Skinnee Labs - GTX480 Block Round-Up (working thread)

Quote:

---

Originally Posted by Waterlogged

I don't know what size those holes are but, as long as it can still pass the volume of a 1/2" I.D. tube, does it really matter what shape they are? ;):shrug:

---

I measured and calculated the cross section of the watercool connector a while back, and it comes out to a diameter close 6mm ID tubing, i.e. really bad. If you look at the new Watercool 58x0 block you'll see that the corrs sectionw as greatly increased, I don't think Rico did that just for fun, especially since he's talking about the new block being flow optimized (which tests show it is).

The AC block seems to have a cross section at the connector similar to that of the old Watercool blocks.

That small narrowing of cross section can really make a difference as well, if you look at Cathar's testing [URL="http://www.xtremesystems.org/forums/showthread.php?t=147767"]here[/URL,] you'll see a sizeable difference (close to 10% flow) between push-ins and regular 3/8" barbs, all because the push-ins don't reduce the cross section slightly like the barbs do.

Basically what I'm saying is that it's a shame because it would be so easy to make the cross section bigger and significantly increase flow.

XSPC block makes it way to the lab...

http://skinneelabs.com/assets/images...0_contents.jpg http://skinneelabs.com/assets/images...0_prodshot.jpg http://skinneelabs.com/assets/images...TX480_base.jpg http://skinneelabs.com/assets/images..._internals.jpg

http://skinneelabs.com/assets/images...internals2.jpg http://skinneelabs.com/assets/images...internals3.jpg http://skinneelabs.com/assets/images...X480_inlet.jpg http://skinneelabs.com/assets/images..._prodshot2.jpg

Waiting for one more part to arrive from the Egg and testing can finally begin!

if its not asking too much can you also do a brief writeup on the ease of mounting of each block, i know some blocks can be a real PITA to mount: requiring, reinforcement plates, back plates, slipping when trying to assemble etc...

Quote:

---

Originally Posted by Coldon

if its not asking too much can you also do a brief writeup on the ease of mounting of each block, i know some blocks can be a real PITA to mount: requiring, reinforcement plates, back plates, slipping when trying to assemble etc...

---

Actually, that is why I am waiting for a part from NewEgg. :D

I sat down this weekend to video the stock heatsink removal as well as the installation of each block only to discover the webcam I planned on using absolutely sucked for recording the removal and mounts. 720p webcam is en route. :up:

oh epic - cant wait, i kinda watched the video for the installation of the aqua computer block and decided that it was too much effort dealing with a backplate, the DD installation guide was quick and simple so it climbed to the top of my list...

Anyone want to guess the order of performance of the blocks? My money is on the swiftech & dd blocks being the top two performers, with the koolance & ek being close seconds, and that the BP block will be the worst of the lot...

depending on which temp you will rank them i say the koolance,Aqua and swiftech will come at the top for core temp. As for overall performance, i think it will be between the koolance and the aqua.

Skinnee

you getting the ac backplate? i didn't see it in the pics...

**i see a revised version..nevermind

It's in this pic:
http://skinneelabs.com/assets/images...0_contents.jpg

:)

Quote:

---

Originally Posted by EnJoY

Yum...micro-channels. Should perform well, very curious to see the results of this one.

---

I hope this one performs very well. Seeing as Aquacomputer is 2nd to the EK in terms of cooling performance (In regards to current 5870 waterblocks), I'm confident they'll do pretty good.

I'm debating whether I should go with EK (even with the issues going around of cards getting shorted) or should I go with Aquacomputer?

Quote:

---

Originally Posted by skinnee

XSPC block makes it way to the lab...

http://skinneelabs.com/assets/images...C_prodshot.jpg

---

Am I the only one that feels disappointed with XSPC these days? I mean seriously they are supposed to be competing with EK in terms of their GPU blocks but in regards to their 5870 blocks, they are near the bottom. Sure they make great reservoirs and even great CPU blocks but they need to either focus more on improving their GPU blocks or just abandon making GPU blocks altogether.

Whatever the case. I will definitely not be buying the XSPC block unless it miraculously performs EXTREMELY well. -- Am I also the only one that finds it kinda "ugly"? It ain't bad but it's no beauty either.

Quote:

---

Originally Posted by Coldon

Anyone want to guess the order of performance of the blocks? My money is on the swiftech & dd blocks being the top two performers, with the koolance & ek being close seconds, and that the BP block will be the worst of the lot...

---

I like guessing games....lemme try :cool:

I'm gonna do a performance score though....10 being best, 1 being stock.
9-10 - Swiftech
9-10 - Koolance
9-10 - AquaComputer
8-9 - EK
8 - XSPC
7.5 - DangerDen
3.5 - Bitspower
1 - Stock

Skinnee hasn't even put the card on water yet, so I have no insider info and these are purely guesses :p: And these are just core temps, the Swiftech heatpipe design is too much of a curveball for me to begin to predict.

