![]() news news archive SEND NEWS!
articles
PC Database
SETI@HOME Team
|
Review by: James "Agg" Rolfe ![]() click to enlarge HWLabs are a new company operating out of Manila in the Philippines. I believe this "Black Ice" radiator is their first product. I received a pre-production prototype so some details are different to the production models which will soon be available. When I received it, I thought at first it was a converted automotive cooler but after inspecting it closely it does seem to have been custom made. In fact, the product page claims it is the first radiator to have been custom made especially for watercooling PC's. The entire unit seems to be copper, not adonized but perhaps sprayed black. It's obviously been designed with a 120mm fan in mind, being slightly wider and about an inch taller than most fans and having corner mounts for screwing a fan onto. The size of the unit and the orientation of the hose connectors means you can fit it most places where a 120mm fan would loosely fit. It fit fine inside the lower front of the AOpen HX08 and Macase K10 I have here- provided you rip the PC speaker mount etc out. I'd think any front 120mm intake could be modified to accomodate this radiator and the existing fan. ![]() click images to enlarge The design is actually quite complex. The water does 2 full "laps" up and down, describing a W or M shape. The water passes through groups of 3 flat blades instead of conventional tubes. Between the blades are finely corrugated fins and each of those fins is segmented into tiny sections. The flat tubing/fin design looks very similar to motorcycle radiators and you can see at a glance it massively increases the surface area of the tubing compared to a round-tube-through-fins design like the Cool-Computers unit we reviewed earlier. As noted earlier this is a pre-production prototype and most of the niggles I have with it are already slated to be changed in the production units. First up is that the radiator fins were a little crumpled in places, possibly due to the packaging but I suspect more likely due to construction difficulties - the prototype is hand-made. More evidence of lower-quality construction or shipping damage showed in the worrying form of a slow seepage leak in one of the corner seals. The packaging method will be different and the final units are being produced in a factory so both those issues should be resolved. A dab of RTV Silicone Sealant (still got the same tube from back in the original Watercooling Experiment!) fixed it - not that it was major anyway, perhaps a drop an hour. The mounting holes were a point of concern for me, in that a long screw could go through the plate and into the fins behind it, but apparently there's no danger of hitting the actual tubing and there will be nylon screws and thumb-nuts supplied for fan mounting anyway. The hose connectors are short straight tubes but in the final design they will be barbed connectors as per most of the watercooling gear we see these days. Speaking of the hose connectors, this unit thankfully uses the 9mm standard fishtank tubing - it seems Senfu really are off on their own with the thinner tubing now. The flow rate through the Black Ice is pretty good: ![]() click to enlarge The test system is the same as for the Cool-Computers radiator review, so you can compare the flow rate to the 3 systems shown in that review. It seems comparable to the Cool-Computers unit, perhaps a little less, but that might be explained by the flat-section tubing. Performance This is not going to be the most scientific test, as my thermal probes are both on loan to other people who are doing heatsink reviews at the moment. However, this system completed a 3DMark2001 batch run that lasted over 8 hours with no errors - which is close enough to "indefinitely" for me. The P3V4X's thermal monitoring is unfortunately really dodgy and the core temperature swings wildly between some clearly false extremes. The core seemed to sit at about 11C over the ambient temperature of the room though, which is not too bad at all. The water in the bucket remained cool and the back of the slocket felt only barely warm to the touch - compared to almost unbearably hot with the FKP-32 that normally cools this CPU. This is all with a Panaflo fan remember, which is designed with the goal of low noise rather than incredible airflow - if you were to slap a Sunon or similar high-powered monster on this radiator it would work even better. Out of interest, I ran the system with no fan at all for a while - the water heated up rapidly as did the radiator, but the system didn't crash for the 90 mins or so I left it running - seems the convection effect on the radiator was good enough in this instance - of course, mounted inside a case it'd be unlikely to last long like that. With the radiator completely removed from the water loop the system crashed in about 40 minutes. All this leads me to believe the unit is coping extremely well with the current heat load despite there being plenty of room to improve the efficiency of this test system (better pump, higher airflow fan). Conclusions Other Recent Reviews:
|
Major Sponsors:
|