interactive waterblock comparator

[BillA]

some really fine work by BalefireX and Groth on procooling

select the wbs of interest, the units = see them
http://www.procooling.com/html/pro_testing.php

be cool

[slater3333uk]

Thanks for the heads up Bill :)

[ZachM]

Neat. That could come in handy.

[Joe Camel]

great start!!

the next page needs to be the "pump vs. WB" so you can estimate the flow your going to see with given pump/WB combo. then add Rad & tube restriction estimates and we have a loop simulator :) (that would help A LOT with all the "will this pump be better than..." or... "do i need to upgrade my WB or pump" type of questions.)

[RedViper]

Nice! :thup:

[Firtol88]

great start!!

the next page needs to be the "pump vs. WB" so you can estimate the flow your going to see with given pump/WB combo. then add Rad & tube restriction estimates and we have a loop simulator :) (that would help A LOT with all the "will this pump be better than..." or... "do i need to upgrade my WB or pump" type of questions.)


:eek: That would help a lot with overall WC adoption as many people have set up systems with unmatched components and are forever after vocal WC detractors.

[pHaestus]

Baby steps.... A lot of test data required to build any sort of simulator. ~7 months of effort tied into the graph Bill posted (and now the setup changes completely and I start over).

[rogerdugans]

Hmmm. Looks good, but doesn't seem to work on my linux machine using Firefox...
I'll have to try it out on one of the Windows machines.....

Edit- works great in IE. :)

Very nice job, guys- once again. :)

[pHaestus]

that's odd; I use firefox on Win32 and it is fine there.

[pelikan]

xp and firefox works.

[ZachM]

2K Pro and Firefox works fine here. You probably need a java plugin.

[magick_man]

lets just blame it on linux :)

~Magick_Man~

[rogerdugans]

Sorry about the erroneous post folks-
I just tried on another linux box with Firefox and the graph loaded up fine.

Kinda weird: the first linux box I tried it on is my main rig and never gives me any trouble with any website stuff not loading.....until now that is. ;)

My mistake. :)

*Edit- As far as blaming linux...blame it on my rank instead. :D

[pHaestus]

I can't see any reason (other than javascript being turned off maybe?) for it to render differently on firefox in Linux. I'll check out my Mandrake box today; it's running an old mozilla version atm I believe.

[Joe Camel]

Baby steps.... A lot of test data required to build any sort of simulator. ~7 months of effort tied into the graph Bill posted (and now the setup changes completely and I start over).


please dont think i was putting what you have done down...this is an awsome project!!

who's data is being used? Bill's, yours?...Cathar seems to have a lot of graphs posted, is his data a source of input?

[rogerdugans]

I can't see any reason (other than javascript being turned off maybe?) for it to render differently on firefox in Linux. I'll check out my Mandrake box today; it's running an old mozilla version atm I believe.

Hmmm: I just checked my other machines- the only one the comparator does NOT work on is this one.

There is no doubt in my mind that the problem is caused by something I messed up. ;)
As with all things, I constantly tweak, test and screw up various things messing about. :D

[pHaestus]

All the data being graphed is mine. No offense was taken; it IS the logical conclusion to the testing

[oc613]

awesome, just awesome work!

thanks for graph, i knew the results individually from the respective reviews, but seeing them all together is really just something else.

what i'm wondering though is if any of you know where my dtek tc4 rev. 2 would stack up at around 1 gpm. i have scowered high and low and could only find one site that said the block compares to the spir@l within .5 of a degree. so if any of the teseters out there can give me a rough estimate of the c/w in relation to the test conditions used for that graph i would be very. very thankful.

i know it would be next to impossible to determine that, but a *rough* estimate is all i need.

i was thinking about a new block and i guess i would see a 4 -5 c. improvement, perhaps more, would just like and opinion based on the experience you gurus of h2o has acquired after years of testing/tweaking.

details are in my sig, and thanks again for the wonderful work you guys put into the passion i now share also. (well using, not really reviewing)

:clap:

[bizhar]

Nice I like.

Works fine in Konqueror???
Only most KDE people will get that. LOL

[greenman100]

awesome, just awesome work!

thanks for graph, i knew the results individually from the respective reviews, but seeing them all together is really just something else.

what i'm wondering though is if any of you know where my dtek tc4 rev. 2 would stack up at around 1 gpm. i have scowered high and low and could only find one site that said the block compares to the spir@l within .5 of a degree. so if any of the teseters out there can give me a rough estimate of the c/w in relation to the test conditions used for that graph i would be very. very thankful.

i know it would be next to impossible to determine that, but a *rough* estimate is all i need.

i was thinking about a new block and i guess i would see a 4 -5 c. improvement, perhaps more, would just like and opinion based on the experience you gurus of h2o has acquired after years of testing/tweaking.

details are in my sig, and thanks again for the wonderful work you guys put into the passion i now share also. (well using, not really reviewing)

:clap:


I'd put the TC-4 rev2 about 4-6 C hotter than a Dtek WW.

Not sure though.

[pauldenton]

I'd put the TC-4 rev2 about 4-6 C hotter than a Dtek WW.

Not sure though.
billA tested it a while ago -
http://www.thermal-management-testing.com/bench_vs_system_testing.htm
in a set of data in which this block also appears..... so you should be able to guestimate from that....
http://www.procooling.com/reviews/html/innovatek_rev3_waterblock_revi.php

[greenman100]

billA tested it a while ago -
http://www.thermal-management-testing.com/bench_vs_system_testing.htm
in a set of data in which this block also appears..... so you should be able to guestimate from that....
http://www.procooling.com/reviews/html/innovatek_rev3_waterblock_revi.php


so I was about right?

