Water block Writeup with the big dogs!!!! - Page 3

#Rotor · 08-09-2002, 03:58 PM

ok, so I imagine, you will be taking all the blocks, go through the motion with each one, and do all of them for result #1.

then restart from scratch, and do all of them again, in the same fashion, for #2....

and so forth, 5 times, I can pretty much think how much work that is going to be, but I can also see that you will pretty much have a very realistic representation of the performances. even with lesser than NASA approved equipment to test with.

pHaestus · 08-09-2002, 04:05 PM

Why throw out the low and the high? Aren't they relevant to the average? Throw out results if you believe you skewed them in your testing; the rest is relevant to estimations of variance in the numbers.

In fact, the std deviation from the mean (and %rel. error) turns out to be one of the things most useful to show when collecting data. Take those and make error bars on your points for the results. Make x and y error bars to account for flow rate error too and you'll start seeing why perpetual equipment upgrades are a reality for some.

#Rotor · 08-09-2002, 04:07 PM

there is a big difference between equipment being faulty, and equipment having a certain amount of tolerance, if the tolerance is grater than the difference between the blocks performances it would appear as if all the blocks are on par. with another, how can that cost someone a sale....

it would just mean that my block indeed will be held company by the BTMS.....or vise fersa errmm ya maybe you have a point..... heheh [joking... again]

if an error was to be made, the quality of the equipment it is going to be made with, will have no influence on that error.

in responce to the post below me

I was responding to your first post....

gone_fishin · 08-09-2002, 04:09 PM

Quote:

Originally posted by #Rotor
ok, so I imagine, you will be taking all the blocks, go through the motion with each one, and do all of them for result #1.

then restart from scratch, and do all of them again, in the same fashion, for #2....

and so forth, 5 times, I can pretty much think how much work that is going to be, but I can also see that you will pretty much have a very realistic representation of the performances. even with lesser than NASA approved equipment to test with.

I don't see how five errors could result in an accuracy.

#Rotor · 08-09-2002, 04:18 PM

well I have to say, and I'm not one to take error lightly, I can't afford to, but everything in the world around us, has error in them, even the test Labs at NASA has equipment with ever so slight an amount of "error" in them..... thinking that there is such a thing as 100% error free equipment out there, well there is no such a thing.....

even my spell checker has error in them.... see

BillA · 08-09-2002, 04:25 PM

ooohh, quick !
call the nurse with the big tits

gone_fishin has caught the BillA disease

gone_fishin · 08-09-2002, 04:26 PM

If you were to change your goals instead of saying this is the order, block A, B, C, ...... then you would have a more valid review. If you were to say for example that your goals are to find a group of blocks that SEEM to perform up to standards where as this other group of blocks do not SEEM to perform up to standards then you would have results more closely matching your abillities to test. You could make your review also focus on the physical aspects of all the blocks, how they mechanically work, how they mount, how flat is the base, how they look etc, etc. This would be a very interesting review then instead of the piss-ant types that try to declare a winner that we are used to seeing.

#Rotor · 08-09-2002, 04:37 PM

that would indeed be a wise thing to do, given the fact that all the blocks in his lineup did come out to perform exactly on par. Which, if that was the case, I'm sure would have triggered some little beeper in WebMedics mind as to maybe getting some more accurate measuring tools. though how would one go about upgrading the on-die thermal diode inside the CPU... we are still limited by what AMD is putting in there, not so?

[edit]
and yes, on the aspect of how they function, there I do agree, especially the retention mechanisms, it should be part and parcel of the blocks, test. I mean if a block can not even hold itself effectively in place to render its service to the CPU, it should not be deemed a high performing block, even if the block itself has the ability to outperform all others....[/edit]

gone_fishin · 08-09-2002, 04:43 PM

Quote:

Originally posted by unregistered
ooohh, quick !
call the nurse with the big tits

gone_fishin has caught the BillA disease

Hey, do you got one of them

pHaestus · 08-09-2002, 04:44 PM

G_F presents a point that should be at the forefront of everyone's mind when testing:

"What are the goals?"

