!

Nothing about that PSU is adequate for 400W, from the 330uF I/P caps through (i.e. including) the O/P wires. I'm not sure 4 AWG #20 wires are rated to carry 22A (and even if they are, those wires are going to get very very warm).

As a 200W PSU it'd be an unsafe POS. Try drawing 400W, and , loud noise, bright flash, and unpleasant smell.

[TechGeek]

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

And a beeping smoke detector.

[momaka]

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

Thanks for the answer, i have another question, there are some cheap psus with no places for pi coils, can more capacitance compensate for that, or the ripple will be through the roof regardless? for example 2x2200uF 16V instead of 1000uF

Yes, increasing the capacitance will definitely help lower the ripple. But as far as lowering it enough to keep it within ATX specifications - that, I don't know. Only way to find out the answer to such a question is to load the PSU with a load tester and measure the ripple/noise on the outputs with an oscilloscope.

As a 200W PSU it'd be an unsafe POS. Try drawing 400W, and , loud noise, bright flash, and unpleasant smell.

No worries. I don't use it even for powering small electronics. Have a better gutless PSU for that, which I actually might upgrade some day, since it's based on the CWT ISO platform and has stable voltage regulation.

Two things I did not see are a 7912 for the -12V and a 7905 for the -5V. Are those O/Ps semi-regulated? Or did I some how miss the 3Ts?

No regulation for -12V and -5V rails. Just 2x FR104 diodes > toroid inductor > filter cap > ATX wires. Oh, and big-ass loading resistors. I suppose that's your regulation right there!

Those "Y-caps" look like generic ceramic disc caps, maybe rated for 2KV, not safety agency approved parts.

Yup. 1000 pF, 2 kV. No writing on them, so we can safely assume no Y2-rating.

And a beeping smoke detector.

I hate that noise!

If I know I will be testing something that may blow up or produce smoke, I always take down the smoke detector in my room beforehand.

[TechGeek]

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

...and wait for an electrical fire to start when you least need it to happen

[momaka]

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

...and wait for an electrical fire to start when you least need it to happen

I put the smoke detector back when I am done testing, so no.

[PeteS in CA]

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

No regulation for -12V and -5V rails. Just 2x FR104 diodes > toroid inductor > filter cap > ATX wires. Oh, and big-ass loading resistors. I suppose that's your regulation right there!

Yep. Without those pre-loads at light or no load the output caps would peak charge and the O/P voltages would go out of regulation high. And if the snubbers weren't carefully designed (Design? We don't need no steeenking design!) there would probably be considerable ringing in the voltage out of the rectifiers, meaning the peak-charged voltage could be really high.

[momaka]

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

And if the snubbers weren't carefully designed (Design? We don't need no steeenking design!) there would probably be considerable ringing in the voltage out of the rectifiers, meaning the peak-charged voltage could be really high.

I've seen something like that happen to a Deer (and surprisingly, a well-built Deer PSU), where the snubber ceramic cap on the 12V rail shorted and blew the resistor that was in series with it. The PSU appeared to work, but would occasionally shut down. Owner said his system was very unstable. I suppose this "ringing" is probably what caused the PC's instability.

[PeteS in CA]

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

I've seen something like that happen to a Deer (and surprisingly, a well-built Deer PSU), where the snubber ceramic cap on the 12V rail shorted and blew the resistor that was in series with it. The PSU appeared to work, but would occasionally shut down. Owner said his system was very unstable. I suppose this "ringing" is probably what caused the PC's instability.

Snubbers limit voltage stresses on rectifiers or switch devices. They also limit high frequency ringing, which can be unpredictably disruptive.

While not snubber networks, those little ferrite beads you sometimes see on MOSFET and rectifier leads also limit ringing by slowing down the turn on and turn of of the devices. This is usually done to reduce noise conducted from the PSU back into the AC line.

[momaka]

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

While not snubber networks, those little ferrite beads you sometimes see on MOSFET and rectifier leads also limit ringing by slowing down the turn on and turn of of the devices. This is usually done to reduce noise conducted from the PSU back into the AC line.

I've seen those on a few old Bestec and Lite-On PSUs. The funny part is the beads were held with that conductive tan/brown glue . EMI/RFI noise on the line surely wouldn't have been a problem at all when that glue went conductive and shorted out the leads (after all, a dead PSU produces no EMI/RFI. ) Looks like I caught many of them on time, so I was able to remove the glue before it got bad. Left the beads alone to just dangle around on the legs.
From what I measured, they weren't conducting between the leads of the parts, even when they were touching.