What do you want to see in a Water Block???

[jaydee]

Price? Performance? Both? What kind of designs do you like the best?

I am starting a business building water blocks and would like some input from actual users as to what they would like to see. i have a few of my own designs and would like some of your thoughts so I can make the best blocks possible.

Any and all input is appreciated.

Thanks

Update: already started making blocks and here is the inside design. read through the thread or check out my website for testin results.

[WebMasta33]

I would like to see a place sell a 3/4"-14NC Tap... I need one to make threads for my hose barbs, and I can only find a 3/4"-10NC or a 3/4"-18NC tap....

Seriously, a equal blend of performance/cost is what I would think 80% of people would want.

Cost always matters, but if it doesn't work well, it's a waste of money.

[Brad]

they are designed for performance, if you want cheap, go buy a hsf instead. I wouldn't like to go too far about US$60 myself.

Compatibility with socket 370 is in my mind now too

[redleader]

Good clamp. In the end how tightly you can clamp without killing your core makes a world of difference. Not to mention that I'd like to be sure that my CPU is safe and secure when I mount it.

[jaydee]

All the blocks I plan on building will be mounted through the holes around the socket. This may be a problem though for SOC370 boards as I never really thought about those boards and I am not sure if they have holes or not. I have been using AMD since the K6's came out so I will have to look into the other boards.
If necassary I will come up with a good clamp for boards without holes. I also noticed the some of the Tyan duellies don't have mounting holes.

Half of my Mill came today, the rest should be here by Friday!!!

[ 12-19-2001: Message edited by: CustomCooling ]

[simon]

I think price performance is a big problem in getting right. But Most of the waterblocks available are all based around a few central designs and I can't imagine why they are so expensive. Several sites seem to sell there own versions of common waterblocks for a similar price but they don't have to pay lots of money in the research area so why aren't they cheaper. Many sights do a maze type block but most are copies. I have decided to buy my own because the materials will cost me aroun £5/$8. How can people be charging £60???? I know my manufacturing isn't CNC but I will probably get a more jagged edge causing turbulance and increased surface area so, is CNC maching really worth the money??? Sorry for the rant but unless someone can detail all the costs involved then I won't be buying them.

[jaydee]

You bring up exellent points!!!! It actually will cost more than you think. You have to add in tooling, mounting hardware, solder(or what ever you use to attach the top and bottom), and time.

What most manufactuers are charging for though is either their overhead(cost of equipment, employees, bills, ect..) or they are just not selling enough blocks to keep going so they have to jack up the prices. Water cooling is not all that popular yet but it is getting better!!

I plan on removing alot of the variables as I have no overhead, as my equipment is bought and paid for, and I have no employes.

The cost for my blocks should be about 10-15% less than most AND I will make them in Aluminum and Copper, not just my best ones in Copper. Anyone seen a Maze 2 in Aluminum??? By using Aluminum the price will be a lot less and the performance will still be good enough. Especially if you use a pelt. For you that want the maximum I will still use Copper aswell. I plan on doing all my designs in both metals. If things work out the way I hope then it will not cost you much more to buy my blocks as it will to make one yourself.

AND YES, the CNC was worth it as it is not that expensive for the one needed to make small stuff like water blocks.

Originally posted by Skippy[the bush kangaroo]:
I think price performance is a big problem in getting right. But Most of the waterblocks available are all based around a few central designs and I can't imagine why they are so expensive. Several sites seem to sell there own versions of common waterblocks for a similar price but they don't have to pay lots of money in the research area so why aren't they cheaper. Many sights do a maze type block but most are copies. I have decided to buy my own because the materials will cost me aroun £5/$8. How can people be charging £60???? I know my manufacturing isn't CNC but I will probably get a more jagged edge causing turbulance and increased surface area so, is CNC maching really worth the money??? Sorry for the rant but unless someone can detail all the costs involved then I won't be buying them.

[Tweekster]

A good socket type clamp system for Athlons that use all 3 prongs.........most of the new 760MPX dualie mobo's coming out don't have mount holes, and I have a feeling that dualies and going to get real popular over the next year

[jaydee]

If that is the case I will have to come up with something as I plan on building a duely.