BillA, webmedic, and my goals are all somewhat different and demand varying precision (and expense accounts).

Regarding simply increasing sample size to produce the needed accuracy:

Not likely. Without good measurements then no number of replicates will help. As mentioned you can probably separate "crappy" from "good" as long as you are only making 2 piles but no better.

The most useful thing you could do with a typical roundup would be to take a lot of pictures and detail the mounting/remounting of the block. These blocks for the most part do perform pretty similarly, so ease of use and bang for buck are valid things to use in comparison.

webmedic · 08-09-2002, 05:09 PM

Not sure what has you all fired all of a sudden and I don't remember asking you for a block (gone fishin). I am atemting to show two different things here.

First there is a difference in the equipment that billa has and home users have. Agreed? If this is so they will never get the results that bill does. The first part of my testing is to show more or less with the same equipment that home users have what kind of results they are likely to see. The margin for error is not entirely but partially an issue as it will be a real world analysis. As a result there will be more of a tempature variation in block design because the pumps billa uses are far beyond any pump you will see used in a computer water cooling. The expense is enough to go refridgerated.

Second I will build the setup like bill has I've already started with some of the parts but don't have them all yet. The results gathered from this will be of more use to those that design blocks. Why because it will be able to show weeknesses in thier design. However as noted above no home user will be able to produce these results because they will not be using a pump that flows 11.4gpm with a 66 foot head. By the way that is the actual flow of a real pump. Bill is using something simmilar but I don't think he is using that exact one.

For me me reasons are strictly informational. I want to know. I never liked hype. I have to know. As a result the testing will have to be done so that I know. My intention is not to lead anyone astray and also I see the need for both types of testing.

Note: fixed an error with the gpm of the pump it now reads correctly.

webmedic · 08-09-2002, 05:39 PM

Quote:

Originally posted by gone_fishin
If you were to change your goals instead of saying this is the order, block A, B, C, ...... then you would have a more valid review. If you were to say for example that your goals are to find a group of blocks that SEEM to perform up to standards where as this other group of blocks do not SEEM to perform up to standards then you would have results more closely matching your abillities to test. You could make your review also focus on the physical aspects of all the blocks, how they mechanically work, how they mount, how flat is the base, how they look etc, etc. This would be a very interesting review then instead of the piss-ant types that try to declare a winner that we are used to seeing.

Sorry I was gone for the big fight. I was watching all day and just when I turn my back look what you kids are up to.

Ok in the final part I will be rating the blocks against each other and at that time they will be ranked on different merrits. Mounting mechanism, fit, finnish, construction, cost, performance, and wheather or not it helps ocability. These are examples of what could be covered. You see the whole reason I've posted the test procedure is to ask how it should be modified to better test. How I should do this or that. This is all open for debate not set in stone.

What whould any of you like to see them rated on.

Ok phman I'll keep all 5 and show the margin for error for each set but If I get something way out I will through out all data and redoo it. Is that better?

#rotor Ok since It will not be as exact as wanted. I will use the mounting that comes with the blocks. If it makes them perform badly thats their fault and they will recieve bad marks for it. Can;t wait to ge tmy hands on a btms. I'm not sutre what will happen but it should be interesting.

Come on guys this is what I wanted or needed but lets keep it constructive. I cant just stop with it now. There are to many things involved so the first round will be for the consumer but I still need to know if I should make canges to my testing procedure. Any responses welcome.

gone_fishin · 08-09-2002, 05:49 PM

Throwing out five bads to get five new bads will still not get you close to anything accurate for individual block analysis.

pHaestus · 08-09-2002, 05:51 PM

No but it keeps the error bars suitably wide to demonstrate that the testing procedures aren't exact

webmedic · 08-09-2002, 05:54 PM

Quote:

Originally posted by gone_fishin
Throwing out five bads to get five new bads will still not get you close to anything accurate for individual block analysis.

