How much would you spend, on test your waterblock?

[nicozeg]

I voted for $0, I know is not an easy task, but the concept of DIYers paying someone to judge their job is completely flawed. It's totally against the Do It Yourself idea.

For example look at BillA tests at oc.com. How many DIY blocks were tested?

-3, two form Owen and one from Hoot, and I seriously doubt they paid the full $80 rate. All the others were comercial tests.

An good question here is: What is testing usefull for?

If you're trying to make the perfect block, you need to make lots of tests whith different variations. It's not practical nor economical to wait for shipping, testing and return on each variation. So you need to have a way of testing by yourself. Only when you reach your perfect block, and want to show all the world how good it is, you need an external test.

But the problem is: What DIYer is going to spend $100 on satisfying his ego? Not very much.

Only if you are interested on selling copies of your block it can be usefull, but that makes you another category than the typical DIY: The WBM, or Wanna Be Manufacturer.

So if DIY tests should be free, the problem is who dares to do the task? Just imagine the flood of blocks that pH is going to have if he announces: Free tests at ProCooling!

Even in that case I doubt I'm going to send mine for the test; I don't have a replacement for the test period, nor the time to make a copy, and completely refuse to switch back to aircooling. :shrug: But I can imagine some bored noob sending each week his new invention

But a good idea could be a monthly test Competition: People apply their blocks to be tested with a couple pictures and a brief description; Participants are displayed somewere on the site; Users or a jury vote for the winner, and he can send his block for a full test that will be news on the main page.

That can make a very nice site feature.

[nicozeg]

WOw its easy? how come someone didnt tell me this sooner?! Its always been such a bitch every time I did testing like this... if only someone woulda told me it was easy I wouldnt have had to waste all that time.

I mean exactly the opposite, I know it's not easy.

I don't smoke crack, just normally use spanish languaje, and my english grammar can be weird. At least with spanish, if you start the phrase with not, it means the opposite.

I'm going to edit it for more precicion.

By the way, if you read the complete post you could get the idea.

[Joe]

hehe I deleted my post, I realized I read it wrong. Yeh I figured it was some language barrier thing.

[jaydee]

Originally posted by nicozeg


An good question here is: What is testing usefull for?

Simple answer. It shows how a certain design works. If the results look good a DIY'er might want to use that design or make a variation of it on his own. If the results suck he won't. Isn't that the whole point on this water block construction forum? Designing, building, and advancing block design? Hard to do if your results are +/- 10C eh?

[nicozeg]

Originally posted by jaydee116
Simple answer. It shows how a certain design works. If the results look good a DIY'er might want to use that design or make a variation of it on his own. If the results suck he won't. Isn't that the whole point on this water block construction forum? Designing, building, and advancing block design? Hard to do if your results are +/- 10C eh?

Fine, I can imagine cathar sending each revision of WW to bill, paying $80 plus Shipping arround the globe, and waiting for results. It would take years and tons of money to develop.

[bigben2k]

Maybe this is slightly misconstrued, and that may be my own fault...

This test procedure (and tools) is for a DIY'er to test his own block, himself, to prove that a design revision is actually better.

The reason we're going through this exercise, is to point out that the normal system test is not repeatable, and therefore, of little value. The results are not transferable to someone else.

[jaydee]

Little disapointed at the replies at overclockers.com but certainly not unexpected. Buy a $9 compunurse to get within 10C of reality eh? Hummm, how would you really know your block sucked if your temps are -+10C? Could be a costly mistake paying Cathar $120 or so when what he gives you may not be a lot better than what you have.

As for AMDMB, well certainly not suprized to see all but one reply from members here. Most the people there are one;s that ask "What should I get?". Oblivious to what testing even is.

Anyway I don't think most people care about testing period. As long as they have a number that "visually" looks good then they have no problem bragging about it or believing it. Of course those same people can't figure out why they can't overclock and bitch and moan because "MBM says it good so it has to be something else" and they piss away 100's of dollars in new mobos and ram to be "leet" overclockers when all they really needed to so was drop the temps.

Bah, screw it.....

[bigben2k]

Well then!

Obviously not everyone is interested in testing their DIY properly: we can't help them!

But for those who do, I think we can put something together. I'm still waiting for Crane's write up.

Judging from the responses, there seems to be a split between the reasonable, and the unreasonable . I'll leave pHaestus, Groth, since87 and others tackle that last one: they're already way ahead of everyone.

[jaydee]

Originally posted by bigben2k
Well then!

Obviously not everyone is interested in testing their DIY properly: we can't help them!

I still would like to see an article written up on what tests sould be measured and then a addon article on procedure and tools. Something we can link the "leet temp kiddies" when they spew there BS temps around. Geting tired of hearing all these temps 2C above ambient from the "leet" air/water coolers.

I am getting a new mobo here soon and will be cutting the socket and putting the thermocoupler I just got on the bottom of the CPU. That will get me going untill I can get the die sim done.

[jaydee]

Originally posted by jaydee116


I am getting a new mobo here soon and will be cutting the socket and putting the thermocoupler I just got on the bottom of the CPU. That will get me going untill I can get the die sim done.

