Time to take the Warlock apart and find out what makes it tick. Literally.
From these two shots, it is not immediately apparent who the OEM is for this unit. I hear some of you asking if those are Nippon Chemi-Con capacitors you see in the lower shot, and to this I reply that's indeed what they are. With a couple exceptions on the non critical capacitors, every capacitor is a Japanese name. Matsushita on the primary, Chemi-Con on the secondary. What's upsetting to me is that with what I've seen so far, this is actually a decent unit. Apevia had no need to overrate this design, it would have stood on its own very well as, say, an 800W unit.
But, I'm getting ahead of myself a bit here. Let's fully take it apart first before I give Apevia what for.
Here we see the green flexible PCB that encircles the 135mm fan to provide the four tri-color LEDs with power.
Speaking of fans, here's our wind monster. The UL number traces to... get this... U-Lucky Electronic Co. I lucky? I just reviewed this unit, I don't feel so lucky. Perhaps if Apevia had included a free box of Lucky Charms I'd feel more lucky.
My first step in disassembly was to remove this here PCB from the center of the unit where it was bolted to the secondary heatsink. This is the fan LED controller, and it finally tells me who made this unit - Wintech. Hmm... I didn't know they were capable of something that could hit 87% efficiency.
PCB removed, we can see parts of the transient filter attached to the backside of the receptacle: an X capacitor and two coils.
Also with the PCB removed, we can see how the primary capacitor area is silkscreened for the Empire State Building of capacitors, but where it got a smallish Matsushita cap instead. That little daughterboard sticking up next to the cap is the NCP1653 based PFC controller - no combined PWM/PFC all in one solution here.
Looking on from the front, we see the rest of the transient filter which includes a MOV, two Y capacitors, and one more X capacitor. The small heatsink on the far left above the yellow X capacitor holds the bridge rectifier, a single Rectron RS1506M good for 15A.
You might be wondering what that little heatsink on the secondary is with the coils attached. I'll get into that in a minute. But first...
a shot of the solder side of the PCB. Which has another sticking point for me, namely various components that have not had their leads entirely cut properly. Several lead segments had been left dangling, and were then soldered in place. Sloppy, sloppy. Say, do these pics look a bit off to you? I think this unit is making Mr. Fuji sick.
By the way, there are various markings on the top of the PCB that lead me to think this is an 800W design. That explains a lot, doesn't it?
What's this little card? Why it plugs into the PCB right between that large coil and the standby transformer. It is responsible for both OCP and fan control duties.
Looks a little barren there, with the heatsinks out. That daughterboard in the center of the unit, next to the main transformer, houses the PWM controller itself, a UCC2894. Nice to see 14 gauge wiring being used there on the secondary outputs.
Moving on, here we have the primary side heatsinks. The PFC section, top right, consists of two 20N60C3's and one LQA08TC600 diode. On the top left, we find the switching section, and it is a configuration I can't immediately identify. First, we have two 11N80C3's that appear to be in a double forward arrangement. I say they appear to be, because I couldn't find the diodes. Instead, that one 08N80C3 you see facing you is seemingly attached to the source pins of the two switchers. I'm sure I'll be hearing from a few of you the moment the review goes live, anxious to tell me all about it.
Things get interesting on the secondary. At the top of this picture, you can see four F29077 transistors. These appear to be providing the entire 12V output for the unit, using synchronous rectification. Ah, that's why we hit 87% in test two - this is potentially a very efficient method of regulation. The module on the bottom takes that 12V power and buck regulates it down to 3.3V and 5V.
So, essentially what this PSU needs is a more powerful 3.3V/5V module to sustain the 900W numbers. Well, that and better ripple suppression. Let me just take that module apart, and...
...grumble about the inconsistent quality control that saw only one too short standoff for three screws, allowing the module's PCB to bow in the middle. Hmm... maybe that's what caused the lousy 3.3V/5V numbers. At any rate, that little IC there is an NCP5435 dual buck controller that works with a total of six NTD4804N's (three for each rail) to provide 3.3V and 5V output. Why they used thermal pads and thermal grease is beyond me. It's redundant. It's redundant.
Let's move on to the conclusions before Mr. Fuji gets sick enough to work a different kind of magic all over my carpet.