I'm eagerly awaiting the results :) At the moment, I'm considering either the EK nickel + plexi or the Aquacomputer all nickel version whenever they are released. I like the Danger Den but the weight of the block is a concern and the lack of a backplate. Does anyone know when EK will be releasing their backplate? Watercool are supposed to be releasing their block this week and it's also supposed to be coming with a backplate, so that will be interesting to see. From the look of the blocks, my thoughts would be that aquacomputer, koolance and EK will be the top performers. It's a shame bitspower didn't put more effort into their block design...

Hey just so you know your links on the first page to each block. Your Koolance link goes to the DD block page.

Quote:

---

Originally Posted by Vapor

I like guessing games....lemme try :cool:

I'm gonna do a performance score though....10 being best, 1 being stock.
9-10 - Swiftech
9-10 - Koolance
9-10 - AquaComputer
8-9 - EK
8 - XSPC
7.5 - DangerDen
3.5 - Bitspower
1 - Stock

Skinnee hasn't even put the card on water yet, so I have no insider info and these are purely guesses :p: And these are just core temps, the Swiftech heatpipe design is too much of a curveball for me to begin to predict.

---

Lets start a pool! :up:

Quote:

---

Originally Posted by HuffPCair

Hey just so you know your links on the first page to each block. Your Koolance link goes to the DD block page.

---

Fixed, thanks for reporting! :up:

Quote:

---

Originally Posted by Vapor

I like guessing games....lemme try :cool:

I'm gonna do a performance score though....10 being best, 1 being stock.
9-10 - Swiftech
9-10 - Koolance
9-10 - AquaComputer
8-9 - EK
8 - XSPC
7.5 - DangerDen
3.5 - Bitspower
1 - Stock

Skinnee hasn't even put the card on water yet, so I have no insider info and these are purely guesses :p: And these are just core temps, the Swiftech heatpipe design is too much of a curveball for me to begin to predict.

---

as per your format:

9-10 - Swiftech
9-10 - DangerDen
8-9 - Koolance
8-9 - EK
7-9 - AquaComputer (not sure about those inlet holes)
6 - XSPC
3.5 - Bitspower
1 - Stock

I'm hoping that the heatpipe transfers a decent chunk (but not all) of the VRM heat to the block to having the rest is dissipated by the heatsink and as a result offering better core temps.

Performance in FC vga blocks isn't just about core temps, in my opinion. Mem and vrm's are important, especially when overvolting...

If u want best core temps buy a MCW-80...

Quote:

---

Originally Posted by JoeBar

Performance in FC vga blocks isn't just about core temps, in my opinion. Mem and vrm's are important, especially when overvolting...

If u want best core temps buy a MCW-80...

---

I agree. For full cover blocks the total package wins it for me. I like cool vrms, memory, core and a fancy backplate if they can be made.

I feel that everything has to be relatively cool in order to achieve the best overclocks and long term stability, and that's the benefit of full cover blocks.

Hmm, asked Gabe when the blocks would be released, he said you have to go through EVGA. Meaning have to buy the cards with the blocks on them already or be able to buy just the blocks?

Also, who has the 470 blocks out? I see Bitspower has one at Sidewinder and EK has one at PPC. Any others out there for the 470s?

Which one is better in YHO? The EK one is a lot cheaper.

We also have a block for the GTX 470:

http://www.abload.de/thumb/gtx470_1vk7b.jpg http://www.abload.de/thumb/gtx470_21moq.jpg http://www.abload.de/thumb/gtx470_3j7tc.jpg http://www.abload.de/thumb/gtx470_5u7l9.jpghttp://www.abload.de/thumb/gtx470_6yuy8.jpg

http://ac.shoggy.de/youtube_aquagrafx_gtx_470.jpg

Shoggy, where is for sale in the US?

Nowhere :D

But as always you can also order it directly from us ;)

and is there a possible way to get some single slot pci bracket for the gtx 470 or gtx480 ?It would be a very nice accessory to include with full cover block.

We offer that as as accessory.

Quote:

---

Originally Posted by Shoggy

Nowhere :D

But as always you can also order it directly from us ;)

---

Shipping must be really expensive!!

Quote:

---

Originally Posted by Vapor

I like guessing games....lemme try :cool:

I'm gonna do a performance score though....10 being best, 1 being stock.
9-10 - Swiftech
9-10 - Koolance
9-10 - AquaComputer
8-9 - EK
8 - XSPC
7.5 - DangerDen
3.5 - Bitspower
1 - Stock

Skinnee hasn't even put the card on water yet, so I have no insider info and these are purely guesses :p: And these are just core temps, the Swiftech heatpipe design is too much of a curveball for me to begin to predict.

---

Wooh, wait a second... The EK is in 4th place? And XSPC is tied for 4th?? I can see the Swiftech being the top performer but why is the Koolance > EK? IT should be the other way around, in my mind that is xD

Quote:

---

Originally Posted by Conumdrum

Also, who has the 470 blocks out?