[pauldenton]

so I was about right?

less i think -
at 1gpm pHaestus has the (dtek)WW as 2C better than the innovatec rev3, which billA has matching the TC-4 (at 1gpm).

looks like the WW gains a bit more with better flow, and the resistance of the TC-4 is very close to the WW, so i reckon 2-3C better depending on the pump....

[greenman100]

okay, I see.

too bad we don't have a better way to compare than across testbenches

good logic, however, and I am done with the threadjack

[pHaestus]

I'm sure I could get a TC4 from D-tek if it's of enough interest. I have a Maze3 I will eventually run too just for fun :)

[BillA]

I still have the TC4 I tested, after days of lapping
that was 2 test benches ago, hard to compare without a lot of 'corrections'
better to just test, you want it pH ?
- it is the original version

[oc613]

i'm definately interested :)

the thing about the tc4 is there isn't a "real" review that compares it to newer blocks under the same parameters. i know that d-tek is offering the blocks at a VERY compettitive price, so i'm sure a lot of people might be considering them, but are just a little afraid to take the *plunge* because they can't find data on the block. i got mine about a year ago and i know i was a bit reluctant.

i bought the tc4 because i'm, well, cheap. i scored the aluminum top for around 15 US new, when an online retailer that will go anonymous was clearing out all there water cooling gear (possibly because it wasn't selling?)

seeing as how i don't have much money invested in the block and upgrade just might be in order if i can shave off a decent amount of degrees. i have a silent 120 mm at around 30 cfm on an original procore and i am not willing to part with silence. recently i upgraded to a mobile barton, and as such have it clocked relatively high, thus i ahve experienced an increase in temps.

i am loading consistently at 50 celcius.

i reworked the flow order of my wc system, which necesited the lengthening of some tubes. i had hoped to gain a couple of degrees by placing the rad right after the pump, as opposed to the waterblock after the pump, but i have seen no difference in temps (~ 1.5 celcius) that can be accredited to the refinement as opposed to error in my crude measurements (read MBM5).

what i want to know is if a block upgrade could bring me down to say 43 - 44 load, them i would seriously consider it, and the only way to know that is to extrapolate it from the data you guys can collect on my current block.

it would be greatly appreciated by myself and i'm sure all current tc4 owners as well as perspective purchasers would be pleased also.

thanks again guys.

:attn:

[orionnt]

Im a noob at watercooling, can somebody please explain how exaclt to read this graph, how to know witch block is better?

[slater3333uk]

Ok. Well first off the graphs cant tell you which is 'best' They can only point you in the right direction. This is because they only show parts of the data needed to tell you which is the 'best' block for your given situation.

Basicly the deltaT (on the Y axis) is the temp of the cpu minus the ambiant temp (that of the air entering the radiator). This is done to eliminate the factor of abmiant tempreture in the testing.

As you can see the graph shows this DeltaT temp against Flow rate (that being the flowrate of water through the waterblock). As you can also see the DeltaT veries with flowrate. Its this relationship between the flowrate and deltaT tempreture which is of interest.

A very basic way to explain is the lower the line on the graph the better the waterblock is. (for example the cascade ss is the lowest and therfore gives you lower deltaT and therfore lower temps overall)

As you can see some of the lines cross which means some blocks will be better than others at low flowrates but at higher flow rates the situation may be reversed. A good example of this is the swiftech MCW6000a vs the LR white water. At a lower flowrate the swiftechs line is lower on the graph hence it works better but at a higher flowrate the white waters line is lower and hence the WW is better. This shows that its not a clear cut 'block A is better than block B'

However this graph is not the full story. The big problem is flowrate is in turn effected by the waterblock. Some waterblocks are more restrictive to water flow than others and therefore when used with the same pump some flow less water than others. This adds a whole new dimension to disiding which is best. As and example If waterblock A has a line lower on the graph than waterblock B then at first it looks like A would be better. BUT what at the same time waterblock A was more flow restrictive than waterblock B. This differance could mean that B is attualy better but without hard data we can relly only guesstimate.

At this time procooling cannot provide data on how restrictive each block is but some information on this is avalible at the oc.com front page.

I hope that helps you out a bit (and hope you understand) :)

[Etacovda]

Just to clarify;

Delta-T on the side of the graph is Cpu temp minus water temp at the block inlet, not ambient temp at the radiator... ie, if your water inlet is at 23 degrees C, and your block has a delta of 10 degrees at the flow rate that your pump gives, your cpu will be 33 degrees.

Ambient temps are not handy for this kind of testing. When you start talking ambient temps, pump temperature comes into play etc and it gets very messy. Using Cpu temp minus water temp is the cleanest way of removing variables.

[slater3333uk]

ahh yeh, my bad. I was in air cooling mode when i wrote that :)

[CAPiTA]

Nice! This is where being colour blind is really annoying. Is there a difference between the colour of the TDX and RBX line representation?

[SwampThing]

Color is the same for those two; TDX has diamonds, the RBX triangles.

[CAPiTA]

Good :) Maybe I'm not as colour blind as I though!

[pHaestus]

MCW6002 added tonight. Headed to bed now

[oc613]

any news on possible testing a tc-4?

thanks :)

[Jibjo]

This is a great graphing device. Eye-opening to me were the stats for the two Swiftech 6000 series vs the RBX and the Dtek Whitewater.

[JTanczos]

IDe like to see more noozles tested for the rbx. Not that it matters a terribly large amount since I just bought one but I was going to use #5 when your test only used #1. Ide like to know where it stands and im sure others would like to know in case the RBX matches the others better with different noozles.

JT



PS sorry if the post is seems off in any way. I am suffering from side effects of medication. Insomnia mixed with nausea is a wonderful thing.

[oc613]

any news on possibly testing a tc-4?

thanks :)



:p