Radical CPU coolers from CoolIT

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
forums are to speak the truth - why did tom's let such a poor product slide with a ok review?
advertising dollars?
big business relationships?

Um... nope. Mostly because I don't think it's a poor product, the reasoning behind which I stated in the review.

Sorry to blow the air out of your conspiracy theory, but here's some sobering news for you: just because someone don't agree with you doesn't make them wrong, nor does it prove there's a conspiracy. :roll:


I'm still using the Eliminator and it's working fine, in fact I'll be using it in the overclocker's marathon in a week or so. Not only will I be pushing it to much higher limits, but if it dies - or performs well - I'll sure as hell be telling people about it.
 
Cleeve -

If this product competes against water and water is touted as a way to cool an ultra overclocked CPU, wouldn't it make sense to compare it at high OC?

Ok, ok, you've said that the stock wouldn't cool it past 2.25GHz. That's fine. So you have a 2.4GHz graph and it says Intel Stock N/A, then you have your 3.0GHz graph and it says Intel Stock N/A. See... not hard. You just show the Stock wouldn't even boot, thus showing value, in that, you can go from a 2.25GHz OC to a 3.0GHz. That's worth something.

Now, my Tuniq can boot the e4300 at 3.06GHz but I haven't tweaked the voltages so it isn't stable. It is stable near stock volts at 2.94GHz, but I run it at 2.7GHz just because I'm lazy and get good temps (55C load, measured in TAT, probably around 35-40C if I was gay and measured it from the board).

Question: How did you measure your temps? As far as I'm concerned, Intel's TAT program is the only way to measure. Taken from the silicon, it will give you a higher reading than the board. My board will read something ridiculous at boot like board 35C, CPU 20C. If that was near true, I'd be condensing all over the place with an air cooler. It's not, and the board sensor is just a simple function, like T_silicon - 15C (or something more complicated which is approximated very well as something like a simple offset, not sure).

I think from the review you gave the clear winner was the Tuniq Tower and similar air coolers.
 
i found this to be quite an informative article. but i have to agree with previous posters that ur overclock was a bit lacking. my e4300 runs with stock cooling at 3GHz. temps r pretty much same as urs. it idles higher but stays under 55C at full load. perhaps its just the motherboard? i use a msi p6n nforce 650i sli motherboard.
 
I just don't know why these results are so different:

http://www.guru3d.com/article/processor/438/

http://www.extrememhz.com/freezone-p1.shtml

http://www.hardcoreware.net/reviews/review-327-1.htm

http://www.techpowerup.com/reviews/CoolIT/Freezone

http://www.driverheaven.net/reviews/CoolIT/index.php

Maybe the freezone tested in guru3d was running evian water...
(Note that the super pi 1M result in the guru3d review suffered with the OC, probably by not being stable, the last iteration takes 5 seconds)
 
Nice article, this review brings me to the day of the pc3 2000 and the i820 chip-set. If people remember, toms review of the pc3-2000 mainboard. Toms said it was the fastest board they ever tested with lots of good features but did they mention that there was a problem with the MTH causing data errors. Me as being a loyal reader of toms and trust what they test and say was a foolish mistake on my part. However I do appreciate the work toms do for readers like the stress test of the PSU and others, but I’ve learned not to take all too deep from the mistake of buying an i820 mainboard on the advice of toms write up in 2000.
criticize an article is a good way to say maybe users can do the same job for cheaper or maybe look into the fact that someone else has tried it with bad results, like I and others did with the pc3 2000.
 
I think this article is not realistic, like others said, too little oc, with my ZALMAN 9700 I can push me Brisbane 4800+ x2 to 3.0 GHz, only air cooled. I live in tropical island with hot summer, with temp of 91 to 96 F degrees and me computer is used for gaming a lot, movies, video and no problem. So too much money for that.
 
was this review done by a bonafide/extreme overclocker? if you gonna plunk that much on a cooler you might as well do extreme overclocking.
 
You're totally right, I missed including that. The ambient temp was 21 degrees.

The watercooling system in theory is limited to ambient temp, although in reality we can see it go a couple degrees below in this case. I'm not positive how that worked - either air movement over condensation on the the radiator caused evaporation to cool it below ambient, or the temp sensor was out a few degrees.

I don't believe too much on evaporation at those temps. So I have to believe all measurements (because of the sensor/software) are way wrong, and I don't think I could compensate using a +3 degrees since that seems to be a curve not a constant.

And as I say, a Pentium D 820 (which is good to have on winter) is measuring 65 degrees using a ThermalRight XP-120 so I don't believe a C2D with a TEC should get to >60 degrees.
 
It is sad that the state of discourse in this world has devolved to rants. It is not actually communication to label something or make conclusory statements without factual support. Merely because you get all emotional about some inanimate object does not contribute the knowledge base of the reader of your posts. It actually defines those who pour their feelings out only for the sake of venting their emotions, as an immature person lacking self-control.

The purpose of the article clearly state that it had a limited scope. It also said it was limited (for now) to the OC that the author felt comfortable with initially using the stock cooler. It was not purported to be an exhaustive study of all available types of CPU coolers, nor of extreme overclocking.

