Wintsch Labs - Arctic Web 437W TEC

I didn't read it yet, but I feel it's ok to say this:

Thank you very much! I've been dying to see how this thing performs, but nobody seems to have one!
 
<4Ghz was all you were able to get? What about overclocking a donthan, and an fx57, or a venice cored 3000 or 3500?
 
Guys, thanks a lot for finally reviewing a peltier block.

I've been mystified by what exactly these things are and just how well they work. Not only did you answer these questions but you also provided that exerpt on WHY they work. A+.

For some reason, as I was looking through the first few photos of the block itself, I got the feeling that it was not made with very good quality. You had to muck up the thermal paste and all... And in comparison to how clean and sleek other blocks look, e.g. dangerden... this one seems kind of sloppy. Not that you go for looks with a peltier... e.g. your Bike Tube! that was truly cool (no pun intended).

Once again, you guys kick ass when it comes to comprehensive reviews. Keep on the stuff.
 
I wonder how people outside of Alaska would enjoy the added heat generated by a 430W pelt?
 
krizzle said:
For some reason, as I was looking through the first few photos of the block itself, I got the feeling that it was not made with very good quality. You had to muck up the thermal paste and all... And in comparison to how clean and sleek other blocks look, e.g. dangerden... this one seems kind of sloppy.
Click to expand...


I agree. The copper seemed really rough, but, this is the next best thing to phase change and the results spoke for themselves.

Imagine putting a pelt on the Swiftech Storm.... haha
 
I found it an amazing performing block... however the companie's stock sealing methods didn't work... that worries me because if they didn't get the stock sealing methods right, what else did they screw up?

I think it has alot of potential and that it should be developed more, the company would be smart to listen to you guys :). Give em a ring (as I bet you already have) and tell them the issues you had. I'd buy it if they fixed it and demonstrated they could handle it, if I were in the position to do so verses a dangerden water block.

~Adam
 
robberbaron said:
I agree. The copper seemed really rough, but, this is the next best thing to phase change and the results spoke for themselves.

Imagine putting a pelt on the Swiftech Storm.... haha
Click to expand...

I was actually thinking the same thing, though I don't think it would perform that well. You need a wider cooling base.
 
Yes, it seems this Arctic Web block is primarily focused on contact surface area and water flow rate. The Storm doubtfully has any of those close.
 
Wow, it's been a while since I have read anything about Pelts. It's good to know that people haven't given up on the idea. Seems like the last waterblock with a built in pelt was years ago - the Swiftech MCW462-UHT.

Anyways, great review. I would have liked to see performance on a newer platform, but as far as heat loads go, a S478 Prescott is more than enough to give an accurate picture of what the Arctic Web is capable of delivering.

On a side note - It's funny how everyone and their brother bashes Koolance products, yet their performance is right up there with Danger Den and PolarFlo. (Admittedly, not the newest DD product, but it's not the newest Koolance either.)
 
Kyle and Gang,

Thanks much for reviewing this product. I've been eyeing it for a while. I've had some pretty mediocre experiences with pelts in the past, and I'm glad to know this combo actually works properly. It may be time to get out the Conformal Coating and the Luberex yet again.

Other than the Neoprene, did the kit come with Luberex or any other type of dielectric?
 
I actually think that the pins they used was very restrictive and the 2 inlets/outlets a stupid idea.

If they had some form of impingement design such as Cathar's but spread out to provide a wider rush in of water and the outlets taking the water from the far ends then the cooling performance would have increased.

The overclock to 3.96 should have not increased the temperatures that much ..

Oh well .. it performed well, but I'm sure they could have done a lot better. .
 
