Best Combination of Blocks?

[HiJon89]

I'm going to put an MCW5002-64T w/226W Pelt on my CPU, and my loop will be MD-20RZ -> 2-302 Core Weapon modded -> MCW5002-64T -> GPU Block?? -> MD-20RZ. For the GPU Block I'm deciding between a Maze 4-1 , or an MCW50-T, either will have a 172W pelt. Which of these two blocks would work best in my loop?

[CrimeDog]

it really doesn't matter, especially with tecs

i think that heatercore is gonna get overwhelmed though

[HiJon89]

Well I think with TEC's maximum flow is more important than block design and I know the Maze 4 is super low restriction so I think I'm just going to go with that :)

[MaxxxRacer]

go with the maze4. the mcw50T is aluminum so its best to stay away from it. sadly the cpu block is still aluminum so you're gonna need a good amount of zerex in there.

as to the heatercore.. its gonna get pretty warm. to keep decent water temps your gonna need a good shroud and some delta HE's on that thing. the 140cfm delta Tri blades will work too.

[HiJon89]

go with the maze4. the mcw50T is aluminum so its best to stay away from it. sadly the cpu block is still aluminum so you're gonna need a good amount of zerex in there.

as to the heatercore.. its gonna get pretty warm. to keep decent water temps your gonna need a good shroud and some delta HE's on that thing. the 140cfm delta Tri blades will work too.
Its aluminum???? :mad: I had no idea, well that sucks.

[MaxxxRacer]

yep. both of those swiftech blocks are aluminum. actually with that said the maze4 cpu block would probably bea better option aswell. just get a brass top for it.

[HiJon89]

yep. both of those swiftech blocks are aluminum. actually with that said the maze4 cpu block would probably bea better option aswell. just get a brass top for it.
Alright, I'm going to get the Maze 4-1 CPU block with brass top and 226W pelt and team it up with the Maze 4-1 GPU block with acetal top, thanks for the advice as always Maxxx :D

[MaxxxRacer]

no prob. remember to put some goood sized fans on that radiator.

[HiJon89]

no prob. remember to put some goood sized fans on that radiator.
Right now I have 2 120x38mm Sanyo Denkis, would it be worth it to add a shroud on the other side and add 2 more fans?

[MaxxxRacer]

got a picture or if not a description of how the fans are connected to the radiator. in any case if ur not using atleast some sort of shroud it would behove you to do so.

[HiJon89]

got a picture or if not a description of how the fans are connected to the radiator. in any case if ur not using atleast some sort of shroud it would behove you to do so.
Using a weapon shroud, fans are pulling.

[EnJoY]

got a picture or if not a description of how the fans are connected to the radiator. in any case if ur not using atleast some sort of shroud it would behove you to do so.


Behove eh Maxxx? Using fancy words are we now that we're a big watercool testing big shot? :p: :D

[MaxxxRacer]

lol enjoy. its just the correct word to use in this particular situation.

hijohn, no reason to upgrade then. u got the best shroud u cna get :p:

[HiJon89]

Is the Maze 4 CPU Block low restriction like the Maze 4 GPU Block?

[MaxxxRacer]

yah its very low restriction. you should mate it up with a pump that goes for high flow vs high pressure. especially considering ur gonna be using a weapon heatercore which is one of hte least restrictive rads on the market.

[HiJon89]

yah its very low restriction. you should mate it up with a pump that goes for high flow vs high pressure. especially considering ur gonna be using a weapon heatercore which is one of hte least restrictive rads on the market.
Pump that goes for high flow eh? How does my MD-20RZ sound :p:

[MaxxxRacer]

ironicly ud do ALOT better with a MD20RLT... the Z is not suited for your setup. but u should get around 2.3gpm. 2.5gpm is the max for ur pump.

[CrimeDog]

i'm not sure why you think these blocks need high flow to perform. with a massive pump you'll only change the hot side of the TEC by one or two degrees and that does not translate into one or two degrees less on for the cpu.

not trying to rain on your parade i'm sure you're going to have fun with those tecs but it's just not the style of those blocks to benifit from high pressure pumps.

