Empire EMP-680

[370forlife]

Empire is a brand we are all very familiar with....maybe. Well, anyway, it's not a brand you will find in the US, lets put it at that.

First lets get a overall shot of this unit:



It has a attractive exterior, matte black all around. Only the 20+4 connector is sleeved though, and the sleeving isn't that good of quality. It is starting to pull apart in areas and fray near the unit itself. Cables all use 18ga wires. Speaking of which, these are the connectors you will find on it:

1 20+4 pin motherboard
1 4+4 pin CPU
3 4 pin molex
1 4 pin floppy
6 sata power
2 6+2 pin PCI-E

Now a label shot:



30A on the 5v and 32A on the 3.3v are kind of high for a newer unit, and with only 480W for 12v. Definitely looks to be based on a older design.

Speaking of which, lets open 'er up and see what makes it tick:



Hmm. Anyone care to guess who makes this one? Another thing, think it's a real 680W? Bridge rectifier is a 10A, 600V with a heatsink. Transient filter has a huge NTC thermistor, shrinkwrapped fuse, .68uf X cap, coil across live, second coil across both, and two Y caps. It's kind of a unique unit, as it uses two transistor forward without APFC. I believe most members here would consider that a good thing. Primary switchers controlled by a UC3845A. Primary capacitors are two CS 200v, 1200uf. Secondary capacitors are a mix of Asia-X and Nicon. On the secondary, the 12V is produced by two 40A, 100V schottky diodes. 5v by a 30A, 40V schottky, and same for the 3.3V. The really sharp would have noticed something here. The cable that connects the AC outlet to the PCB was plugged in backwards. I have a question, would reversing live and neutral do any damage to the unit?

Here's a list of caps:

12V: 1500uf, 25V Nicon, 2x 1000uf, 16V Asia-X
5V: 2x 2200uf, 10V Asia-X
3.3V: 2x 2200uf, 10V Asia-X
5vsb: 2x 1000uf, 16V Asia-X
-12v: 1000uf, 16V Asia-X



Ooo, everell will like this one. They are using the DM311 to generate the 5vsb.



IGNORE THE BIG BLACK BAR. Ehh hem. Anyway, protections taken care of by a Weltrend WT751002



Strange brand fan. From what I can find, it's either made by CEMEC or Evercool. It's quiet though.

Well, anyone have any idea who makes it?

[shovenose]

Re: Empire EMP-680

whats with the black bar?

whos the oem-out with it LOLOLOLOL!

[Toasty]

Re: Empire EMP-680

Deer?

Where's the BEEF! 680w my azz. Lucky if it could do 1/2 that.

>>I have a question, would reversing live and neutral do any damage to the unit? <<

No. One of my systems is running on a backwards wired socket for many years. The only complaint is from the UPS LED which indicates the site wiring fault.

Toast

[everell]

Re: Empire EMP-680

My vote is Solytech/L&C/Deer/Allied, etc.

Why? This power supply has updated circuit ideas, like the DM311 pwm chip for 5vsb, and a 3845 main pwm chip rather than the "chip of the year", but they are still using old style thru hole resistors and components....which are machine inserted. They have not made the move forward in technology to surface mount components. Most of the better power supplies have surface mount technology. Even Bestec uses surface mount components!

[shovenose]

Re: Empire EMP-680

Shit caps.

[momaka]

Re: Empire EMP-680

Good sized transformer, looks big enough. The DM311 generated 5vsb is definitely a good thing. Too bad the caps are crappy. Looks like a decent PSU otherwise.

As for who made it...
Well, it doesn't look like the typical Deer/Allied/L&C so probably not a Solytech. I want to say Sun Pro, but I doubt they would make something this good, and if they did, I'm sure they would label it as over-9000 Watts. But let's go with Sun Pro, just for fun.

I have a question, would reversing live and neutral do any damage to the unit?

No, no at all. AC is AC, so it doesn't really matter. In some parts of Europe and Asia, the plugs are reversible/bi-directional - that is, Live and Neutral can be swapped, depending on how you plug in the cord. Of course, ground is always ground, and that's all that matters.

[Toasty]

Re: Empire EMP-680

Transformer ERL marking is from the Solytech/L&C/Deer/Allied group, no?

>>The really sharp would have noticed something here. The cable that connects the AC outlet to the PCB was plugged in backwards.<<
So the locking tab was turned around?

Main switchers p/n's are....?

It'll never make 104.8 amps. I doubt it would make 75% of that without smoke. Regulation is probably out the door so far when you get there, it won't make a difference anyway.

What's the rail cross-loading specs say?

Toast

[370forlife]

Re: Empire EMP-680

Transformer ERL marking is from the Solytech/L&C/Deer/Allied group, no?

So the locking tab was turned around?

Yeah, it was turned around. But I think the ERL markings are just a standard thing a lot of transformers use.

Main switchers p/n's are....?

Two IXTW26N50P's, can do 26A at 25C, 17A at 100C, 78A in pulse mode.

I'll give everyone a big hint. The blacked out part number is KK-9966D.

[POM_MJ]

Re: Empire EMP-680

Two transistors forward topology with low RDS(on) (0.260ohm), this unit must have high efficiency around 80% I thought.

[Th3_uN1Qu3]

Re: Empire EMP-680

Any chance of a side shot of the secondary? Maybe not 680W, but it looks like it'd do 500W. I'm just curious about the size of the main toroid as i can't figure it out from above.

[Toasty]

Re: Empire EMP-680

KK Power Tech

Clube do Hardware tested it quite successfully too. Sister site of Hardware Secrets.

At 80% (545w) load it did 79.2% efficiency, at 100% (682w) - 73.2%, at overload 106% (724w) - 66.5%.

I still don't like it. The output filtering looks small and I'd like to see how it was after 6 months of use.

Their testing only loaded the 3.3 and 5v to 33% compared to the label ratings, so don't be misled by 100% loading statement. Overall load was at 100% of label, but individual rails were not. Only the 12v1 & 12v2 were at 100%+ of label during the testing.

For "real world" usage, I guess the numbers are fine, but I wouldn't load it to more than 75% for an extended period of time.

Toast

[370forlife]

Re: Empire EMP-680

Their testing only loaded the 3.3 and 5v to 33% compared to the label ratings, so don't be misled by 100% loading statement. Overall load was at 100% of label, but individual rails were not. Only the 12v1 & 12v2 were at 100%+ of label during the testing.

Right, even KK power's website lists this unit as a ATX v1.3 unit, but hardware secrets/Clube do Hardware (btw, this is the exact unit that they tested) test per the new intel v2.3 loading guidelines.

I think 680W is nearing the top of the KK-9966D. It would need beefier primary switchers, more 12v capacity, and more filtering for the 12v to get more out of it. The Empire 500W they tested (btw, I have that unit too) kept ripple down somewhere at ~25mv on the 12v rail at 500W. On the 680W, which uses the same filtering on the secondary as the 500W, the ripple skyrocked up to ~100mv at 680W. I think some higher capacitance secondary caps would help quite a bit.