Go Back   Pro/Forums > ProCooling Technical Discussions > General Liquid/Water Cooling Discussion
Password
Register FAQ Members List Calendar JavaChat Mark Forums Read

General Liquid/Water Cooling Discussion For discussion about Full Cooling System kits, or general cooling topics. Keep specific cooling items like pumps, radiators, etc... in their specific forums.

Reply
Thread Tools
Unread 11-05-2002, 02:20 AM   #101
g.l.amour
Cooling Savant
 
g.l.amour's Avatar
 
Join Date: Mar 2002
Location: on da case
Posts: 933
Default

wise words, they are...

still, using the right y's and pump can help. but having a low flow block makes it that much harder indeed.
g.l.amour is offline   Reply With Quote
Unread 11-05-2002, 02:52 AM   #102
MadDogMe
Thermophile
 
MadDogMe's Avatar
 
Join Date: Aug 2002
Location: Just shut up ;) ...
Posts: 1,068
Default

Quote:
Remember the modding discussed in regard to Danger Den blocks? If I split the flow just before the blocks but run 3/8" to the NB and Gfx it might be neat to do the insides of both of them with the dimpled "floor" ect. Thoughts?
I'd have the GPU&NB blocks as 'wide open'(dimple floor yes!) as I could!.
If you ran them parallel with the CPU you'd want them more restrictive, to force more of the flow into the CPU block...

The fan noise can happen in cases as well I believe(especialy 'airtight' ones), but is antagonised by being closer. it's only when they are run very fast it happens, 7Volt is 'DeRiguer' for rad~fans, I use a POT to vary/slow the speed so the noise stops. until I can afford a decent rad&fan.
MadDogMe is offline   Reply With Quote
Unread 11-05-2002, 04:53 PM   #103
Blackeagle
Thermophile
 
Join Date: Oct 2002
Location: U.S.A = Michigan
Posts: 1,243
Default

Here is why I was thinking in terms of splitting the flow before it reaches the CPU.

If I buy and use the high psi MD-30 pump there is going to be a 5/8" line coming off the pump to the rads. Now I am thinking that the rads would be set up to be end/end with the four inlet/outlet ports as close together as possible. The large line from the pump comes to a custom Y that splits this large line into two .5" lines going into the two rads. Then rejoin the two .5" lines into one large one again, as we don't want to send 1/2 the flow to the NB & GFX. But there are no blocks that I've looked at that can accept such a large line. And wouldn't the loss of flow caused by having a reducer just before the CPU block waste in back pressure much of what you were trying to save for the CPU by not splitting the flow before the CPU? Now if we were talking about only the Atlantis block we could consider modding the blocks inlet/outlet barbs to a over spec size. (Or perhaps the higher speed of flow is what you think would offer best gains?) But this won't work on Cathar's Little Rapids water block, the impingment jet will be the pressure point in this block not the barb, agree? And lastly, wouldn't total system flow be higher if the 5/8" lines flow were split into 3/8"&1/2" before the CPU. Then encourage in every way possible all the flow that the .5" line can possibly take to go that way by the custom Y and tubulance mods to the NB and Gfx blocks. Then both of the lines coming off the blocks to go to the res near by saving need for another Y, the outlet line of the res would be .75" instead going back to the pump. this res will be attached to the right side case wall between the MB and the 5.25" drive bays. I'm trying hard to figure ways to keep this system as tight as possible for best preformance and minimum of restrictions. And there is not a single 90 barb in it, (2 45* on the Gfx).

the MD-30 has at the beginning a flow of 4.6 gpm @ 4' which is 2X the height of anything that will be in this water system in the case (18-20"). So we would have a slight gain in flow due to that reduction in the high point of the system but it won't be much. Now with every restriction of the cooling system the flow rate is going to fall off. With a start point of 4.6+ gpm we can't afford more than a 35% reduction or I'll be below the prime flow suggested by BillA of 3 gpm. And with 2 or at least 1 large rad + CPU, NB and GFX w/bs and perhaps as much as 5' of tubing of various sizes the total restriction figure is beyond my math to figure a estimate of. Can a MD-30 handle all the above without a loss larger than 35%, is it that strong?