People you should appreciate the effort. Take what is useful to you from the article and control yourselves.


cleeve,

Old saying: Never expect gratitude.
another not so old saying: No good deed goes unpunished.
 
friend of mine who is a regular here (Hi Kevin!) told me about this review and asked me to put in a few words, since I had that product for a few months now.
so rather than commenting on a review itself I will simply list my own findings.

first off, I knew right away that eliminator has issues, simply because so many people on the boards were moaning and groaning despite generally positive reviews. I was planning to do some heavy duty modding and was sure that I could overcome any problems the factory setup had. (this turned out only partially true).

anyway, on with the show...

out of three settings eliminator has (low/medium/high) low and medium cannot keep up with a 2.8GHz CPU (overclocked E4300). With a 100% TAT load it takes about 15 minutes to overheat it and get to a point when system just shuts down. In terms of noise these settings are somewhat acceptable, even though far from what I'd consider a quiet system.
When you switch it to High, Eliminator can keep up with 100% TAT load on 2.8GHz overclocked E4300. It is very noisy though and I don't consider 2.8GHz a good overclock, so something needed to be done.

I took everything apart (bye-bye warranty) and replaced that noisy 92mm SOB with a 120mm Delta GFM fan (its very, very noisy...) that I had from a previous project. It can put out about 160cfm. I basically just wanted to see how far eliminator can go with a real blower on it. Well, it could sustain steady 3GHz overclock, but 3.2 will make it crap out in about 15 minutes. Again, I don't consider this to be a good overclock for a watercooler.

So at this point I wanted to see why is this silly thing cools so poorly - I took a t-balancer from my main box and wired it up with sensors. I got temps from waterblock, coolant (at one of the pelts), radiator, air inside of the box, air exiting the radiator, and, obviously, good old Coretemp to give me an insight at what's happening with CPU.

After spending an evening testing, here are my findings.
At idle Eliminator cools great :). It can take a coolant and a cpu block to about 16C from 25C ambient in under 10 minutes. (Then again, at idle E4300 barely produces any heat).
Things change when you put it under real load. I found that because water loop is so tight (it's probably about 150ml of coolant altogether), and has realtively little inertia, pelts can be easily overwhelmed by even a modest overclock. It goes like that - you start at 16C idle, and watch the coolant temperature climb slowly to about 30-31C when CPU shuts down due to overheat. It takes 15 to 20 minutes to go through this cycle.
You can help it by keeping a radiator very cold (I stuffed an ice pack inside just to see what will happen). Yes, that helps. With a radiator being ice-cooled peltiers could handle 3.4Ghz (well, for 10 minutes). Except this is, of course, stupid.
Overall I find radiators design so poor that even a Delta fan did not make as much difference as I was hoping for.

Anyway, long story short - kinda ok pump, pretty decent waterblock, but puny peltiers and a really bad radiator design make this unit a NO GO for overclocking. It's OK if you need to cool a stock CPU, in particular if you are in enviroment with high ambient temperatures. For $150US, though... I don't know.

So here's what I ended up doing with my Eliminator:
1. I dumped its 92mm fan. It's a joke anyway, in terms of cooling capacity per noise produced. Right now I have 120mm Panaflo with 92-120mm adapter attached to it. It's controlled by Speedfan.
2. I plugged the pump where fan connector used to go, on that little control board. Now I can control waterflow with Hi/Med/Low settings. It's set on Medium.
3. I powered pelts with 16V from an old power supply I found in my parts box.

overall it can do 3.2GHz now, at 100% TAT load. It's not great but you know, whatever, you've got to stop at some point. When/if I find a cheap and decent water radiator I'll probably add it to a loop, then it'll be something worth having.

Hope this helps.
 
I was very pleased to see a comparison with the Exos 2. There aren't many decent comparisons around.

Having used an Exos 2 for about 18 months, in many different systems and configurations, and having tried many DIY water cooling systems for comparison, I will say that anyone who thinks an Exos 2 is a "not so good" water cooling solution doesn't know what they're talking about. My own experience with the Exos 2 is that its the best complete system on the market and hard to beat with a DIY system for the same price. The Exos 2 rules for:
1) the ease of setup
2) reusable hoses and thumbscrews
2) the ability to drain the system from outside the case
3) the cooling capability with 1/2" hoses (because most computer enthusiasts don't seem to grasp the fact that cooling is dependent upon turbulent flow moreso than radiators, pumps and fans, but any decent injection molder does)
4) reliability

As far as the CoolIt system, I won't be trying one based on the many complaints. That info is priceless.

As a final note, I'm sick of smelling coolant in the air. I am running a E6400@3.68GHz stably on air (that's right, AIR), I won't be in the market for a liquid cooling solution in the near future. My advice: Get a Thermaltake Big Typhoon, replace the 120mm fan with a Thermaltake Smart Fan II (3500RPM), and when you see and can't hear the results of your $60 investment you'll never go back to stinky/messy/fugly water cooling. 🙂