Great review. For 140 bucks though, you would think they'd have the condensation issues thoroughly worked out. Hopefully they will take a [H]ard look at the modifications you made, and implement them on their retail product.............
 
nice review. personally i would have just gone nuts with liquid electrical tape and silicone sealant to fix the bad seal, but thats just me.:-D

Can you share anything about the water temperature? especially comparing 1v2 radiators. While the second may not have had much of an impact on CPU cooling under most conditions it may have added value of keeping water temperature low enough that a northbridge or GPU TEC could be accomodated in addition to the CPU block. It's a pretty popular option.

and just FYI there's a botched URL under "System Configuration and Testing Procedure" and a "was would" in the 3rd textblock on the condensation page ;)
 
I liked the article, but a diagram of how you plumbed the 4 port block and the 2 rads would have been nice.
 
Quote from your review:

"First, you must prep your motherboard socket to prevent condensation. There are several ways to prevent moisture from getting into the socket area; the most popular method is to use thermal compound or dielectric grease in the socket area."

When I first read that, I almost crapped. Well, not really, but alarms went off in my head that many people might not realize that 'some' thermal compounds are electrically conductive to some extent or another, and putting this in your socket will hose your parts in a hurry when you turn the system on.

Dielectric grease is the correct stuff to use in most cases. I extolled the use of this back in my chilled water/peltier days with the Celeron 300a's, BH6, BX6-2 and dual Celeron 433's @ 715 (BP6) and it worked great until I froze the grease as well and it became a solid and cracked. Condensation was able to get in through the cracks. Really not a concern unless you are in the negative 20/30 degree range like I was.
 
I'm too much of a LimpGawd to respond to this article. Incredible stuff. Glad someone has the $$$ to try it out. Why no AMD Athlon64 variants?

IronChefMorimoto
 
krizzle said:
For some reason, as I was looking through the first few photos of the block itself, I got the feeling that it was not made with very good quality. You had to muck up the thermal paste and all... And in comparison to how clean and sleek other blocks look, e.g. dangerden... this one seems kind of sloppy. Not that you go for looks with a peltier... e.g. your Bike Tube! that was truly cool (no pun intended).
Click to expand...
the DD maZE 4-1 is not any cleaner to put together and install, since it comes in pieces and you have to apply your thermal compound and fit your neoprene yourself during the insall.

the swiftech pelt solution is a bit of a different animal, since it compes pre-assembled, but all three look pretty similarly messy once installed and running.
 
IronChefMorimoto said:
I'm too much of a LimpGawd to respond to this article. Incredible stuff. Glad someone has the $$$ to try it out. Why no AMD Athlon64 variants?

IronChefMorimoto
Click to expand...

Because we need to know how it performs under extremely hot conditions and AMD is just not up to par ;)
 
101 said:
I wonder how people outside of Alaska would enjoy the added heat generated by a 430W pelt?
Click to expand...
probably only a little more than i enjoy the heat from my 226 and 80 watt combined.

you learn to adapt.

:)
 
IronChefMorimoto said:
I'm too much of a LimpGawd to respond to this article. Incredible stuff. Glad someone has the $$$ to try it out. Why no AMD Athlon64 variants?

IronChefMorimoto
Click to expand...
have a look at DNA's thread.

his temp monitoring is quite comprehensive.
 
robberbaron said:
I agree. The copper seemed really rough, but, this is the next best thing to phase change and the results spoke for themselves.

Imagine putting a pelt on the Swiftech Storm.... haha
Click to expand...
mohammedtaha has it right. the storm only cools the cenrtal area of the block well. using one with a pelt would be pretty silly.
 
Neutraleyes said:
Wow, it's been a while since I have read anything about Pelts. It's good to know that people haven't given up on the idea. Seems like the last waterblock with a built in pelt was years ago - the Swiftech MCW462-UHT.
Click to expand...
naw, swiftech has the 5002 mated to a pelt and DD has the maze 4 mated to a pelt.

neither block is anywhere near this performance level, though.
 
mohammedtaha said:
I actually think that the pins they used was very restrictive and the 2 inlets/outlets a stupid idea.

