Printable View
Alot of Independent Hardware tests have the samples supplied by
companys that make the product.Due to not being telepathic i
wrongly assumed you supplied the samples.But i'm sure if you send
him a sample he would gladly test it as you seem to like his results
and you use them for advertising your rads at any opportunity when
a new user is looking to buy a Radiator.
:D
As Radical undertakes this testing for free, unpaid, we aren't prepared to drop an entire range of rads on him to test, as that would take up weeks of his time... we supplied the PA120.2. We weren't approached to provide a PA120.3 for the initial tests - he sourced it himself. If/when he wants/has time to test a PA160, sure he'll ask fer one...
@digz: I purchased the PA 120.3 because I wanted to know how well it performed. As TC had no local import here, and many people hesitate to order internationally, I didn't see much sense in asking them to send me one.
Now, after the review, I saw that many people are interested in it and want to have one. Also, as it was so much better than competition, the idea came up to compare a PA .2 to the #2 rad of the competition, the Swiftech MCR320.
As the numbers show, even a PA .2 can compete with any other .3 rad.
PA 160 would be interesting to test also, even if I think a slightly different testbed would be just fair for it to compete on. I'm still waiting on some stuff to arrive here, if that works out I'll drop Marci a mail for sure :)
yah very hard to order internationally, xpensive air freight, not all accepting int'l CC, etc etc.
by sponsoring a said test/review people/some sites tend to be more bias to make a positive review for said item...
this is not good if you're to remain an unpartial product tester...
Daam... i got an tripple pro (got an tripple to compensate for a bit lower performance) cus that was thin enought to mount on the top of my case (internaly) but i could easily mount a double rad in front of my case instead...
oh well atleast i now can easily get cold air from outside straight into my rad from the window....
@septim: Well, that woulnd't even have been the problem in this case. All EU, no customs and Paypal transaction. Doesn't get easier than that ;) Still you're right if it's overseas, also as you said shipping is quite pricy anyhow then.
I agree with you on the sponsoring, at least if it's more than sending in some parts and giving support (in a technical way).
@exa: That's one point why I thought it would be interesting. It's way easier to fit a .2 rad in most cases than a .3. I hope that results like these do give the folks more options when choosing a case, even if they want to put all the fancy hardware in there like dual GPU and such.
Questions for Radical_53:
1.In the first test, the mcr320 had a temp of 8.4c@600rpm vs 9.2c in the second test. The temp was 6.9c@900rpm in the first test and only 6c in the second. The results .8c hotter then .9c colder between the two tests, Why is that?
2.How much (if at all) did the i.d. of the barbs differ between rads in the flow test.
3. What do you consider the margin of error to be for the test results?
Thanks, FLB
I used the same barbs on all rads for the tests, all EK 1/2" barbs.
You can't directly compare the both tests, as the setup was somewhat different. The second setup hat a higher flow rate and one block less on the loop.
Margin of error would normally be 0.1°C +/- for each sensor for what we're measuring here (a temperature difference), but as we took multiple measurements (logged), it would only change the absolute results, but not the relative result from one to the other contender.
With enough data you can still make fairly reliable indirect comparisons. Reducing load and increasing flow rate will reduce temps as we clearly see at 900rpm.Quote:
Originally Posted by Radical_53
The water drops .8c@.855gpm at 900rpm and 1200rpm for both rads. This temp drop represents 180 watts +/-22.5 watts (.8c +/-.1c).
At 600rpm the mcr320 drops 1.2c while the Pa drops .7c. This suggests that the mcr320 hasn't reached a steady state (or that you were stepping on the hose, j/k). I'm guessing that you preheated the rad before turning on the fan. Is this correct?
Were the alphacool rads new or used? The pressure drop of these rads seems to hint that some of the channels are blocked.
Thanks for taking the time to answer my questions,
FLB
The Alphacool rads were new, as all others except for the two Black Ice Extreme rads, were one was mine and one came from Jupiler. Before I used them initially, I cleaned them to some point (no week-long vinegar rinsing thing, but basic cleaning).
About the test itself: I didn't heat the rads up before measuring took place. I heated them up with running fans, logging temperatures and took the final values as soon as they were steady for at least 5-10 mins. After that, load was taken off with fans still running, until the temps went to a steady idle state. Then fan speed was changed and the next heating period started again. For the temp readings, I don't know if that's written here somewhere: I used one probe on all "normal" temps, but two for the air inlet temp (to be sure). All sensors were calibrated before the measurements took place at different temps to be sure that all worked correctly.