I wish my math background was equal to the task of understanding BillA's work better, but it isn't. I regret I don't understand nearly half the terms he uses, and in some posts less than half. I glean what I can but it isn't enough. If anyone here on this board can come up with a ballpark figure of the reduction in flow caused by the above system's restrictions I'd be VERY gratfull indeed.

g.l.amour,

Who is bigmomma and what sort of rad has this (site, company ect.) come out with? Where to find?

Thanks


Thoughts and further advise guys?

Last edited by Blackeagle; 11-05-2002 at 05:03 PM.
Blackeagle is offline   Reply With Quote
Unread 11-06-2002, 02:49 PM   #104
bigben2k
Responsible for 2%
of all the posts here.
 
bigben2k's Avatar
 
Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
Default

Here's the P/Q curve, for the 2-MD-SC, which has the same specs as the 2-MDQ-SC:
Attached Images
File Type: jpg 2-md curve.jpg (47.2 KB, 252 views)
bigben2k is offline   Reply With Quote
Unread 11-06-2002, 03:03 PM   #105
#Rotor
Cooling Savant
 
#Rotor's Avatar
 
Join Date: Feb 2002
Location: Dione, sector 4s1256
Posts: 852
Default

now just imagine for a moment.......

110v Ac to 110v Ac PWM....... With a frequency knobb on the outside
__________________
There is no Spoon....
#Rotor is offline   Reply With Quote
Unread 11-06-2002, 08:34 PM   #106
murray13
Cooling Savant
 
murray13's Avatar
 
Join Date: Sep 2002
Location: Omaha, NE USA
Posts: 216
Default

Quote:
Originally posted by Blackeagle
Here is why I was thinking in terms of splitting the flow before it reaches the CPU.

....Thoughts and further advise guys?
Two things:

1. Have you considered how large the md-30rzt really is? It is 3.5" wide 4.66" tall and 11.25" long. Are you going to have room inside the case for such a large pump?

2. If you split the flow before the CPU you won't have max flow going thru the CPU. Whether you split the flow before or after the cpu, if the split uses the same components (tee's, hose, etc.) the total flow will be the EXACT same.
Code:
                    ----2gpm---CPU----
                  /                   \
>pump-----3pgm---[                     ]-----3gpm----Rad>
                  \----1gpm--NB--Gfx--/


                      --------2gpm-------
                     /                   \
>pump---3gpm--CPU---[                     ]----3gpm---Rad>
                     \---1gpm--NB--Gfx---/
sorry about the bad diagram but can you see what I mean?

Just some food for thought. I know this doesn't answer your question. IMHO the md-30rzt will be enough to get great flow, not sure if it will meet your 3gpm goal but it will come closer than any other pump I know of.
murray13 is offline   Reply With Quote
Unread 11-06-2002, 08:50 PM   #107
bigben2k
Responsible for 2%
of all the posts here.
 
bigben2k's Avatar
 
Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
Default

The MD-30 is still overkill.

I don't see the point in having a 2gpm bypass. If you're going to put the CPU in the mainline, then split the rest between the GPU and the NB. They should be relatively unobstructive (little restriction).
bigben2k is offline   Reply With Quote
Unread 11-07-2002, 03:15 AM   #108
MadDogMe
Thermophile
 
MadDogMe's Avatar
 
Join Date: Aug 2002
Location: Just shut up ;) ...
Posts: 1,068
Default

They are restrictive though are'nt they?. for normal GPU/NB blocks it's a good idea. for Uberflow ones... I'd go series(or try it at least, mines a gaming rig, cool that GPU! , system trade off ), or series~parallel (CPU> NB&GPU in parra')
the flows would'nt both be 3gpm(forcing it all through the CPU is more restrictive) but you can gaurentee the flow in the CPU block would be maxed out in the second setup, above 2gph!...

Anything but putting the CPU in a parallel loop! ...

Tell the truth, he could prolly put them all in series or triple parallel!, use brake line for hose, and still have uber cooling with a 30series!! ...
MadDogMe is offline   Reply With Quote
Unread 11-07-2002, 11:02 PM   #109
Blackeagle
Thermophile
 
Join Date: Oct 2002
Location: U.S.A = Michigan
Posts: 1,243
Default

Well I've really enjoyed reading up on the articles you fellas suggested to me today.