---

We launched our 470 block at the same time as our 480 block.

Quote:

---

Originally Posted by blazarcher

why is the Koolance > EK?

---

I can't confirm anything because I haven't compared EK's 480 to ours. But I think the logic assumes microfins > less complicated shapes, which is often correct.

Tim

Quote:

---

Originally Posted by blazarcher

Wooh, wait a second... The EK is in 4th place? And XSPC is tied for 4th?? I can see the Swiftech being the top performer but why is the Koolance > EK? IT should be the other way around, in my mind that is xD

---

Look at the microchannels on the Koolance :slobber:

Compared to the EK Wave design from like...2007? EK shouldn't keep up with the micropin array of the Swiftech or the microchannels of the Koolance or AquaComputer in terms of performance (if they're all done right, that is....a too-thick base or poor contact can kill the performance of any design quickly). Kind of like saying a Whitewater can keep up with a Heatkiller LC :shrug:

Quote:

---

Originally Posted by Vapor

Look at the microchannels on the Koolance :slobber:

Compared to the EK Wave design from like...2007? EK shouldn't keep up with the micropin array of the Swiftech or the microchannels of the Koolance or AquaComputer in terms of performance (if they're all done right, that is....a too-thick base or poor contact can kill the performance of any design quickly). Kind of like saying a Whitewater can keep up with a Heatkiller LC :shrug:

---

Didn't I see a pic of you somewhere :ROTF:

Quote:

---

Originally Posted by Vapor

Look at the microchannels on the Koolance :slobber:

Compared to the EK Wave design from like...2007? EK shouldn't keep up with the micropin array of the Swiftech or the microchannels of the Koolance or AquaComputer in terms of performance (if they're all done right, that is....a too-thick base or poor contact can kill the performance of any design quickly). Kind of like saying a Whitewater can keep up with a Heatkiller LC :shrug:

---

I'm a bit confused. How can you explain the EK (5870 waterblock) performing better than EVERYONE. Here is a link, not sure about the validity of the source...
http://translate.google.com/translat...ndup%26hl%3Den

Yeah a mega-link xD

Anyways here is the graph that I am arguing from.

http://img294.imageshack.us/img294/1440/gput.jpg

That's just the GPU temps, the VRM temps don't matter as the EK only wins because it has a backplate. -- I guess we can guesstimate that the AquaComputer will excel here xD

Here is a shot of the EK block showing the EK wave design. Although old, I guess it did pretty well here in that regard...

[IMG=http://img405.imageshack.us/img405/6290/ekg003t.jpg][/IMG]

So is the danger den block the absolute best in terms of cooling performance (ignoring the Swiftech)? Or is Koolance just as good?

----
Sorry about the questions but I was SERIOUSLY considering the EK block and now am a bit confused xD

Again sorry about the so-called large pic, I would have assumed everyone would be running at least 1920x1200 xD

Skinnee's gona test it all milord !

blazarcher, just calm down a bit... people are just placing their bets as to how everything is gonna turn out. They are not telling you which block you need to get or which leg you need to put in your pants first when putting them on. And resize that giant photo please, I hate the sideways scrolling.

Quote:

---

Originally Posted by skinnee

And resize that giant photo please, I hate the sideways scrolling.

---

Get image zoom addon for firefox and you can resize :up:

Quote:

---

Originally Posted by Tacops

Get image zoom addon for firefox and you can resize :up:

---

Is that for the massive crazy pic size poster or for Skinnee? I responsibily resize my pics before linking to them on Photobucket.

Need a uber detailed pic? I make a link for the visually needy with 90" monitors.

Quote:

---

Originally Posted by blazarcher

Shipping must be really expensive!!

---

Always 34.90 EUR for a parcel up to 31.5kg.

Conundrum: i second skinnee's opinion - even though i use 1900x1200 27" screen, mostly i use browser window of size at half of that resolution for eg. several windows side by side. Problem with larger images is, that text gets wrapped with them in posts in same longish lines, and gets on my nerves each time when i need to maximise browser just to be able to read text. Your case is not that bad, i've seen some WIDE pics posted (like 2000-3000 and more px wide :D ), but you might note as future reference, that it would be polite to others in forum to use pics no more then 1024x768, but preferably 800x600 or even less.., even if rules allow, it's simply respecting others.

personally my choice in gtx480 block is based as follows: 60% overall performance (GPU, mem + VRM) + 40% ease of assembly/dis-assembly, the back-plate on the AC block has already ruined it for me, I see absolutely no reason for it apart from aesthetics, one can argue it provides a more uniform mount but i think the effect will be minimum and isn't worth the extra hassle IMHO (see how i said IMHO there, so don't flame me).

Quote:

---

Originally Posted by Vapor

Look at the microchannels on the Koolance :slobber:

Compared to the EK Wave design from like...2007? EK shouldn't keep up with the micropin array of the Swiftech or the microchannels of the Koolance or AquaComputer in terms of performance (if they're all done right, that is....a too-thick base or poor contact can kill the performance of any design quickly). Kind of like saying a Whitewater can keep up with a Heatkiller LC :shrug:

---

... eh.