[MaxxxRacer]

crime dog you entirely missed the point. first you said u say high flow isnt good and then u say the blocks dont benefit from high pressure. two contradictory statements as the high pressure pumps are lower flow compared to the high flow medium pressure pumps.

in any case the maze4 blocks are low restriction blocks that LIKE higher flow. and getting high flow in them is relatively easy as they are low preessure drop. and when u can get more performance pretty easily then why not go for it. and considering he is using TEC's heat dump isnt much of an issue as there will be over 500watts being dumped into the water.

[HiJon89]

ironicly ud do ALOT better with a MD20RLT... the Z is not suited for your setup. but u should get around 2.3gpm. 2.5gpm is the max for ur pump.
Great...know where I could get one without paying a fortune? How much better are we talking?

[MaxxxRacer]

well with your heatercore and those blocks you could prolly get around 3.5gpm or maybe more. with ur z you will get max of 2.4gpm.. thats being optimistic too as at 2.4gpm the head pressure would be very ow on the pq curve.

ironicly here though a D4 would perform better than ur Z i think. dont quote me on this but its worth looking into.

[HiJon89]

well with your heatercore and those blocks you could prolly get around 3.5gpm or maybe more. with ur z you will get max of 2.4gpm.. thats being optimistic too as at 2.4gpm the head pressure would be very ow on the pq curve.

ironicly here though a D4 would perform better than ur Z i think. dont quote me on this but its worth looking into.
:mad: Guess its an MD-20RLT for me :(

[Craig]

Great...know where I could get one without paying a fortune? How much better are we talking?

You can simply replace the front impellor housing & impellor of the Z to ones for the RLT which will convert it over to the RLT configuration. I would expect this to be lower cost than buying another pump & selling off the Z type.

But the real question is how much differance in the CPU temp would be found by the increase in flow from 2.3 gpm to 3.5 gpm? If the gain is under 1C then I'd say stick with what ou have.

[HiJon89]

You can simply replace the front impellor housing & impellor of the Z to ones for the RLT which will convert it over to the RLT configuration.
But then I would have to spend at least $50 on that, if I just sell my MD-20RZ and buy an MD-20RLT I shouldn't lose much money at all.

[Craig]

But then I would have to spend at least $50 on that, if I just sell my MD-20RZ and buy an MD-20RLT I shouldn't lose much money at all.

If you can trade for under the cost of the parts then of course you should go for it.

[HiJon89]

If you can trade for under the cost of the parts then of course you should go for it.
I can't imagine that 1GPM would only make a 1C difference in the CPU temp, and with 500W of heat being put into the water keeping the water temp down is going to be very important.

[MaxxxRacer]

lol.. considering ur not going to have to worry about noise i would go with the WMD20RLT as it is only 100 dollars compared to 200 for the jap version. it would perform mighty well. I will go hook up two maze4 gpu's to my test system and let u know the flow that i get. It will have a 2-199 weapon core on there too so the flowrates will be very comparable.

well i did the test and i got ~3.6gpm. so my prediction was just about right. and considering the max flow of the Z is 2.5 gpm i would most certianly suggest that you switch pumps.

[Craig]

I can't imagine that 1GPM would only make a 1C difference in the CPU temp, and with 500W of heat being put into the water keeping the water temp down is going to be very important.


As I've never used any TEC's I will freely say I've very little to base a guess on regarding them.

But once you pass 1.5gpm with straight water rigs the performance curve begins to flatten out with most blocks & above 2gpm you don't see much gain for further increases. In a straight water rig if you went from 2.5gpm to 3.5gpm your gains wouldn't be great, a 1C gain at these levels of flow is pretty hard to get.

So what would the expected/hoped for gains be for the extra flow in a TEC rig like you're planning to set up????