I now better understand why a single large rad would be better than duals due to the restrictions being added at each added Y or other connections involved sith going to duals. I've reread a bunch of differant threads and ideas today (day off from work) and have now seen info in a number of peoples posts that I didn't get before.

Murray 13,

A very good question on the pump's size vs. space.

Yes I've cconsidered it carfully and that is why I'm going to buy a large full tower/server style case, the Lain-Li PC76 to be exact. This is a huge case that has the floor area to handle what I want to do and that is the only reason I've decided to go with it instead of many less costly cases. The floor area is a huge 10X24" and that will, by pulling the lower drive cage or cages give me the space I need where I want it to be in the case, down low. With all the upper 5.25" bays this case has I can live without the lower ones.

edit : I guess I should have added one other fact as to the case choice. The location this will be set up in preculdes the use of a cube type case as it wouldn't fit as it would be to wide. The PC76 is the best compromise I can find.

Last edited by Blackeagle; 11-07-2002 at 11:27 PM.
Blackeagle is offline   Reply With Quote
Unread 11-08-2002, 03:51 AM   #110
MadDogMe
Thermophile
 
MadDogMe's Avatar
 
Join Date: Aug 2002
Location: Just shut up ;) ...
Posts: 1,068
Default

Using oversize gets rid of all the inherent restriction fittings suffer from, apart from the actual bend(s). don't forget the 1/2 fittings people normaly use(and scream restriction!) are in fact 9.9mm inner!, far cry from 12mm tube inner, then add that to the bend...

Use 5/8~14mm fittings with 1/2" hose for best fit...

If you can afford it, experiment with the rads, heaytercores are cheap, but bare in mind that PC rads claim to be low pressure drop/unrestrictive nowadays ...

Did you read Robotechs graph from the last page?. http://w3.one.net/~robotech/images/W...ump_curves.jpg

It was a revalation to me ...
MadDogMe is offline   Reply With Quote
Unread 11-08-2002, 04:29 AM   #111
murray13
Cooling Savant
 
murray13's Avatar
 
Join Date: Sep 2002
Location: Omaha, NE USA
Posts: 216
Default

-->Blackeagle,

Great! I just didn't want you to pay USD$200+ for a pump and find out that it is half the size of some cases.
murray13 is offline   Reply With Quote
Unread 11-08-2002, 08:20 AM   #112
Blackeagle
Thermophile
 
Join Date: Oct 2002
Location: U.S.A = Michigan
Posts: 1,243
Default

Yeah Murray,

I know what you mean, even in this beast of a case and with both the lower drive cages out it will still eat up 40% of available space. but the MD-30 is no larger in that respect than a MD-20.

In regards to the flow routing, I see the point several of you guys have made in referance to not depriving the CPU of any flow. And have several thougths.

I'm going to try it several ways when I get it built, won't cost anything more than time and a bit of tubing.

I've thought of one thing that I'd missed up to now. With Cathar's block with it's dual outlets there would be no Y needed to do as BigBen2k & Maddog suggested in splitting the flow to both the NB and the Gfx as it leaves the CPU, just run one of the outlets to each, then lines coming off those blocks go direct to the air trap/res. This would be a short dual loop and a tight one with only ONE xtra fitting (xtra barb at res, compared to running all in series) as the line out of the res would restore the flow to a large line again. Nice.

With the Atlantis I'd still have a Y and a extra fitting at the res. But with this arrangment I'd be able to send more cooant to the Gfx vs. the NB by having a offset Y with a .5" line going to the Gfx and a 3/8" line to the NB. Even with a added fitting I may like this even better than above, as the Gfx needs more of the flow by far vs. a NB. Nice part of making the custom Y here would be that I could play with the offset Y's smaller outlet to the NB within that Y by drilling a undersize outlet hole in the NB side of the offset, 1/4 or 5/16" to force more to the Gfx and all I'd have to do to tune it is change the size of that restricted outlet.

This could get to be real fun. (someone at this point should call 911, he's lost it ! ! Another one around the bend and over the top !)