At this point, I'm not taking any bets. There's simply too many extra factors in GPU block performance. :shrug:

can't wait for results, my DD480 block with CPU included in loop on a Feser 360 stays around 34c idle and highest I've seen the 480 get to even while overclocked and overvolted to 900/2150 I've only hit 52c load :D:D

You always need to distinguish between cooling performance and flow restriction, both being important for the block's overall performance. And what's more important to you depends on what kind of loop you're running and on your preferences. I for one am almost sure that the EK will probably come out on top for me as usual. Lowest pressure drop and temps that aren't much worse than those of the micro channel blocks. If I were to go for a dedicated GPU loop though, I wouldn't care much about flow and only consider thermal performance. But for a CPU + GPU loop I'll gladly sacrifice a few degrees on the GPU for higher flow through the CPU block.

Vapor was only talking about thermal performance I think.

EK now have the backplates for the 480 blocks up on their store. They look real nice and definatly made my decision on which blocks to get a whole lot easier :) It's a shame they arent bundling them with the blocks like Aquacomputer but it still works out cheaper for us Aussies anyway.

http://img97.imageshack.us/img97/9137/31788035.jpg

Still waiting for the Nickle plated GTX480 aquagraFX... Asked in the Aqua-computer's English section of their forum, but its like a ghost town in there... :(

Shoggy, can I pay a down payment to pre-order one??

Quote:

---

Originally Posted by eternal_fantasy

Still waiting for the Nickle plated GTX480 aquagraFX... Asked in the Aqua-computer's English section of their forum, but its like a ghost town in there... :(

Shoggy, can I pay a down payment to pre-order one??

---

A-C-Shop in Germany are selling some nickel plated Aquagrafx 480 blocks. I got my nickel plated heatkiller 5870 blocks from there a few months back.

http://www.a-c-shop.de/Aquacomputer-...ckelt-A-C-Shop

It's tempting but I really need plexi for my build. Another company needs to make some plexi tops for their blocks to give EK some competition on the plexi front.

@Coldon: keeping in mind that the GPU can be bend (like with our card) you will gain a lot by using the backplate since it pushes the GPU against the copper and can compensate that problem.

@eternal_fantasy: will still take some time. As long as we sell the copper one like hell we will not hold back a few to nickel-plate them since that takes a bit while others scream for blocks ;) Since that nickel plated variant does not exist in the system I can't also do an order confirmation for it.

Quote:

---

Originally Posted by converge

A-C-Shop in Germany are selling some nickel plated Aquagrafx 480 blocks. I got my nickel plated heatkiller 5870 blocks from there a few months back.

http://www.a-c-shop.de/Aquacomputer-...ckelt-A-C-Shop

It's tempting but I really need plexi for my build. Another company needs to make some plexi tops for their blocks to give EK some competition on the plexi front.

---

Thanks for the link!! Just put down an order with help from Google Translate. :D
Hope the site is reputable... Somehow got a 15% discount code already entered, so it came out to be around 100 euro sent to the UK. :up:

Quote:

---

Originally Posted by Shoggy

@eternal_fantasy: will still take some time. As long as we sell the copper one like hell we will not hold back a few to nickel-plate them since that takes a bit while others scream for blocks ;)

---

Thanks for the reply. :)

Quote:

---

Originally Posted by Shoggy

@Coldon: keeping in mind that the GPU can be bend (like with our card) you will gain a lot by using the backplate since it pushes the GPU against the copper and can compensate that problem.

@eternal_fantasy: will still take some time. As long as we sell the copper one like hell we will not hold back a few to nickel-plate them since that takes a bit while others scream for blocks ;) Since that nickel plated variant does not exist in the system I can't also do an order confirmation for it.

---

shouldnt the 8 screws around the GPU area be sufficient to prevent a bad mount due to bending? I can totally understand why it's nice to have the option of the back plate but while I personally feel like it may be unnecessary, skinnees testing will ultimately tell...

i just received my gtx480 today and after hearing that abortion of a stock cooler at full blast, i need to get a waterblock ASAP.

The fans on GPU's are just annoying and I am deaf as a post too. I've been playing with the clocks on my card as I wait for the new camera to arrive and last night just couldn't take the fan noise anymore and shut the test bench down... yes, I really want that NewEgg order to hurry up so I can get the card on water!! I will have some numbers/clocks to share from pushing the card on air later this evening though.

edit: That EK backplate looks good. I was surprised to see only AquaComputer supplied a backplate. To me backplates are like standoffs (especially for the longer cards)... they should be the norm and not the exception.

