Failed power brick to power hard drive

[edy]

Hi folks,

Another day, another failed electronic. This time it is a cheap $8 hard drive USB connection kit which lets you power up any IDE and SAT drive (HD, optical, floppy) and connect to USB to access.

The kit comes with a separate power brick (like the one shown here) which outputs to 4-wire power connector like you see from CPU supplies... and also a SATA adapter if you have a SAT drive instead. There is a separate data cable which has USB on one end and the other end has 3 different options to connect to IDE or SATA or other IDE-like connection with fewer pins (floppy?).

Have a look at the power supply. Nothing but bare bones and no regulatory/safety markings of any kind. It takes 120V mains and outputs 5v and 12v to power the 4-pin standard CPU power connector for IDE drives (or SATA with a conversion cable). The LED used to power on when it was plugged in the wall. No more.

Smell test seems to point to the side near the mains. I don't think the fuse shorted, and I can test the diodes and cap no problem (haven't got to it yet). But yet somehow I suspect that power MOSFET with the crappy heat sink and corroded legs?

I had the think plugged into a drive overnight for a backup and even the whole day, I unplugged it from the drive when backup was done, and when I went to plug it into another drive it was dead. I wonder if it got blown then through some short failure. I'll be to play around but I may not want to bother... I have some CPU supplies in the garage which can be used instead but this was more convenient (but piss-poor design).

Thoughts anyone?

[edy]

Notice no grounding at all yet they decided to use a 3-prong plug? And the side that powers the drive also missing black wire. Huh?

[BradC]

Notice no grounding at all yet they decided to use a 3-prong plug? And the side that powers the drive also missing black wire. Huh?

Really common. That way they can use the standard IEC cables that seem to breed around anything PC related. It's in an entirely plastic enclosure and probably qualifies as double insulated.
Even the PSU that came with my macbook has a 3 pin plug on one end and a 2 pin socket on the other.

The yellow wire is gnd, white is +5v and red is +12v. Nobody says they have to conform to a specific colour code. The unit is sealed after all.

[codeboy2k]

As said, you don't need the ground, it's all plastic. They did well with the AC side, there's lots of clearance and creepage distances maintained.  There's no filtering, no spark gaps, and no MOV protection, which might explain why diode D1 looks like it exploded and blew the end off. This is part of the bridge rectifier.

[BradC]

As said, you don't need the ground, it's all plastic. They did well with the AC side, there's lots of clearance and creepage distances maintained.  There's no filtering, no spark gaps, and no MOV protection, which might explain why diode D1 looks like it exploded and blew the end off. This is part of the bridge rectifier.

Which D1 are you looking at? The only physical damage I can see is a bit of cracked insulation on the power resistor. The 4 bridge diodes look in-tact to me.

[codeboy2k]

As said, you don't need the ground, it's all plastic. They did well with the AC side, there's lots of clearance and creepage distances maintained.  There's no filtering, no spark gaps, and no MOV protection, which might explain why diode D1 looks like it exploded and blew the end off. This is part of the bridge rectifier.

Which D1 are you looking at? The only physical damage I can see is a bit of cracked insulation on the power resistor. The 4 bridge diodes look in-tact to me.

OK. Maybe it was just the angle and lighting in the photo. D1 of the bridge has one end that looks, rough, cracked, like it was blown apart there. If the diodes check out then it's not a problem.

[eas]

Even the PSU that came with my macbook has a 3 pin plug on one end and a 2 pin socket on the other.

You ever wonder why there is a metal button that fits in the channel on the socket side of the mains cable, rather than, say, plastic?  I always assumed that it was used as a ground.  According to this tear down, it is on iPad adapters, and from online photos of the inside of MacBook adapters, it seems to be connected to something internally in those too.

[BradC]

You ever wonder why there is a metal button that fits in the channel on the socket side of the mains cable, rather than, say, plastic?

I actually never wondered at all. I figured it was metal as it is quite structural and plastic tends to break when handled by ham fisted hacks like me.
 

I always assumed that it was used as a ground.  According to this tear down, it is on iPad adapters, and from online photos of the inside of MacBook adapters, it seems to be connected to something internally in those too.

Well, well, well (three holes in the ground). You are absolutely spot on. The socket has 2 spring loaded segments that contact the button. There we go, it's a poor day when you don't learn something new!

[tom66]

There's nothing "well" about the AC side of this. That glass fuse is unlikely to be able to interrupt fault current properly.  The lack of an NTC also increases the stress on the fuse. And it's a self-oscillating converter with very marginal regulation properties.

[codeboy2k]

There's nothing "well" about the AC side of this. That glass fuse is unlikely to be able to interrupt fault current properly.  The lack of an NTC also increases the stress on the fuse. And it's a self-oscillating converter with very marginal regulation properties.

I only said the PCB was done well, re: creepage and clearance. Then I followed that with no protection devices. 

Also, the self-oscillating converter is not THAT bad.. it is often used in these bricks, because it's cheap and good enough. I can see a TL431 and an opto-isolator, so there's feedback to the oscillator, it probably has at least 5% load regulation, and likely better.  This is good enough, since it's designed to go with a product that includes its own point-of-load voltage regulators anyway.  I don't need gold standard power supplies powering my consumer devices, but I do wish the cheap PSU makers would put overvoltage zeners on the outputs, I absolutely don't want a cheap PSU to KILL my device.

[cosmicray]

The power resistor (2.2 ohm ?) next to C4, is is cracked at one end, or is that a blob of solder ?

The other thing is the tape on the top of C1 (the large cap), is that there to provide additional insulation from the aluminum end of the cap ?

[edy]

The double-sided tape on the cap may be to make it stick to the top of the case. Honestly I have no idea what use it has, unless they expect that cap to blow out and they want to minimize the pop not to scare people. I will need to check the components and report back... Haven't had time yet. I feel that MOSFET let out the magic smoke because it has a crap heat sink and the leads look awful. I don't know what that glass fuse is, there are no markings even but it is easy to check all these components.

One thing I'd like to do is reverse engineer the circuit and do a Dave-CAD type drawing, more to help me learn this awful design and why the safety issues are being violated (I presume to keep costs at a bare minimum, not because of a lack of knowledge).

[SeanB]

Mosfet will be shorted on all 3 leads and the 0R22 resistor will be very open circuit. All the SMD transistors under the board will also be either shorted or blown open, seeing as R5 is definitely blown open. Sad to see the slightly conuctive foam pad defeating the isolation barrier, that foam degrades into something conductive with time and toasting, probably was the initial cause of cooking semiconductors, though high ESR of the cheapolytics did not help either.

[simplifyiyuk]

could be inrush current from your description of the failure. Looks very stripped back and some reused/recycled parts in there for sure. might has well put the regulatory mickey mouse symbol on there 

[rdl]

I would just throw that in the trash and be thankful it didn't start a fire when it went bad.

I have a similar device, except the power supply it came with claims to be UL listed, CE and TUV approved. I think I paid a whole $15 for it, but it was on sale. I would estimate the power supply accounted for more than half of the total cost. If you only paid $8, well...