You know you have a problem. Is that I never asked for one of your blocks? I'm not sure but this was not inteded to set you off. Read a couple posts above and YOU WILL SEE NO HOME USER CAN REPRODUCE BILLS, PH, or MY FINAL RESULTS. THEY WILL SIMPLY NOT HAVE THE EQUIPMENT TO DO SO. They will not be able to get the flow rates bill tested at because there pump will not do it.

gone_fishin · 08-09-2002, 05:59 PM

I'm not really set off but I see the truth has kind of set you off

State your goals to match your honestly achievable results. What does the man who really designed your block think? Perhaps he could be invited here to give his view?

BillA · 08-09-2002, 06:03 PM

slow down there webmedic

3gpm is quite attainable with available pumps and wbs
while I do collect some data beyond 3gpm, it is not 'published'
(as it is not relevant for virtually all users)

be cool

webmedic · 08-09-2002, 06:07 PM

Oh man you don't want to know. Hes been talking to his prof that teaches fluid dynamics and the prof said the equipment we are using is within 2% margin of error and this whole discussion is rediculus. He belived so before talking to the prof. I'll get him on here for you. If you like but I've been holding him back because I see the need for both types of testing and thinks its overkill.

You just dont get it do you. You will not get bills results in a normal water cooled compter. Not because of the margin of error but because of the equipment he uses. I have never went back on anything. I believe this both types of testing are benifical. I'm doing this first and when I have all the parts collected will do the next phase. Who else is even trying to collect these parts and do that? I only know of three people. Are you one of them? I don't think so.

webmedic · 08-09-2002, 06:09 PM

Yes bill but it is much more dificult to atain on some water blocks with very restrictive flow.

gone_fishin · 08-09-2002, 06:09 PM

No, I am not. But one who does should state their goals to match their abillities, that is all I am saying

webmedic · 08-09-2002, 06:15 PM

Ok I already said in my writeup that the results are what they are. Did you read it? I'm not making this out to be anything it is not. I understand the limitations of the equipment used. I posted this on the testing methods paper I posted. If you see an error let me know so that I may make it more clear what this test is to achive.

ps: I'm not afer world domination. I just wanted to do a water block shootout. So how can I make it more clear if I stated something to the contrary point it out so that I can correct it.

the article is here if you missed it:

http://www.water-cool.com/modules.ph...owcontent&id=3

gone_fishin · 08-09-2002, 06:24 PM

As I stated before, the best you could come up with is two camps or groupings of blocks. One that seems to perform good and one that does not seem to perform good, given your testing abillities

BillA · 08-09-2002, 06:26 PM

g_f
those with an interest should be encouraged

webmedic has several areas in which he needs some practical experience,
and believe me, he is going to get mega-helpings of such

no biggie, lets be critical in a constructive sense - and see how it plays out

I started this kind of testing a bit over 2 years ago, and I learn a bit more every single day

I too think the present tack has some shortfalls - but the results will speak more directly to that than any 'pronouncements'

go to work webmedic

EDIT: a point of clarification
"You will not get bills results in a normal water cooled compter."
not so, the wbs performance is quite independant of the measurement 'system' (or should be)
it will perform as it is able to perform
the sole question is our capability to accurately measure how it is performing

pHaestus · 08-09-2002, 10:01 PM

I can (by and large) reproduce the order of JoeC's heatsink comparative ranking table with my diode reader. In fact, I e-mailed him about one heatsink that seemed anomolous and he said it was a typo. Now in absolute terms I can't get the same C/W (his are a bit lower), but the comparative rankings I can reproduce. This should be the same for me and anyone with a decent testbed and experience.

Waterblocks may be harder though; it's difficult to totally isolate all the variables and get numbers that can even be compared.