About the wattage, that was right on the second test. My readings gave me roughly 195W heat output.
About the comparison: Sure, a rough comparison is possible. Sadly I coulnd't keep the exact same test setup, direct comparison would have made it easier. I'm still not through the process of having just the testbed I'd want, so I'm still changing something here and there. As I'm also changing things like sensors and measuring equipment, that means that each test run itself is so far "correct", but comparison to another test run can lead to problems in detail here and there.
Thanks for your quick reply. Was the first test ~224W? (195/6x6.9).Quote:
Originally Posted by Radical_53
What are/were you using for the heat load?
Is there a link to were you describe your setup so I don't have to bother you with so many questions?
Thanks, FLB
The description of the setup is in the article ;) Basically I used my SLI setup (-> sig) as a heat source, yet the second test had slightly lower CPU clocks + voltage and a missing chipset block. I used a different CPU block on the first setup also.
As Marci said, I don't get paid for these tests, so I do have to use the stuff that I got here. My "main" test rig doesn't have enough heat for a rad test, so I had to use my own rig for that purpose.
So are you happy with the PA120.2? That is the only one i could find in stock and ordered it. I wanted the 120.3 but no one had them in stock....Quote:
Originally Posted by Radical_53
BTW Marci and Radical thanks for all your tests and commitment. Tests like these help us consumers buy items : ).
Yes, totally satisfied. For me personally, I upgraded my case like a year ago to use a bigger rad, as my old .2 rad wasn't capable of giving me good temps even with a single card.
So now, I haven't finished that case yet, but had always planned to use a .3 rad in it. No matter which .3, it would have been really close, blocking quite some of the slots in the front. With a PA .3, I'd even have to let my PSU stick out a little as it's so long.
Now, as I've seen of what the PA .2 is capable of, I'll certainly use it for that rig. It doesn't block my slots in the front, so that I can use other stuff I might want to put it, I don't have my PSU stick out and I'll still have at least the performance on a level that no other than a PA .3 can provide. I'm hoping that a custom made 140mm shroud will unleash even a tad more performance, but that hasn't been tested yet. Gives me the opportunity to run a SLI setup, all overclocked, with really silent operation in a very slim "package". I'd say it won't get much better than that :D
Seen this yet d00d??
Am assuming this is the family from which the earlier-posted 120.4 is from - same as BIP is to BIX, so GT is to GTX...Quote:
The next step in the Black Ice GT evolution is coming...
Enter the Black Ice GTX
Yes, seen it and waiting for some... they claim it's not just a double-thickness conversion, let's see what's in it.
What kind of heatload did you use to test PA120.2 vs MCR320? Do you think that the PA120.2 would surpas MCR320 with higher CFM fans like sanyo denkis or would it be worse?
Heat load wasn't that much different on both setups I had. The first setup, were all the .3 rads were tested, had a lower CPU clock of 2.8GHz, but higher voltage (1.55V) and the chipset was on the loop. The second setup was with 2.9GHz and 1.41V, but without the chipset on the loop.
Heat load was ~194-202W according to my sensors on the second test, should have been somewhat higher on the first test (225-230W).
About the higher cfm on the PA, I'll see :) I'm still about to get a shroud made for 140mm fans and then I'll do a test on that, see if it does improve performance even further. But for now, I do only have the data of the normal 120mm YL D12SLs, so I'd really have to test that out. The improvements from more air flow got smaller the higher the rpms of the fans got, so it might not be too much that you can gain there.
Hmm... Just a 2.9ghz 1.41v DC opty? Very low heatload... Trying to pick a new rad for kents/4x4/k8l whatever. Maybe a single R600 in the loop. PA120.2 is what I really want but PA120.2 + shroud + gaskets + 2 denkis is almost 175 dollars. Hell of a lot of money for a radiator.
Oh, you should read the rest of what the setup consisted of ;) Those were the changes from the setup with all .3 rads to the comparison of the .2 PA and the .3 MCR. There was a SLI setup of 2 7800GT cards (full cover block), running at 500/1400MHz on the loop also ;)
Should be plenty of headroom if you plan to use only a single card then.
Ah alright :) Thanks!
Wonder what ever happend to Cathar bench testing the new rads ?
Wonder what happened to this being a sticky?
I unstuck it. I never planned it to be stuck forever.
Australian summer + spectacular motorbike crash.Quote:
Wonder what ever happend to Cathar bench testing the new rads ?
Would rather he concentrated on getting himself fully fixed up than hassle him for further testing results.