Death of a Gutless Wonder III: The Labors of Hercules

REVIEW INFORMATION
SUPPLIED BY: Hercules
MANUFACTURER: Hercules
PRODUCT: HRC512F 500W
PROD LINK: N/A
PRICE: $19.95 @ eBay
Price is at the time of testing!

Don’t… even start. I don’t want to hear a single one of you whistling that Wham song. I’m not going to name it, because then I’m giving it power. It’s that one that always popped into my head as a kid while I was bowling and cost me big points on the score because I hated it so much. Seriously, I may hate Neon Trees songs now, but those pale in comparison to THAT song. I loathe that song.

Anyway, this is the fan.

Oh, dear Lord. You guys know what the term “cost down” means, right? Well, in electronics, they use that term whenever they want to come up with a way to make something as cheaply as possible. You come up with a nice, decent power supply design first. Then, you start removing and/or replacing components with cheaper stuff as needed until the thing stops working. Then, you undo the last thing you did and sell it to the public.

From what I can see here already, this unit has been cost-downed almost to the point I wouldn’t even plug it in if I wasn’t reviewing it. Let’s get more into it, and see what’s been compromised for the sake of making it cheap.

Honestly… units like this scare me.

First, we start off at a very cheap looking AC receptacle. Usually, you’d find some AC line filtering here, but not this time. Cost down. Don’t need any filtering for a unit to work, you just need it for any radio in your house to work. Who cares? Not the OEM of this unit.

Insulators under the PCB? Cost down. Don’t need them either. Ground wire to the case? Naw, we’ll just run the wire to the PCB and ground the case to it second hand via the standoff holes. Don’t need proper grounding when you’re not looking for UL approval.

To my surprise, the soldering on this unit is actually not too bad. It’s nothing I’m going to call better than very good, but it’s not terrible. They did something right, at least.

Only one of the four standoff holes provides any proper grounding for the housing. And it’s not the hole closest to the actual AC receptacle ground, either. Safety? Bah, who needs it. We’re makin’ money, here.

I’m surprised they didn’t cost down that standoff right out of the unit, honestly. This is the cheapest unit I’ve ever seen.

And hey… see that discolored spot there? That’s not a solder quality blip, folks. That’s heat damage. Something got toasty enough on the other side of the board to do that. More on that in a minute.

Oh, man. Look at that gigantic discolored spot on the board where the secondary heatsink was. I’m upping this unit from gutless wonder to fire hazard.

Seriously. Think about it. All it took was my load tests to do that. Figure half an hour on the cold tests, half an hour on the hot ones. One hour. This piece of crap burned the hell out of the PCB in SIXTY MINUTES. Really, what do you think this unit would do, if you put it in your main rig and ran it at 200 watts for months? True, this unit did have working overpower protection, but think about that, too, and how far that protection allowed this unit to run. You have this pile of junk bathing your components in out of spec ripple for weeks on end, not shutting down. Maybe some of the voltages are out of spec too, and your rig is a little unstable. During this time, that PCB is getting burned, and burned, and burned. Suddenly, one day… one fateful day, something gets a little too hot. Maybe the insulation on the coil windings. Maybe a diode. It catches fire. The fan’s right there, hovering above whatever’s burning. The fan catches fire, too. Where does it stop?

I recently saw a story posted on the Register via our very own forums about a Jersey CP4-420WS that had very dramatically flamed out. It burned so hot that it melted the cheap housing it was built into. Look at the above picture again… not saying this is the same OEM as that Jersey, but do you see how that might be possible? Forget the ripple. Forget the performance. Forget the price. You don’t want this anywhere in your house at any price. Your safety has to be more important than saving a buck on a unit like this.

No doubt about it, this unit has been cost-downed to the point it is a safety hazard. Whether you arrive at this conclusion based on the terrible grounding or the big discolored blotch on the PCB, it is simply THE worst power supply I have ever tested to date. I’m glad it’s dead.

Let’s carry on.

We’re cost-downing here, so there’s no point using decent capacitors, is there? And look, they’re even undersized for the board silkscreens.

And I find it hilarious that this thing actually has proper bleeder resistors to discharge those caps. That’s a safety thing. Safety costs money. What are we, millionaires?

Looking at this picture, we can clearly see that the source of the major heat damage we saw underneath is that diode, there, D21. I de-soldered and pulled it off the board… the PCB is totally black there. Surprisingly, the diode measured good.

Checking out at all this heat damage, you might be surprised that our exhaust temps were so low. There’s one simple explanation for it, though. The unit got no cooling where it mattered. The secondary is in a hot spot, where there was no airflow. Another reason it’s a safety hazard… they just stuck a big fan in there thinking the unit would automatically keep cooler than an 80mm at the exhaust. Wrongo. No airflow to the right parts, parts overheat, parts catch fire, fire spreads to the fan, fan seizes, fire spreads further, case melts… see how this goes?

Oh, and let’s not forget the suckeriffic H.Q. capacitors here. Pi filters? Who needs those, they might have brought the ripple under control. Let’s continue our cost down march into oblivion and sell it to people who otherwise don’t know what they’re getting into, see the price, and think, “Oh, great! I’m a little too tapped out to finish my build with the power supply I want, but I’m saved! I’ll just grab this one for a bit until I can afford better. It’s cheap, but maybe it’ll do temporarily…”

Cost down, cost down, cost down the 5VSB rail too. Let’s use junk capacitors and go with an old two transistor design. Let’s face it though… with this unit, taking the risk of having the 5VSB rail hit 12 volts and fry the mainboard because of this crappy design are the very least of your potential problems. Uh… fire hazard, anyone?

That chip you see there is a Weltrend WT7520. I wouldn’t be surprised if it’s the most expensive part in this wonder of wonders. It provides PWM control and very basic protection. How basic? Over voltage and under voltage only. If the error condition can’t be detected by those functions, it’s not going to shut down. There is an auxiliary input that runs to the primary side to provide overpower protection, but there’s nothing at all for the standby rail. It’s free to overvolt itself to the moon in this unit if it wants to.

Now, I’m getting angry. Really angry. “For continued protection.” Like you have much of that to begin with. Like that fuse is going to stop this unit from lighting up like a cheap candle due to profound cost-downitis. Excuse me while I squish something.

Ow. I squished the cheap housing, bent it all to hell, and got a paper cut. A paper cut. From metal. Ever get one of those? Not pleasant.

Speaking of the housing, three of the four screws that hold the cover on were stripped out. Not the screw threads, the tapped holes in the housing. Remember, the housing is only grounded by a standoff hole on the PCB. That screw strips, the board lifts away from the standoff, and suddenly the case is NO LONGER GROUNDED.

Oh, look – a halfhearted stab at a line filter: two Y capacitors. That’s all. And four general-purpose diodes instead of a dedicated heatsinkable “might actually be able to handle the power” bridge rectifier. That fuse is good, of course. The unit died to protect it. Facepalm.

These are the secondary side parts. From left to right, we have 12V, 5V, and 3.3V. Not one of these measured bad.

In fact, I cannot find the problem with this unit. Something failed, but not definitively enough to test bad. Everything I put my meter on tested good. My money’s on that burned diode, not that I’m wasting another second of my time on this unit to replace it.

The primary side. All parts good. Let’s score this embarrassment to power supply engineering now, shall we?