BillA · 08-09-2002, 11:08 PM

VERY interesting pHaestus

question:
are ALL diode readers going to produce the same results ?

if not,
how should the diode reader be described (speced) to indicate/assure its accuracy ?

08-09-2002, 03:58 PM	#51
#Rotor Cooling Savant Join Date: Feb 2002 Location: Dione, sector 4s1256 Posts: 852	ok, so I imagine, you will be taking all the blocks, go through the motion with each one, and do all of them for result #1. then restart from scratch, and do all of them again, in the same fashion, for #2.... and so forth, 5 times, I can pretty much think how much work that is going to be, but I can also see that you will pretty much have a very realistic representation of the performances. even with lesser than NASA approved equipment to test with. __________________ There is no Spoon....

08-09-2002, 04:07 PM	#53
#Rotor Cooling Savant Join Date: Feb 2002 Location: Dione, sector 4s1256 Posts: 852	there is a big difference between equipment being faulty, and equipment having a certain amount of tolerance, if the tolerance is grater than the difference between the blocks performances it would appear as if all the blocks are on par. with another, how can that cost someone a sale.... it would just mean that my block indeed will be held company by the BTMS.....or vise fersa errmm ya maybe you have a point..... heheh [joking... again] if an error was to be made, the quality of the equipment it is going to be made with, will have no influence on that error. in responce to the post below me I was responding to your first post.... __________________ There is no Spoon.... Last edited by #Rotor; 08-09-2002 at 04:10 PM.

08-09-2002, 04:18 PM	#55
#Rotor Cooling Savant Join Date: Feb 2002 Location: Dione, sector 4s1256 Posts: 852	well I have to say, and I'm not one to take error lightly, I can't afford to, but everything in the world around us, has error in them, even the test Labs at NASA has equipment with ever so slight an amount of "error" in them..... thinking that there is such a thing as 100% error free equipment out there, well there is no such a thing..... even my spell checker has error in them.... see __________________ There is no Spoon....

08-09-2002, 04:37 PM	#58
#Rotor Cooling Savant Join Date: Feb 2002 Location: Dione, sector 4s1256 Posts: 852	that would indeed be a wise thing to do, given the fact that all the blocks in his lineup did come out to perform exactly on par. Which, if that was the case, I'm sure would have triggered some little beeper in WebMedics mind as to maybe getting some more accurate measuring tools. though how would one go about upgrading the on-die thermal diode inside the CPU... we are still limited by what AMD is putting in there, not so? [edit] and yes, on the aspect of how they function, there I do agree, especially the retention mechanisms, it should be part and parcel of the blocks, test. I mean if a block can not even hold itself effectively in place to render its service to the CPU, it should not be deemed a high performing block, even if the block itself has the ability to outperform all others....[/edit] __________________ There is no Spoon.... Last edited by #Rotor; 08-09-2002 at 04:42 PM.

08-09-2002, 05:09 PM	#61
webmedic Cooling Savant Join Date: Dec 2001 Location: Cheney, Wa Posts: 367	Not sure what has you all fired all of a sudden and I don't remember asking you for a block (gone fishin). I am atemting to show two different things here. First there is a difference in the equipment that billa has and home users have. Agreed? If this is so they will never get the results that bill does. The first part of my testing is to show more or less with the same equipment that home users have what kind of results they are likely to see. The margin for error is not entirely but partially an issue as it will be a real world analysis. As a result there will be more of a tempature variation in block design because the pumps billa uses are far beyond any pump you will see used in a computer water cooling. The expense is enough to go refridgerated. Second I will build the setup like bill has I've already started with some of the parts but don't have them all yet. The results gathered from this will be of more use to those that design blocks. Why because it will be able to show weeknesses in thier design. However as noted above no home user will be able to produce these results because they will not be using a pump that flows 11.4gpm with a 66 foot head. By the way that is the actual flow of a real pump. Bill is using something simmilar but I don't think he is using that exact one. For me me reasons are strictly informational. I want to know. I never liked hype. I have to know. As a result the testing will have to be done so that I know. My intention is not to lead anyone astray and also I see the need for both types of testing. Note: fixed an error with the gpm of the pump it now reads correctly. __________________ www.water-cool.com Last edited by webmedic; 08-09-2002 at 05:42 PM.