(To me backplates are like standoffs (especially for the longer cards)... they should be the norm and not the exception.) EXACTLY ... But im sure a company like EK that gets so many customers knows they can still sell a backplate seperately

That EK backplate is really nice looking :toast:

@blazarcher...check the internals of the 5870 rev1 Koolance block vs. their GTX480 internals, there's been a massive update. And my performance conjecture is merely just conjecture of course :p:

Quote:

---

Originally Posted by blazarcher

I'm a bit confused. How can you explain the EK (5870 waterblock) performing better than EVERYONE.

---

As Vapor pointed out, our original VID-AR587 didn't use microfins. The VID-AR587T2 (for later HD 5870's) does, though.

There are certainly other factors, like quality of contact, cold plate thickness, etc. But all things being equal, I will be surprised if the microfin coolers don't top out the list. ... And the Koolance 480 block uses microfins on both the GPU *and* VRM area. ;)

Tim

Quote:

---

Originally Posted by skinnee

blazarcher, just calm down a bit... people are just placing their bets as to how everything is gonna turn out. They are not telling you which block you need to get or which leg you need to put in your pants first when putting them on. And resize that giant photo please, I hate the sideways scrolling.

---

I am calm xD Sorry about the pic, I'll get that resized.
And it's not really the fact that "they are telling which block you need to get" it's a personal preference. I was just a bit skeptical about what you said concerning the EK blocks (not having micropins = not THAT great performance).

I just want to see which one performs the best that's all.

Quote:

---

Originally Posted by Koolance

As Vapor pointed out, our original VID-AR587 didn't use microfins. The VID-AR587T2 (for later HD 5870's) does, though.

There are certainly other factors, like quality of contact, cold plate thickness, etc. But all things being equal, I will be surprised if the microfin coolers don't top out the list. ... And the Koolance 480 block uses microfins on both the GPU *and* VRM area. ;)

Tim

---

Speaking of prior blocks.... What happened to making blocks that were gold plated? Like the one's you did for the 2900XT/3850/3870, Koolance VID 290, 385, and 387 respectively?

Working on a build right now and am having trouble picking which waterblock to go with. First I was going towards EK's Nickelplated waterblock (w/ backplate) then hearing that the Aquacomputer version performs better, I then thought why not go with Aquacomputer... Finally hearing that Koolance is doing very well with their micropin design, I'm thinking now either Danger Den or Koolance (or AquaComputer).

Now if you were to release some gold plated waterblocks, now that would be the deciding factor.

----

If that doesn't work, I came across this site http://www.goldplating.com/Products/...-brushgold.htm

They supposedly offer solutions that allow you to "gold plate" bare copper. And I was thinking of doing this on one of EK's waterblocks. Does this seem like a viable option or should I just forget about the gold (bling) and just go with a nickel plated waterblock instead?

Quote:

---

Originally Posted by eternal_fantasy

Thanks for the link!! Just put down an order with help from Google Translate. :D
Hope the site is reputable... Somehow got a 15% discount code already entered, so it came out to be around 100 euro sent to the UK. :up:


Thanks for the reply. :)

---

No worries. Yeah the store is great, Jochen is a pretty nice guy and sorted the order out for my hk 5870 blocks a while back :up:

Quote:

---

Originally Posted by skinnee

edit: That EK backplate looks good. I was surprised to see only AquaComputer supplied a backplate. To me backplates are like standoffs (especially for the longer cards)... they should be the norm and not the exception.

---

Yeah I agree, backplates are a must. I heard watercool are bundling backplates with their hk blocks as well, whenever they appear...

Quote:

---

Originally Posted by converge

Yeah I agree, backplates are a must. I heard watercool are bundling backplates with their hk blocks as well, whenever they appear...

---

Yeah if they ever appear lol. Seriously, they are taking way too long. Might as well go for AquaComputer or Koolance instead.

I got my 480s today and noticed the ek backplates were now in stock so I decided to bite the bullet and ordered the last 2 ek nickel plexi blocks and 2 nickel backplates :up:

If/when I change the theme of my build, I'll most likely try out the aquacomputer blocks. The microhannels on the koolance block looks nice but the block is just too flashy for me and is missing a backplate. I'm sure you could make the ek one fit though.

Quote:

---

Originally Posted by converge

No worries. Yeah the store is great, Jochen is a pretty nice guy and sorted the order out for my hk 5870 blocks a while back :up:

---

Just received order dispatch confirmation from a-c-shop.de for my Nickel plated aquagraFX GTX480!!!! The ETA on the site for the nickel plated version is for the 26th, was not expecting it to be dispatched so soon! Had to change my order, added a MIPS RAM block which came out much cheaper then anywhere else with their 15% off everything before end of May. Great service and awesome shop. :up:

Will post some shots of the plated block when it arrives hopefully early next week. :D

Skinee. I read that you are not using a flow meter in testing. But as I look at these blocks, I can definitely see which ones will perform better, but at the same time restrict more flow. To me the deciding factor is more than just the best temps, it's also how much flow the block will restrict. I thought this was an important aspect of all water blocks - maybe I am wrong.

