Discussion in 'Hardware' started by Bloody_Pete, 9 May 2011.
Anyone managed to test this?
I'm planning on checking mine starting from next week. There's 9 different types to get through though. There's a mix of Black Ice GTX, Stealth and SR1's, XSPC RX versions 1 & 2, RS, Swiftech and Aquatuning so might take a while.
I'm looking forward to seeing your results
Just being pedantic but can only use Xeons for dual CPU.
That first vid was an eye-opener.
What a simple test! It never occurred to me to try.
I'll be honest, I didn't know exactly how the flow worked in a rad in the first place.
Same, I thought it when through every channel in series, not half then half.
Anyone managed to test any?
i have done the flow test now on my black ice. it is good. took about 12 seconds for water to come out after filling started.
that would be terribly restrictive if it went snaking through back and forth, though might do a better job reducing temperatures.
also as an aside i tracked down why my temps are staying up. its the pump. ddc runs damn hot. needs its own frakin watercooling. i think it actually makes more heat than the cpu.
What flow test?
like in the article / videos.
If it flows out after 12 seconds thats very bad... It shouldn't flow out at all...
??? if you're pouring water into it, it has to go somewhere. i'd be more worried if you kept pouring and it never came back out again.
unless i missed something it seems like all they were doing was pouring water into one side and seeing if it came right back out the other=fail. ones that took some time for it to exit=pass.
No. Basically you should be able to fill the inlet tube, but the water will only be able to get to the same height as that level in the tube, so if this is below the top of the rad it wont flow anywhere and will just sit there. If you get flow out of the outlet then it means your rad stuffers from this 'bug', the larger the flow rate, the worse the 'bug' and the lower the performance...
I haven't had the chance to test any of mine yet, as I have 3 rads in total and 2 of them are in-use in my main rig.
The other is a cheap 120.1 with fixed fittings that I don't use any more, so I'd be surprised if it was properly welded on the inside, but..
..On the other hand, I get fantastic performance from my loop; so if I discovered that my main rads had this supposed flaw then I'd be wondering what all the fuss is about (except that maybe it could be better).
Have we had official comment from any of the manufacturers or designers on this issue yet?
I accept that it would be counter-intuitive and maybe naiive to just assume that this is how the radiators in question are supposed to be, but maybe that is the case?
As comments in this thread have highlighted, some previously-happy customers didn't even know how the flow in the radiators is supposed to go to begin with.
I can't help but wonder if this issue is really just a few guys who think they've uncovered a massive problem which is in reality a storm in a teacup stemming from a misunderstanding of radiator design.
A check on XS turned up the following thread, which is actually in response to this thread:
Some of the responses line up with what I was thinking, that this may be by design for reasons of pressure control. XS is notoriously heavy on the Thermochill fanboyism, but I'm inclined to believe what's written there, too.
No idea what the guy ranting about bit-tech using a rad is about though, since this is a forum thread and nothing to do with bit-tech/CPC itself?..
More of the same, which appears to be quite easily explained-away by the same information posted in the 2 XS threads.
It seems to me that there are just people out to publicise supposed design flaws in products that they don't understand from an engineering perspective.
That comment isn't aimed at you, Bloody_Pete; but more at the weekend reviewers and youtube bloggers who paint themselves as experts while raising non-issues to get people worried and riled-up over nothing.
So with the funnel or pipe filled it shouldn't be able to come out unless its filled ( higher than the rad) in order to pass, if it just comes out than oh dear. I might do that when I get my watercooling gear, just to see.
Nice to see EKs rad passing althrough there was only one as well as the black ices
Very interesting that so many could fail the test. I hope none of my rads leak like that. My temps are good though, so they seem to be doing their job.
Holy thread-necro, batman.
As I predicted in the post right before yours from 4 months ago, and as confirmed in the two links in that post.. It's by design and nothing to be concerned about, especially if your radiator is already performing just fine as it is.
People getting worked up over this supposed design flaw despite their loops already performing just fine are just falling into the trap of some sensationalist amateurs who decided to make a big deal out of this as if it were a matter of poor quality control instead of doing the professional thing and asking the radiator manufacturers about their findings first.
In fact, if they had done that they wouldn't look so ridiculous now.
I don't want to start a fight but even though i'm not a rad expert i don't think such a leak is good. However i wouldn't focus on the leak itself but those thermal shots they had. See that half the rad is near ambient temperature so excessive bypassing makes the rad to go underused. I think the leak guarantees minimal flow through the waterblock but then they should redesign the rad. Maybe they should let the user control the amount of bypass to achieve best performance in different flow conditions...
ITT: People who don't understand fluid dynamics. The hole in those radiators is tiny. There will be far more coolant being pumped into that radiator than will be able to fit through that hole, so a large majority of it will go through the core. It's by design to make sure that you don't get too much pressure build-up at the inlet, which would reduce flow throughout the entire loop.
TL;DR: Storm in a teacup, and a misunderstanding of radiator design.
Separate names with a comma.