Blackeagle is offline   Reply With Quote
Unread 11-08-2002, 08:29 AM   #113
bigben2k
Responsible for 2%
of all the posts here.
 
bigben2k's Avatar
 
Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
Default

Quote:
Originally posted by Blackeagle
I've thought of one thing that I'd missed up to now. With Cathar's block with it's dual outlets there would be no Y needed to do as BigBen2k & Maddog suggested in splitting the flow to both the NB and the Gfx as it leaves the CPU, just run one of the outlets to each, then lines coming off those blocks go direct to the air trap/res. This would be a short dual loop and a tight one with only ONE xtra fitting (xtra barb at res, compared to running all in series) as the line out of the res would restore the flow to a large line again. Nice.
But if the GPU and NB blocks aren't equally restrictive, then you might throw off the balance of flow within Cathar's block. I think you might reconsider that option.
bigben2k is offline   Reply With Quote
Unread 11-08-2002, 10:54 AM   #114
Blackeagle
Thermophile
 
Join Date: Oct 2002
Location: U.S.A = Michigan
Posts: 1,243
Default

Yeah, I did think about it as a possible problem. And this was a real concern I had at first with the idea as the GFX line, seems to me, would have higher resistances, as I'd run.5" to the NB in this set up to minimize this. Just due to the xtra tubing involved in both reaching down a bit farther to it and then also routing up to the res, before even considering the differances in the block designs. The blocks themselves are differant enough to effect things also. So I agree that there would be a imblance, just not sure of how large a amount.

But given the high pressure jet in Cathar's Little Rapids which would in this case be getting it's water from what would amount to a 5/8" line (I won't reduce it down until right at the block, which would start the velocity increse even earlier.). Once the water is injected by the impingment jet and the fllow splits into the two differant directions I don't see how the imbalance in the lines after they leave the CPU block could be great enough to overcome and effect the much more powerfull force of the inlet area of the CPU block. At least not enough for it to have a messureable change.

A redistribution block could be cut out of plixi with an small open pool inside it with both of the CPU lines coming in one side and then differant lines leaving to the NB and Gfx. Now the redistribution block would offer the advantage of allowing a much smaller line to the NB and a larger line with higher flow to the Gfx. But this redistribution block would have to be another rather large point of resistence to the overall flow would it not? Worth it?

Last edited by Blackeagle; 11-08-2002 at 11:11 AM.
Blackeagle is offline   Reply With Quote
Unread 11-08-2002, 11:20 AM   #115
bigben2k
Responsible for 2%
of all the posts here.
 
bigben2k's Avatar
 
Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
Default

I think that as long as the flow is more or less balanced, it should be ok.
bigben2k is offline   Reply With Quote
Unread 11-08-2002, 11:28 AM   #116
airspirit
Been /.'d... have you?
 
airspirit's Avatar
 
Join Date: Jul 2002
Location: Moscow, ID
Posts: 1,986
Default

You really shouldn't consider anything that'll interfere directly with flow inside of the block. You would be better served to use a four way PVC junction than to try to feed half out to the GPU and half to the NB. Use two adjacent parts of the "X" for input and the other two adjacent openings for output. I'd be barely more restrictive, but would give a place for flow rate normalization before splitting to the blocks. Otherwise you risk hotspots and low flow areas inside of the block which can lead to crispy smokey CPU chips (with extra barbeque sauce for flavor!).
__________________
#!/bin/sh {who;} {last;} {pause;} {grep;} {touch;} {unzip;} mount /dev/girl -t {wet;} {fsck;} {fsck;} {fsck;} {fsck;} echo yes yes yes {yes;} umount {/dev/girl;zip;} rm -rf {wet.spot;} {sleep;} finger: permission denied
airspirit is offline   Reply With Quote
Unread 11-08-2002, 06:13 PM   #117
Blackeagle
Thermophile
 
Join Date: Oct 2002
Location: U.S.A = Michigan
Posts: 1,243
Default

Hey, I like barbeque, but not at hundreds of dollars a pop.

After making my response to Bigben, I got to thinking about it some more. And you and Bigben are right, and the little I might lose in flow isn't worth risking the CPU for. And in addittion that redistribution hub will allow me to send much more to the GFX than it would get if the coolant were evenly split.

the more I think about this system the more excited I'm getting about it.

The thing I find wondering the most is which of the CPU blocks I intend to order will be better in this system.