08-09-2002, 04:05 PM	#52
pHaestus Big Player Making Big Money Join Date: Aug 2001 Location: irc.lostgeek.com #procooling.com Posts: 4,782	Why throw out the low and the high? Aren't they relevant to the average? Throw out results if you believe you skewed them in your testing; the rest is relevant to estimations of variance in the numbers. In fact, the std deviation from the mean (and %rel. error) turns out to be one of the things most useful to show when collecting data. Take those and make error bars on your points for the results. Make x and y error bars to account for flow rate error too and you'll start seeing why perpetual equipment upgrades are a reality for some.

08-09-2002, 04:25 PM	#56
BillA CoolingWorks Tech Guy Formerly "Unregistered" Join Date: Dec 2000 Location: Posts: 2,371.493,106 Posts: 4,440	ooohh, quick ! call the nurse with the big tits gone_fishin has caught the BillA disease

08-09-2002, 04:26 PM	#57
gone_fishin Cooling Savant Join Date: May 2002 Location: Da UP Posts: 517	If you were to change your goals instead of saying this is the order, block A, B, C, ...... then you would have a more valid review. If you were to say for example that your goals are to find a group of blocks that SEEM to perform up to standards where as this other group of blocks do not SEEM to perform up to standards then you would have results more closely matching your abillities to test. You could make your review also focus on the physical aspects of all the blocks, how they mechanically work, how they mount, how flat is the base, how they look etc, etc. This would be a very interesting review then instead of the piss-ant types that try to declare a winner that we are used to seeing.

08-09-2002, 04:44 PM	#60
pHaestus Big Player Making Big Money Join Date: Aug 2001 Location: irc.lostgeek.com #procooling.com Posts: 4,782	G_F presents a point that should be at the forefront of everyone's mind when testing: "What are the goals?" BillA, webmedic, and my goals are all somewhat different and demand varying precision (and expense accounts). Regarding simply increasing sample size to produce the needed accuracy: Not likely. Without good measurements then no number of replicates will help. As mentioned you can probably separate "crappy" from "good" as long as you are only making 2 piles but no better. The most useful thing you could do with a typical roundup would be to take a lot of pictures and detail the mounting/remounting of the block. These blocks for the most part do perform pretty similarly, so ease of use and bang for buck are valid things to use in comparison.

08-09-2002, 05:49 PM	#63
gone_fishin Cooling Savant Join Date: May 2002 Location: Da UP Posts: 517	Throwing out five bads to get five new bads will still not get you close to anything accurate for individual block analysis.

08-09-2002, 05:51 PM	#64
pHaestus Big Player Making Big Money Join Date: Aug 2001 Location: irc.lostgeek.com #procooling.com Posts: 4,782	No but it keeps the error bars suitably wide to demonstrate that the testing procedures aren't exact

08-09-2002, 05:59 PM	#66
gone_fishin Cooling Savant Join Date: May 2002 Location: Da UP Posts: 517	I'm not really set off but I see the truth has kind of set you off State your goals to match your honestly achievable results. What does the man who really designed your block think? Perhaps he could be invited here to give his view?