I guess the pressure monitor will tell us how good the flow is? (Sorry if thats a stupid question)

Quote:

---

Originally Posted by Boulard83

Forgot something ... you NEED to use the right fittings in the pic as teh INLET.

---

Let the test results do the talking instead.

Phil

Armitage, you read my mind. Vapor and I had been talking about restriction and flow rates and we came to the conclusion that I have to monitor and capture flow rate of the loop regardless of my "issues". So, a King 7520 was looped up.

http://skinneelabs.com/assets/images...480_bench5.jpg

not the best pic, but you get the idea. :)

Edit: Philwong... what is that quote from, I can't seem to find the OP from Boulard?

I was refering to this block :

http://skinneelabs.com/assets/images...internals2.jpg

In order to have flow in the VRM section you need to used the right hole as inlet.

Quote:

---

Originally Posted by Boulard83

I was refering to this block :

http://skinneelabs.com/assets/images...internals2.jpg

In order to have flow in the VRM section you need to used the right hole as inlet.

---

I know..

Phil

Quote:

---

Originally Posted by Boulard83

I was refering to this block :

http://skinneelabs.com/assets/images...internals2.jpg

In order to have flow in the VRM section you need to used the right hole as inlet.

---

Shouldn't it be on the contrary? If you use the right as inlet then water goes through both entries of the VRM channel and, as such, there will be no waterflow. Instead, if you use the left one as inlet, once the water has gone through the GPU it will enter the VRM channel and then leave (provided the outlet part of the VRM channel is closed so that no water can get in).

At least this is the way I see it...:shrug:

Quote:

---

Originally Posted by prava

Shouldn't it be on the contrary? If you use the right as inlet then water goes through both entries of the VRM channel and, as such, there will be no waterflow. Instead, if you use the left one as inlet, once the water has gone through the GPU it will enter the VRM channel and then leave (provided the outlet part of the VRM channel is closed so that no water can get in).

At least this is the way I see it...:shrug:

---

You would think that would be the case but there would be a siphon effect from the rest of the water going past the little piece of copper acting as a diverter.

@Prave
I really doubt your right. really !

As Waterlogged told, if you use the left one, you need the outled water to cause a syphon in order to "pump" water in the VRM area.

Quote:

---

Originally Posted by Boulard83

@Prave
I really doubt your right. really !

As Waterlogged told, if you use the left one, you need the outled water to cause a syphon in order to "pump" water in the VRM area.

---

Actually, it looks like it could work the same with the water flowing in either direction.

Still think the best is the right one as inlet.



Skinnee should test this with this particular block.

What we need is a plexi side, drop of colour dye and a web cam. :up:

Quote:

---

Originally Posted by skinnee

Armitage, you read my mind. Vapor and I had been talking about restriction and flow rates and we came to the conclusion that I have to monitor and capture flow rate of the loop regardless of my "issues". So, a King 7520 was looped up.

http://skinneelabs.com/assets/images...480_bench5.jpg

not the best pic, but you get the idea. :)

Edit: Philwong... what is that quote from, I can't seem to find the OP from Boulard?

---

Excellent ! Looking forward to the reviews. I particularly love the look of the EK blocks, enough sometimes to pick them over a slight loss in performance :)

I am waiting for independent testing from phatty labs myself.

Quote:

---

Originally Posted by Boulard83

Still think the best is the right one as inlet.
Skinnee should test this with this particular block.

---

Quote:

---

Originally Posted by eternal_fantasy

What we need is a plexi side, drop of colour dye and a web cam. :up:

---

+1 on both counts!

The XSPC approach to cooling the VRMs should work just fine, IMO. This barb will ensure that flow goes over the VRMs. "Path of least resistance" dictates it will take all paths, which means over the VRMs (if that barb is aligned properly). I doubt that inlet vs. outlet will make much of a difference in end performance unless flowrates are so extraordinarily low that the heat from the VRMs increases the water temp noticeably before it gets to the core.

EDIT: and in case it's unclear, I don't think it matters much if that port is inlet or outlet when it comes to flow distribution (i.e., no observable thermal difference)--I skip that logical explanatory step above.

Quote:

---

Originally Posted by Waterlogged

Actually, it looks like it could work the same with the water flowing in either direction.

---

Yeah lol, either way the water in the VRM section HAS to go somewhere. So either will do. Seriously, don't you think XSPC already thought about this?

Ill wait for real testing. ;)

Skinnee, i want you to test the XSPC in/out importance.

Quote:

---

Originally Posted by skinnee

Armitage, you read my mind. Vapor and I had been talking about restriction and flow rates and we came to the conclusion that I have to monitor and capture flow rate of the loop regardless of my "issues". So, a King 7520 was looped up

---

Nice, because I just saw this on Youtube:

http://www.youtube.com/user/trubrita.../0/yFQRorMZb4k

Quote:

---

Originally Posted by cx-ray

Nice, because I just saw this on Youtube:

http://www.youtube.com/user/trubrita.../0/yFQRorMZb4k

---

Yeah I saw that too. But honestly, I don't care much about flow rate. All I care about is looks and cooling performance xD

Quote:

---

Originally Posted by Boulard83

Ill wait for real testing. ;)

Skinnee, i want you to test the XSPC in/out importance.