If they had some form of impingement design such as Cathar's but spread out to provide a wider rush in of water and the outlets taking the water from the far ends then the cooling performance would have increased.
Click to expand...
you can't have your cake and eat it too.

to get jet impingement happening, you need really fast moving coolant.

you can get that in a core sized area with a normal pump, so it's a really good solution.

to get the high turbulence region happening over this large an area, it would just take too much pumping power.

cathar posted about this stuff in a thread regarding a direct die cooler, and altough he was focusing on what it would take to get a square centimeter's worth of jet impingement from a single outlet, think for a second about trying to get at least 46 square cm of jet impingement even with multiple nozzels.
 
DFI Daishi said:
you can't have your cake and eat it too.

to get jet impingement happening, you need really fast moving coolant.

you can get that in a core sized area with a normal pump, so it's a really good solution.

to get the high turbulence region happening over this large an area, it would just take too much pumping power.

cathar posted about this stuff in a thread regarding a direct die cooler, and altough he was focusing on what it would take to get a square centimeter's worth of jet impingement from a single outlet, think for a second about trying to get at least 46 square cm of jet impingement even with multiple nozzels.
Click to expand...

While I was typing this I started to think of the power needed to get it accomplished .. it's not impossible with a "normal" pump, it's all about tackling it different. Cathar will soon get it done ;).
 
i really like the motherboard retention system that they are using. it looks much nicer to work with than the DD solution, which i can say first hand is somtimes a pain during the initial build phase because lining things up takes a while. it is also a pain every time i clean and re-seat my processor.

as for condensation proofing, i really think that the precautions that you folks take in the review are most like the minimum standards, rather than a safe and long term solution for a computer that will be run at least 8 hours daily.

i have had my accidents with my pelt rig, however my current solution seems to be quite good in terms of long term viability:

before assembling you computer mask off the sockets, jumper/IO pins and slots of the board, as well as any hot running chips/ICs that you plan to put additional cooling on later, apply several coats of spray-on conformal coating, fill all your sockets and slots with dielectric grease.

take similar precautions with your add-in cards and vid card.

use well-fitted neoprene gaskets around all of your chilled surfaces.

after going through this regime, and then fittign everything together, condensation is minimized. if some condensation or frost forms for whatever reason, it's probably not going to be the end of the world.
 
mohammedtaha said:
While I was typing this I started to think of the power needed to get it accomplished .. it's not impossible with a "normal" pump, it's all about tackling it different. Cathar will soon get it done ;).
Click to expand...
his promised VGA block that is pelt compatible?

i eagerly await it, however keep in mind that it will most likelly be intended for use with a wee little 40x40 mm pelt, as opposed to a 68x68 mm monster like this.
 
DFI Daishi said:
his promised VGA block that is pelt compatible?

i eagerly await it, however keep in mind that it will most likelly be intended for use with a wee little 40x40 mm pelt, as opposed to a 68x68 mm monster like this.
Click to expand...

yeah .. well 40x40 is the only way for VGA unless GPU's get bigger. There's no other way around it. We'er also going way off topic ..

I believe I've seen that same pin design somewhere else .. Someone has done the same thing before but I can't recall where.
 
mohammedtaha said:
yeah .. well 40x40 is the only way for VGA unless GPU's get bigger. There's no other way around it. We'er also going way off topic ..

I believe I've seen that same pin design somewhere else .. Someone has done the same thing before but I can't recall where.
Click to expand...
just indicating that i think it might not be faasible on a pelt of this size.

back to topic: swiftech's diamond pins ala 5000 series blocks look pretty similar to me.
 
So does a 226W TEC give half the cooling? What kind of temps would you be looking at with only a 226W TEC? (Temps on, say, the article's setup @3.96 Ghz as compared to 61F with the 437W.)

Also, why did you have to take the block apart? Was something wrong with it? I thought you were just showing us how it was put together for the purpose of the article, but someone said that it had a crappy design b/c you had to to do that. Did you need to clean off the thermal compound?
 
"easiest way to remove thermal compound is to use Permatex Blue Label cream hand cleaner"

Never fail to learn something [H] every time :D
 
oh oh

" it just goes to show that more coolant capacity did have an impact on the performance of the TEC"

Pretty sure you mean "more cooling capacity ..." as in the addition of the extra radiators cooling capacity instead of the additional coolant (water) added in the loop due to the second rad as the wording implies.
 