Thanks


EDIT:

Did a bit of checking at ProCooling & the G5 which is IMO the best performing block to date would gain about 1C if you went from 1.5 to 2.5gpm. Again, not sure just how relavent this is in regard to TEC cooled systems.

Will be of interest to me to find out just how much gain the increase from 2.5 to 3.5 would be.

[Craig]

well i did the test and i got ~3.6gpm. so my prediction was just about right. and considering the max flow of the Z is 2.5 gpm i would most certianly suggest that you switch pumps.

But just how much of a drop will this change offer at the CPU?

[Craig]

A RLT would drive a 1.6gpm flow through a G5, and a Z would add .3gpm for all of .4C lower temps approx. This is a estimate, but should be close.

But if you increased flow a further 1gpm you'd be unlikely to see a full 1C as the increasingly flat performance curve just won't allow it.

One thought regarding the buying of parts vs trading one pump for the other. If you get the parts to convert your present pump then you can swap back around to the Z if you later wished to run a higher resistence set up where the Z would offer a advantage.

[MaxxxRacer]

absolutely no idea craig. the setup i tested this on is not connected to a computer or any temperatuer measurement or heat load. right now im doing straight flow testing.

As well there is absolutely no good testing of the maze4 TEC with flowrates considered. maybe in a few months i could talk to DD about this and get one tested to see how it does. but no promises.

One thing that should be noted is that blocks perform differently with different heat loads. what i am saying is that at one heat load the curve for a block my look a certain way but if u double the heatload or tripple the heatload (the tec he will be using will just about tripple the heatload) the curve will look noteably different. I first saw this at a very good french site that does watercoolign testing.

this is merely a hypothesis but i think that higher flowrates will be more useful than at lower heatload. i hypothesize this because when u have 340watts coming into a waterblock it takes alot more water mass to absorb the heat than it would to absorb 120watts of heat in a given time frame.

[Craig]

absolutely no idea craig. the setup i tested this on is not connected to a computer or any temperatuer measurement or heat load. right now im doing straight flow testing.

As well there is absolutely no good testing of the maze4 TEC with flowrates considered. maybe in a few months i could talk to DD about this and get one tested to see how it does. but no promises.

One thing that should be noted is that blocks perform differently with different heat loads. what i am saying is that at one heat load the curve for a block my look a certain way but if u double the heatload or tripple the heatload (the tec he will be using will just about tripple the heatload) the curve will look noteably different. I first saw this at a very good french site that does watercoolign testing.

this is merely a hypothesis but i think that higher flowrates will be more useful than at lower heatload. i hypothesize this because when u have 340watts coming into a waterblock it takes alot more water mass to absorb the heat than it would to absorb 120watts of heat in a given time frame.


I agree with you on this. Also why pH needs a major upgrade soon.

A good water rig can help the Preshot's double the heat output of his present set up. I'm not sure what a AMD64 O/C'd to 3gig on water is outputing as heat output's but also sure to be well above pH's present set up.

[MaxxxRacer]

yah well pH is building his a64 3000 winchester setup as am I.

but the issue he and myself are having is the calibration of the retarded cpu. the cpu reads temps about 10C high. which is ALOT. i have yet to do full testing but its pretty annoying. also another thing i found is that the waterbath temp readings take forever to stabaliize due to the IHS insulating the sensor.. it makes calibrating the cpu very difficult.

[CrimeDog]

crime dog you entirely missed the point. first you said u say high flow isnt good and then u say the blocks dont benefit from high pressure. two contradictory statements as the high pressure pumps are lower flow compared to the high flow medium pressure pumps.

in any case the maze4 blocks are low restriction blocks that LIKE higher flow. and getting high flow in them is relatively easy as they are low preessure drop. and when u can get more performance pretty easily then why not go for it. and considering he is using TEC's heat dump isnt much of an issue as there will be over 500watts being dumped into the water.

two seperate thoughts max, that's why they're seperated.

*this is seperate from what's up there* get it?

HiJon it looks like you love wasting money so have fun

[MaxxxRacer]

he wont be spending much money on the new pump if he sells his current one or just gets the impeller housing.