Originally posted by Tweekster:
A good socket type clamp system for Athlons that use all 3 prongs.........most of the new 760MPX dualie mobo's coming out don't have mount holes, and I have a feeling that dualies and going to get real popular over the next year

[deeznuts]

a guy named BillA, who wrote this excellent article: http://www.overclockers.com/articles481/

hypothesizes that the open style blocks, like the swiftech 462, are the best performing blocks. because you want absolutely fastest flow, and those provide it.

deez

[jaydee]

I will be testing these blocks against regular style blocks. My theory so far is that the more water you can push through that block the more cool will be transfered onto the block.

Some say it is better to have a slower flow so the water has more time to pick up the heat. But I think that is backwards thinking. I would think more flow would transfer the cold from the water to the block instead of the heat to the water.

Air colling you want to do it backwards by transfering the heat to the heat sink, but with water I think you want to do it backwards and transfer the cold to the block.

I will be testing all this in January once I get everything going.

Originally posted by deeznuts:
a guy named BillA, who wrote this excellent article: http://www.overclockers.com/articles481/

hypothesizes that the open style blocks, like the swiftech 462, are the best performing blocks. because you want absolutely fastest flow, and those provide it.

deez

[futRtrubL]

Cold is just the absence of heat. Heat is the disordered kinetic energy of atoms.

Anyway, people may think slower water will increase efficience but they are wrong. Each cc of water will carry more heat away but with faster flow there are more ccs. Also the water doesn't suddenly increase in temp as it exits the block, it heats up while its in there, greater water temp means less delta T (temp. difference, this is the ONLY thing that drives heat movement) so less "force" pushing/pulling the heat. Concider this the "Voltage" and the rate of heat flow is the "Current" the only thing left is the "Resistance" this is the C/W rating of a material, in this case the Cu or Al of the block (a function based on material conductivity, thickness and surface area). So for any heat transfer aplication you want high temp difference between block and water in contact with the sides of the block (cold water), and low resistance, so a conductive material (Ag better than Cu better than AL), short distance between channels and base, and lots of surface area in contact with water.

On a side note, turbulence can help, which is unused as water is relatively insulating, by getting this cool water in contact with the block.

Edward

[simon]

When you thick in terms of phyics you will find that as water gets hotter it will pick up heat less quickly. So if you have a higher flow rate you will get more of the cooler water passing over the block which will pick up the heat more effectively. I'm crap with graphs and stuff but if you imagine a time/heat of water graph the initail line will be very steep but will get less steep over time. In other words their is no piont keeping water that has been heated up already in the block when their is cooler water waiting to get in. I do appreciate that they may be cheaper than other places, and that I have to pay for your time rather than put mine in for free but I should be able to do my own for about £5-10 but the cheapest block i've seen in the Uk is £40. Design points, trying to increase turbulance will hopefully mix the water up more, turbulance not made bybumps in the base but the spirals you find in the barrel of a gun.(if you see what I mean, I've seen no proof this would be better, but in theory it sounds good).

[simon]

Sorry for reapeating what the other guy said, my page was open for ages so I didn't see it but thats what I was getting at too. I've heard people say the opposite to this is it is harder for the water to lose as much heat in the radiator, but their is less heat to lose so I think it should make no impact compared to that of benifits of higher flowrate.

[jaydee]

Exellent comments guys!!! I got the rst of my mill today and will get started on the test blocks after Christmas.

Any more input is welcome!!!

How much is £40 in US?

[Bollocks]

Well 40 pounds is like 120 australian dollars, so about 60 american bucks at current exchange rates.

I just need my new job to be able to afford a damn waterblock in the first place, got most other things sorted except for finding a radiator that will fit width-wise in my HX08

I would have to say that price DOES play a huge part in watercooling, especially when there isnt a huge demand yet.

Imagine that you could go and get a crappy hsf for 40 australian (20 american) or for 60 australian you could go pick up a waterblock, clamp and heatercore, which one would you pick?

I know those prices are really not too viable at the moment, but using things like aluminium blocks (because some of us dont want performance, we want silence and cheapness) would make this kind of cooling available to more people and increase demand.

Sure, produce hardcore blocks for the psycho's amongst us, but there does need to be some option for the poorer people or those that dont want to spend hundreds on getting a quieter computer.