---

I want you to?:rofl:

Yes i want !

or at least ..

i hope !

Quote:

---

Originally Posted by Vapor

The XSPC approach to cooling the VRMs should work just fine, IMO. This barb will ensure that flow goes over the VRMs. "Path of least resistance" dictates it will take all paths, which means over the VRMs (if that barb is aligned properly). I doubt that inlet vs. outlet will make much of a difference in end performance unless flowrates are so extraordinarily low that the heat from the VRMs increases the water temp noticeably before it gets to the core.

EDIT: and in case it's unclear, I don't think it matters much if that port is inlet or outlet when it comes to flow distribution (i.e., no observable thermal difference)--I skip that logical explanatory step above.

---

That fitting is my biggest concern with the XSPC block. I'd really like to know what size those holes are as that could pose as a really big restriction point if they're too small.

Quote:

---

Originally Posted by blazarcher

Yeah I saw that too. But honestly, I don't care much about flow rate. All I care about is looks and cooling performance xD

---

Cooling performance is end result of flow rate, more flow=better performance, to a certain point anyways .

Quote:

---

Originally Posted by Coldon

shouldnt the 8 screws around the GPU area be sufficient to prevent a bad mount due to bending? I can totally understand why it's nice to have the option of the back plate but while I personally feel like it may be unnecessary, skinnees testing will ultimately tell...

---

The screws around GPU area combined with the pressure from the block around these screws (and over the VRMs and such) will make the card bend. I really wish i had one for my 285 GTX blocks.

Quote:

---

Originally Posted by blazarcher

Yeah I saw that too. But honestly, I don't care much about flow rate. All I care about is looks and cooling performance xD

---

I'm interested in what the end result will bring if you for instance have two Swiftech GTX 480 blocks in the same loop with your CPU and mobo. Considering that GPU temp is rather secondary (with water cooling it's already 40-50c below the stock coolers), I'm prepared to sacrifice a few degrees there in favor for better CPU temps.

Video for stock HSF removal... First one is a little rough, but I'll get better as I progress.

Video wasnt too bad for your first. Im weary though on using goo gone on PC Components

I saw Skinnee's nose ;)

Nice video, got a few suggestions though:
- edit in some close-up pictures, e.g. highlighting the screws (I think Shane did this as well), showing the pad at the back and how far you need to remove it etc.
- talk more about other options besides the cleaning stuff you used, i.e. isopropyl alcohol, acetone etc.
- explain why you're doing what you're doing; I've been doing this kind of stuff for ~8 years now, but I've never used a coffee filter, do you use it because it's lint free? I usually use paper towels and then a lint-free lens cleaning cloth...

Had no problems with GooGone what so ever and have been using it for quite a while.

Good feedback, thanks Nickel! :up:

And yes, the coffee filter is for lint free, cheap and with the amount of coffee I consume daily they are guaranteed to be available in the house. :)

Quote:

---

Originally Posted by skinnee

Video for stock HSF removal... First one is a little rough, but I'll get better as I progress.

---

Do you know any "hot" chicks that could assist you during your videos? :shocked:

If I did do you think you would've seen me on camera? :p:

Progressing along here, since nVidia did not give us VRM sensors we had to come up with our own.
http://skinneelabs.com/assets/images...-vrmprobes.jpg http://skinneelabs.com/assets/images...vrmprobes2.jpg

And yes, I need to clean up those hot glue whiskers yet.

Quote:

---

Originally Posted by Nickel020

I saw Skinnee's nose ;)

Nice video, got a few suggestions though:
- edit in some close-up pictures, e.g. highlighting the screws (I think Shane did this as well), showing the pad at the back and how far you need to remove it etc.
- talk more about other options besides the cleaning stuff you used, i.e. isopropyl alcohol, acetone etc.
- explain why you're doing what you're doing; I've been doing this kind of stuff for ~8 years now, but I've never used a coffee filter, do you use it because it's lint free? I usually use paper towels and then a lint-free lens cleaning cloth...

---

I agree on some points Nickle, but this isn't a watercooling 101 class. Be nice to have one. Skinnee does the deep work and I'd rather him focus on the test than why he uses GOO GONE to pre wipe a CPU for TIM application. We all know that, or ones watching the vid really should know before they get to this level.

And he's very economical with that coffee filter, but doing this all the time, costs add up, even coffee filters!

I have paper towel bits scattered everywhere, and depending on my mood, the same with coffee filters.

For the fun of it, this is a pic of Skinnee at the 2010 XS party at the go cart track next to the Palms Casino. We's just chatting. Nice guy!