I was very glad to finally see a review of some more exotic cooling back on the [H]. I have seen this block show up on procooling forums and there seem to be mixed reviews. I would have to say that if you don't already own a water cooled system you are almost better off looking for a used Vapochill or Prometia...things ALWAYS seem to add up when setting up a H20/TEC cooling solution. That doesn't mean I wouldn't buy one of these if I didn't have to buy a brand new Meanwell...never thought the S320 would be too small.

I would have to agree that the condensation prevention was a bit lacking...dielectric grease is your friend...seriously...lather your socket area with the stuff, make sure all pin holes in the socket are filled...if you can see PCB between the neoprene gasket and the socket...fill it up with dielectric grease. Conformal coating is also a good idea as previously mentioned, you can never be too safe...condensation is your worst nightmare.

To answer the question about the 226W performance...My 226W pelt (DD Maze4-1)...on my 3500+ winchester (non-overclocked at the moment) I usually idle at around 28c...which is a bit high for me, but I haven't really had a chance to tweak yet, plus the four 120 fans on my rad are barely spinning to keep the noise down. For right now, you could achieve the same performance with a Zalman. However, if I crank the fans up it would drop those temps quite a bit.

I would second DFI's claim about the DD (or anyones) TEC/Water block mounting setup...sometimes a PITA, and unless you are a real stickler for looks...it wont look all that good. Danger Den also doesn't have neoprene Athlon64 socket gaskets available (haven't for sometime)..which required me to cut my own, which again was a PITA. Mine turned out OK...

Also if you are constantly at your CPU socket swapping chips...the is not the kind of setup you want.

As for heat generated by pelts...I live in Florida...it usually is 95F and up during the day...it can be 78-80 ambient in my house, but when you walk into my office you notice a definite heat build up...and I don't even want to think about it if I still had a CRT...
 
Russ said:
So does a 226W TEC give half the cooling? What kind of temps would you be looking at with only a 226W TEC? (Temps on, say, the article's setup @3.96 Ghz as compared to 61F with the 437W.)
Click to expand...
if you tried to do this with a 226 watt pelt, the assembly would start to cry, and you would cry shortly afterwards, when temps started to climb up to what you would expect from running the block alone.

an athlon would be another matter, but the current pentiums would just overwhelm that size of pelt.
 
226watts seems like it is alot... is that 226watt dissipation per second or per minute or what?

~Adam
 
CleanSlate said:
226watts seems like it is alot... is that 226watt dissipation per second or per minute or what?

~Adam
Click to expand...
dude, do you not know what watts measures?

and it is not dissipating 226 watts, it is pumping a maximum of 226 watts across itself.

it only ever pumps that much heat when both the hot side and the cold side are at 20C temperature, and a 226 watt thermal load is applied to the cold side. given that coolant temps will never be that low, let alone the hot side of the pelt, it's kind of a "magic" number that you will never see in real world opperation.

what you should actually be looking at to figure out cooling performance are the delta T per watts load graph for the pelt, and the watts pumped per hotside temp graph. then figure out how cool you can realistically keep the hotside with a cooling solution that you can stand the noise of and can afford, and see where it falls on the two graphs.

all that boils down to is that 226 watts isn't going to do a whole heck of a lot on a prescott unless you have quite unusually low coolant temps, and even then it's not a very good value for the bit of money that you are saving as compared to a 437 watt pelt.
 
It was obvious in my post that NO I didn't know what the watts measured, I had an idea though. Learn new things daily.

Yet I never know anything.. oddly enough.

~Adam
 
CleanSlate said:
It was obvious in my post that NO I didn't know what the watts measured, I had an idea though. Learn new things daily.

Yet I never know anything.. oddly enough.

~Adam
Click to expand...
well, i tried to answer the part about 226 watts being a lot. i would suggest the repository of knowledge that is wikipedia, if you want to learn aout what watts are. :|
 
You must log in or register to reply here.
Back
Top