I'm looking at an upgrade from my trusty celeron 466 @ 525, stock cooling, and while i would LOVE an athlon XP, i really will not buy one while the only option is to spend bucketloads of cash on buying a waterblock so that i dont have to run 46dB fans as the only other option.

The only other option then becomes another celeron (thanks but no thanks) or a P4, and yes lets bring on the bashing, but im sorry, silence is my one real reason for watercooling.

Ok silence, a little bit of an overclock, and for the cool factor given everyone i know poo-pooh's watercooling as stupid and over the top, even tho they have never tried it

[futRtrubL]

You said that you will be offering all designs in both Al and Cu flavors. What designs do you plan to offer? Like single/dual/no peltier setups and for which sockets?

Edward

[jaydee]

All the design that I will be making will be in both Copper and Aluminum. The designs I will end up with should contend with the top designs already out. Once I get the blocks milled out there will be pics of the inside design on my website along with my testing of them and independant reviews of them.

My testing will not be a Duron or a 1gig or less T-bird either like a lot of tech sites seem to be using. My test setup will be a Epox 8K7A+ and a T-Bird 1400. I will do both stock and overclock tests on every block. Being the 1400 is probably the hottest running CPU there is, that is what I will use.

Also I plan on making the regular Socket A blocks to start with that will use the holes around the socket. Once those are done I will move to the duel boards as I will have to come up with a hold down device if they won't have mounting holes. Then if there is a demand I will do soc370 and maybe P4.

I also will be doing chipset and GPU blocks.

Needless to say this will take some time to get going. I plan on having a sellable socket a cooler out by Mid January in both Aluminum and Copper.

$60us??? My blocks will not be that much!! I should have a good block in the lower $30's for the copper and the mid $20's for the Aluminum. And the lesser versions in the mid $20's for copper and $20 or less for the Aluminum. That's not set in stone but it looks that way as of now.

Originally posted by fut:
You said that you will be offering all designs in both Al and Cu flavors. What designs do you plan to offer? Like single/dual/no peltier setups and for which sockets?

Edward

[futRtrubL]

Fantastic! I'll be keeping an eye out for reviews. Be sure to post info here when you release them ;']

Edward

[Bollocks]

If you are planning on offering around $20 - $30 US for the blocks then you will most probably find a customer in me

For that kind of price i would definatelly jump on board.

Good luck and we definatelly want to keep hearing from you as you progress

[simon]

£40 is about $60, they are much cheaper from the states but then I have to pay higher delivery prices, import tax, sales tax, duty. Adds up to about the same price.

[Leuf]

Originally posted by Tweekster:
A good socket type clamp system for Athlons that use all 3 prongs.........most of the new 760MPX dualie mobo's coming out don't have mount holes, and I have a feeling that dualies and going to get real popular over the next year

The epox, asus, abit, and msi mpx boards all have the socket holes: http://www.epox.com/html/english/m762Aver2.jpg http://images.anandtech.com/reviews/...y2/asusmpx.jpg http://images.anandtech.com/reviews/...y2/abitmpx.jpg http://images.anandtech.com/reviews/...ay2/msimpx.jpg

The oddball of the group is the epox board that has the sockets at 90 degree angles to each other, which might cause probs if the clamps are too long.

[jaydee]

Cool!!! If the boards have holes then I won't need a clamp as you should be able to use any of my blocks on there. I am trying to stay away from the clamp as it would add to much cost for a decent one.

Originally posted by Leuf:
The epox, asus, abit, and msi mpx boards all have the socket holes: http://www.epox.com/html/english/m762Aver2.jpg http://images.anandtech.com/reviews/...y2/asusmpx.jpg http://images.anandtech.com/reviews/...y2/abitmpx.jpg http://images.anandtech.com/reviews/...ay2/msimpx.jpg

The oddball of the group is the epox board that has the sockets at 90 degree angles to each other, which might cause probs if the clamps are too long.

[jaydee]

I will start testing my first block today. Check my website for the pics. I am not going to show the inside design untill I know the results and have the final version done, but it is a maze style design with a 1/2" wide channel. If this one works out it will be around $20-$25 for the Aliminum and $30ish for the Copper. After this one is done I will start to work on the cheaper versions.

[jaydee]

Well thats interesting!!! how did my name get changed to jaydee116???? Oh well.