His nose is perfectly normal.

http://i245.photobucket.com/albums/g...kinneedone.jpg

Hope you don't mind Skinnee, and there were no crazy women at this party, was typical geek stuff. I live in Vegas, the wild women only hang with the folks with mo' money that any of us have.

LOL, you didn't have to blur out my face... its all good. I do have a big snoz though. :p:

A couple weeks back in Detroit...
http://skinneelabs.com/sample/_V8U4702.jpg

In GTX480 block testing news: Bitspower block was mounted up (video coming), running the first stock clocks/stock voltage run and the system hard locked. Power-cycled, instant FF. Been troubleshooting for the past few hours, and will try the board again tomorrow.

However, we have 2 options if the Classified is dead...
1. Wait for RMA
2. Run with the Rampage II Extreme

I presume you folks would like results rather than wait for RMA, so I'm switching over to the R2E now.

can you move the sensors a mm nearer the vregs? hehehe

Quote:

---

Originally Posted by Hondacity

can you move the sensors a mm nearer the vregs? hehehe

---

:)

They are as close as they can get and still maintain clearance of the chokes. Sensors are reading 30C+ at idle as I get the system changed over the R2E.

WoooW

look goods

Quote:

---

Originally Posted by skinnee

LOL, you didn't have to blur out my face..

---

What a legend :up: such a class act!

I've literally a pair of msi 480 gtxs sitting here I refuse to touch until skinnee has finished his testing, do you have any idea what so ever on an ETA for the results?

& I know it goes without saying but your effort, hours and hard work is very much appreciated

Quote:

---

Originally Posted by Conumdrum

I agree on some points Nickle, but this isn't a watercooling 101 class. Be nice to have one. Skinnee does the deep work and I'd rather him focus on the test than why he uses GOO GONE to pre wipe a CPU for TIM application. We all know that, or ones watching the vid really should know before they get to this level.

---

Well I disagree. The video is/should be a WC 101 since the people who don't need to hear all the details don't even need the video anyway. The point of it is to teach people who have never done this... And yes, Skinnee's time is best spent testing, so making videos you can only understand if you know how to install blocks is a way bigger waste than making 101 videos. As nice as bling videos are, Skinnee making entertainment videos for enthusiasts is really a waste of his testing time.

Quote:

---

Originally Posted by Hondacity

can you move the sensors a mm nearer the vregs? hehehe

---

Quote:

---

Originally Posted by skinnee

:)

They are as close as they can get and still maintain clearance of the chokes. Sensors are reading 30C+ at idle as I get the system changed over the R2E.

---

Would it not be better to use calibrated thermocouple sensors here? With the Dallas sensors you can't get close enough to the mosfets, which is what you want to be actually be measuring. Sacrifice some accuracy in order to get closer to the action^^

Rico sent me some pics today, which i would like to share with you:

http://www.abload.de/img/gpu_x3_gtx480_2kw83w.jpg

http://www.abload.de/img/gpu_x3_gtx480_1kwv80.jpg

http://www.abload.de/img/gpu_x3_gtx480_3klvtl.jpg

As you can see: Backplate not shown on the pics

Looks great, as always!

I wonder if there are any changes to the internals compared to the 58x0 block...

Quote:

---

Originally Posted by skinnee

Video for stock HSF removal... First one is a little rough, but I'll get better as I progress.

---

I would be very leery of using and/or recommending others use localized cold shock via liquid CO2 to loosen up the adhesive on that foam strip... the components and board are designed to handle the gradual heating of the reflow process, making a hairdryer or heat gun the ideal choice for the job. The liquid CO2, if not used very carefully, could cause all sorts of damage (think of it along the same lines as why you don't want to spill LN2 on a board... sure, the liquid CO2 isn't as bad, but it's still not a good idea).

Quote:

---

Originally Posted by Nickel020

Would it not be better to use calibrated thermocouple sensors here? With the Dallas sensors you can't get close enough to the mosfets, which is what you want to be actually be measuring. Sacrifice some accuracy in order to get closer to the action^^

---

That was my initial idea... use the UEI DT302 I have for LN2, but that was going to be more hassle for very little gain. From the runs last night on the BP block, the sensors are working very, very well. Keep in mind, if NV would've just done the proper design and provided VRM sensors we wouldn't be having this discussion in the first place. :)

Just checking back after a weekend full of testing...

Bitspower installation video is in editing, max testable OC found and saved (found OCP for my PSU in the process, and I swear you can burn your fingers off on the GTX480 PCB at the VRM's using Afterburner Extreme for voltage control), 3 mounts and 6 runs completed. The rest of the blocks will be tested in the order received, but I plan on having install videos released along the way. Oh yeah, the VRM sensors are working extremely well!

Also, I did hear back from Rico @ Watercool and they will be sending a block for the round-up this week. I'll post again on that topic once the card has shipped... I am not counting on it until I know the block is en route. I have a busy work week ahead of me, but there will be progress made on block testing guaranteed!

:toast:

Nice skinnee !

We appreciate your work :)