08-09-2002, 06:03 PM	#67
BillA CoolingWorks Tech Guy Formerly "Unregistered" Join Date: Dec 2000 Location: Posts: 2,371.493,106 Posts: 4,440	slow down there webmedic 3gpm is quite attainable with available pumps and wbs while I do collect some data beyond 3gpm, it is not 'published' (as it is not relevant for virtually all users) be cool

08-09-2002, 06:07 PM	#68
webmedic Cooling Savant Join Date: Dec 2001 Location: Cheney, Wa Posts: 367	Oh man you don't want to know. Hes been talking to his prof that teaches fluid dynamics and the prof said the equipment we are using is within 2% margin of error and this whole discussion is rediculus. He belived so before talking to the prof. I'll get him on here for you. If you like but I've been holding him back because I see the need for both types of testing and thinks its overkill. You just dont get it do you. You will not get bills results in a normal water cooled compter. Not because of the margin of error but because of the equipment he uses. I have never went back on anything. I believe this both types of testing are benifical. I'm doing this first and when I have all the parts collected will do the next phase. Who else is even trying to collect these parts and do that? I only know of three people. Are you one of them? I don't think so. __________________ www.water-cool.com Last edited by webmedic; 08-09-2002 at 06:10 PM.

08-09-2002, 06:09 PM	#69
webmedic Cooling Savant Join Date: Dec 2001 Location: Cheney, Wa Posts: 367	Yes bill but it is much more dificult to atain on some water blocks with very restrictive flow. __________________ www.water-cool.com

08-09-2002, 06:15 PM	#71
webmedic Cooling Savant Join Date: Dec 2001 Location: Cheney, Wa Posts: 367	Ok I already said in my writeup that the results are what they are. Did you read it? I'm not making this out to be anything it is not. I understand the limitations of the equipment used. I posted this on the testing methods paper I posted. If you see an error let me know so that I may make it more clear what this test is to achive. ps: I'm not afer world domination. I just wanted to do a water block shootout. So how can I make it more clear if I stated something to the contrary point it out so that I can correct it. the article is here if you missed it: http://www.water-cool.com/modules.ph...owcontent&id=3 __________________ www.water-cool.com Last edited by webmedic; 08-09-2002 at 06:17 PM.

08-09-2002, 06:24 PM	#72
gone_fishin Cooling Savant Join Date: May 2002 Location: Da UP Posts: 517	As I stated before, the best you could come up with is two camps or groupings of blocks. One that seems to perform good and one that does not seem to perform good, given your testing abillities

08-09-2002, 06:26 PM	#73
BillA CoolingWorks Tech Guy Formerly "Unregistered" Join Date: Dec 2000 Location: Posts: 2,371.493,106 Posts: 4,440	g_f those with an interest should be encouraged webmedic has several areas in which he needs some practical experience, and believe me, he is going to get mega-helpings of such no biggie, lets be critical in a constructive sense - and see how it plays out I started this kind of testing a bit over 2 years ago, and I learn a bit more every single day I too think the present tack has some shortfalls - but the results will speak more directly to that than any 'pronouncements' go to work webmedic EDIT: a point of clarification "You will not get bills results in a normal water cooled compter." not so, the wbs performance is quite independant of the measurement 'system' (or should be) it will perform as it is able to perform the sole question is our capability to accurately measure how it is performing Last edited by BillA; 08-09-2002 at 06:52 PM.

08-09-2002, 10:01 PM	#74
pHaestus Big Player Making Big Money Join Date: Aug 2001 Location: irc.lostgeek.com #procooling.com Posts: 4,782	I can (by and large) reproduce the order of JoeC's heatsink comparative ranking table with my diode reader. In fact, I e-mailed him about one heatsink that seemed anomolous and he said it was a typo. Now in absolute terms I can't get the same C/W (his are a bit lower), but the comparative rankings I can reproduce. This should be the same for me and anyone with a decent testbed and experience. Waterblocks may be harder though; it's difficult to totally isolate all the variables and get numbers that can even be compared.

08-09-2002, 11:08 PM	#75
BillA CoolingWorks Tech Guy Formerly "Unregistered" Join Date: Dec 2000 Location: Posts: 2,371.493,106 Posts: 4,440	VERY interesting pHaestus question: are ALL diode readers going to produce the same results ? if not, how should the diode reader be described (speced) to indicate/assure its accuracy ?

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)