MB VIA KT400|6P1A4D KD7 ABIT ATX 1 $39.00
POWER SUPPLY 300W ALLIED AL-A300ATX 1 $18.99

Subtotal $57.99
Tax $.00
Shipping and Handling charge
6.00
Total $63.99


Step 5 of 5

Hehehe. That Abit KD7 sounds like a sweet board. 3 phase Power, Vcore up to 2.35V, nice NB Orbish fan, ect... Couldn't pass it up for $39.00.

[Cathar]

Originally posted by jaydee116
Could be a costly mistake paying Cathar $120 or so when what he gives you may not be a lot better than what you have.

I'd never take money to test a block for someone. I don't deserve that right.

I do dispute the assertion that the relative readings are each ±10C though. Absolute temperatures, for sure, but not relative.

[jaydee]

Originally posted by Cathar
I'd never take money to test a block for someone. I don't deserve that right.

I do dispute the assertion that the relative readings are each ±10C though. Absolute temperatures, for sure, but not relative.

I think he ment he would give you $120 for one of your blocks if his didn't work the way he thought it should.

I agree about relativity, but it does the guy with a 55C temp reading no good if his temps are really 45C. He might buy your block and the temps drop to 52C and he is pissed because he just dropped $120 for a 3C gain and his temps are still way higher than he thinks they should be all because of a 10C off temp probe. He could have just got a decent thermocoupler for $30 and saved $90. Relativity is only usefull for the one's that understand it. Of course maybe we are not worried about them here.

[babyeater]

I am looking for relative temps as when i add a new component to my system i want to know is it performing better or worse i think thats all the DIY'er needs to know . To see if he is going the right direction in improving his temperatures.It would be nice to know his actual temps aswell but that would require a good probe.
What i understand from this thread is that if i buy a compunurse and it is 10C off it will still give me relative temps i.e. my new block performs 1C better than my last. Are my right in this or will the compunurse not be accurate enough for this 1C difference.

[jaydee]

Originally posted by babyeater

What i understand from this thread is that if i buy a compunurse and it is 10C off it will still give me relative temps i.e. my new block performs 1C better than my last. Are my right in this or will the compunurse not be accurate enough for this 1C difference.

Not realistically. They very everytime you warm them up by 1-5C from what I have seen from users of them. One day they might read 32C and two days later under identical enviroment they might read 37C. Better off just using the mobo's probe. They seem to be more consistant (providing they are actually working) as long as you don't allow any airflow around them. If your going to half ass it there is not reason to buy anything IMO. You "should" be able to make a graph and plot your probes linarity. In other words take reading at various overclock's and voltages and plot it. Check this thread out for what I mean: http://forums.procooling.com/vbb/sho...&threadid=6388

Problem with this is stated in that thread but mainly you really don't know what the true wattage output of the CPU is.

[jaydee]

I got my KD7 today. Looks as if there is no onboard probe inside the socket area so it must read from the one on the CPU??? It has a WindBond chip on it aswell. If memory serves thats what controles the temp reading???

Anyway I will be hooking this up tonight. Will see what the TC reads against whatever the board says. Sweet looking board...

[Groth]

The Winbond chip in the KD7 reads the system and socket thermistor temps. Yep, there's tiny little surface mount thermistor in there; doesn't that sound accurate?

Thermal diode temperature and another system temp come from a Nat. Semi. LM90. Happy calibration!

[jaydee]

Originally posted by Groth
The Winbond chip in the KD7 reads the system and socket thermistor temps. Yep, there's tiny little surface mount thermistor in there; doesn't that sound accurate?

Thermal diode temperature and another system temp come from a Nat. Semi. LM90. Happy calibration!

Thanks. I think I see the little thermistor in there. Interesting. That LM90 is +/-3C!

[Groth]

That LM90 is +/-3C!

Then, add another +/- 2, for the manufacturing tolerance on the CPU. Then move that range, depending on whether the nominal ideality factor of your CPU is the same as the LM90's assumed 1.008. But the temps will at least be linear.

Most of that is systematic error - calibrating the reader/CPU should give you +/- 1° C, or a hair better.

[jaydee]

I just got back from WallMart and they have a multimeter with a type K thermocoupler for $17.00.... Will test this against the other one I bought and use them both. One for water temp and one on the backside of the CPU. WIll calibrate them and see how far off they are from each other. Distilled water for $.58 a gallon to.

[liquidcooler]

Check there to see how this guy tests coolers: http://www.dansdata.com/coolercomp.htm










http://www.losconectados.com

[jaydee]

Originally posted by liquidcooler
Check there to see how this guy tests coolers: http://www.dansdata.com/coolercomp.htm

That is interesting, but has not much to do with what we are trying to accomplish here.

[hydrogen18]

am I the only one who has noticed the topic here has a grammatical error?

[jaydee]

Originally posted by hydrogen18
am I the only one who has noticed the topic here has a grammatical error?

No, I just been ignoring it.

[bigben2k]

Corrected, again... thank you!

[Pears0]

I vote nothin, im a poor university student in the uk, i can barly afford to build my watercooling systems, yet alone pay for them to b tested!