[HiJon89]

lol.. considering ur not going to have to worry about noise i would go with the WMD20RLT as it is only 100 dollars compared to 200 for the jap version. it would perform mighty well. I will go hook up two maze4 gpu's to my test system and let u know the flow that i get. It will have a 2-199 weapon core on there too so the flowrates will be very comparable.

well i did the test and i got ~3.6gpm. so my prediction was just about right. and considering the max flow of the Z is 2.5 gpm i would most certianly suggest that you switch pumps.
The thing is the American version is almost 3 inches longer and there's no way I can fit that into my case, the Japanese version barely has 1mm of clearance as it is.

[MaxxxRacer]

oh ic.. that does pose a problem. lol. well id suggestjsut getting a different impeller then.

[HiJon89]

Is this all I would need?
http://www.customaquatic.com/customaquatic/itemdetail.asp?itemid=WP-IKMD0501

[MaxxxRacer]

im pretty sure but contact custom aquatic and ask them if that is everything u need. it would be great if you would let us know what they say as i can put this in the pumps giude

[HiJon89]

im pretty sure but contact custom aquatic and ask them if that is everything u need. it would be great if you would let us know what they say as i can put this in the pumps giude
Well this guy asked Custom Aquatic a similar question and got a not-so-promising response:
http://forums.procooling.com/vbb/showthread.php?t=10778
He wanted to put an MD-20RZ impeller on an MD-20RLT and Custom Aquatic contacted Iwaki for him who said that the parts aren't interchangeable :( Not sure how reliable that information is.

[MaxxxRacer]

interesting. they should be as only the impeller is different.

[HiJon89]

What do you guys think of this pump:
http://www.marinedepot.com/md_viewItem.asp?idproduct=JW1123&ast=&key=
Its supposed to be the equivalent of an MD-20RLT, is that a bunch of BS?

[BRiT]

What do you guys think of this pump:
http://www.marinedepot.com/md_viewItem.asp?idproduct=JW1123&ast=&key=
Its supposed to be the equivalent of an MD-20RLT, is that a bunch of BS?

It's no BS. Maxxx mentions this in his "Best Pump for watercooling?" thread.


BlueLine/Panworld
These pumps are very similar to the Iwakis. The head of the company of Blueline was previously the head engineer over at Iwaki. He branched off and started his own company Blueline. These pumps are very comparable to the Iwakis in performance and have similar part numbers (makes it easy to compare).

Though it seems the cost isn't much/any different than the WMD series of pumps.

[MaxxxRacer]

its essentially almost exactly a MD20RLT, designed by the same guy. but it uses cheaper parts. thats why its alot cheaper.

[HiJon89]

its essentially almost exactly a MD20RLT, designed by the same guy. but it uses cheaper parts. thats why its alot cheaper.
Would it be alright to use it instead of an MD-20RLT to save some money, or is it worth it to just bite the bullet and spend the extra $50?

[MaxxxRacer]

Well i havnt heard bad things about them. infact ive heard quite the opposite from ppl who have used them. I know niksub1 doenst particulary like panworld is a good company but u'd have to ask him why exactly he thinks this.

so i think it would be a safe bet.

[nikhsub1]

I just don't trust panworld... they never list MTBF's, and all the MCP600's and 50z's are panworld. Seen too many panworld failures, however, anyone ever seen an Iwaki failure? Just my .02

[HiJon89]

There's a guy selling an MD-20RT, not the RLT, just RT. Is this just as good? Full part number is MD-20RT-115NL

[nikhsub1]

Same difference.

[MaxxxRacer]

yah its the same pump. the just forgot to put the L in there or it's the old iwaki namiing scheme..

niksub1, now that im thinking about it ive never heard of anyone complain of a dead iwaki. i wonder what it would take to kill one of these things

(dreams of Swiftech doing a deal with iwaki for low cost (75 dollars) 12volt DC pumps with good specs)