Last edited by Blackeagle; 11-08-2002 at 06:35 PM.
Blackeagle is offline   Reply With Quote
Unread 11-09-2002, 05:24 PM   #118
nexxo
Cooling Savant
 
Join Date: Jul 2002
Location: Brimingham, UK
Posts: 385
Default

OK, here are pics of the pump I was talking about: continuously rated, this Johnson pump runs on 12V (just plug into your PSU), 2.2A, 26W; pumps 356 GPH / 22.5 l/min at 3ft. head.

Dimensions are 18cm long by 5.5 cm width.

The only problem is the price... $179,-- i.e. £130,--. Ouch! Regardless, I decided this will be the one for my project.

Pics below:









Whaddaya think?
__________________
"There is a thin line between magic and madness"
nexxo is offline   Reply With Quote
Unread 11-09-2002, 05:33 PM   #119
Puzzdre
Cooling Savant
 
Puzzdre's Avatar
 
Join Date: Aug 2002
Location: Croatia
Posts: 969
Default

Well, I think I should express my opinion with this:


and also express my opinion in availability and price here:


I really, really like it!
__________________
'Out of cheese error...
...please reboot the universe (press the GBL to continue)'
Puzzdre is offline   Reply With Quote
Unread 11-09-2002, 08:11 PM   #120
bigdawginva
Cooling Savant
 
Join Date: Oct 2002
Location: VA
Posts: 104
Default

Looks like Swifty's switching to a new pump as well.

http://www.swiftnets.com/products/mcp300.asp
bigdawginva is offline   Reply With Quote
Unread 11-10-2002, 08:08 AM   #121
nexxo
Cooling Savant
 
Join Date: Jul 2002
Location: Brimingham, UK
Posts: 385
Default

Looks like the Idra from Sicce Pumps. Good pump, small, quiet, powerful. I definitely like the anodised aluminium fittings Swifty put on it. Class!

I'll have to see what I'm going to do with my 3/4"/20mm OD pump connections to connect them to 1/2" tubes... Any suggestions welcome
__________________
"There is a thin line between magic and madness"

Last edited by nexxo; 11-10-2002 at 02:35 PM.
nexxo is offline   Reply With Quote
Unread 11-11-2002, 06:30 PM   #122
airspirit
Been /.'d... have you?
 
airspirit's Avatar
 
Join Date: Jul 2002
Location: Moscow, ID
Posts: 1,986
Default

Ooh. For big projects:

Supreme Pondmaster 1200.

110W, mag-drive (too strong of a field for self containment, I think ... externally mount it or wipe your drives), 3/4" in/out, max head 15ft, 120VAC/60Hz.

1200@0'
1200@1'
1180@2'
1150@3;
1130@4
1120@5
950@6
810@7
710@8
650@9
600@10
500@11
450@12

Need I list more? Tuff pump.
__________________
#!/bin/sh {who;} {last;} {pause;} {grep;} {touch;} {unzip;} mount /dev/girl -t {wet;} {fsck;} {fsck;} {fsck;} {fsck;} echo yes yes yes {yes;} umount {/dev/girl;zip;} rm -rf {wet.spot;} {sleep;} finger: permission denied
airspirit is offline   Reply With Quote
Unread 11-12-2002, 12:55 PM   #123
nexxo
Cooling Savant
 
Join Date: Jul 2002
Location: Brimingham, UK
Posts: 385
Default

Very handy if you plan to buid a 12 foot tower case PC...
__________________
"There is a thin line between magic and madness"
nexxo is offline   Reply With Quote
Unread 11-12-2002, 12:56 PM   #124
airspirit
Been /.'d... have you?
 
airspirit's Avatar
 
Join Date: Jul 2002
Location: Moscow, ID
Posts: 1,986
Default

But I am ...
__________________
#!/bin/sh {who;} {last;} {pause;} {grep;} {touch;} {unzip;} mount /dev/girl -t {wet;} {fsck;} {fsck;} {fsck;} {fsck;} echo yes yes yes {yes;} umount {/dev/girl;zip;} rm -rf {wet.spot;} {sleep;} finger: permission denied
airspirit is offline   Reply With Quote
Unread 11-13-2002, 05:02 AM   #125
nexxo
Cooling Savant
 
Join Date: Jul 2002
Location: Brimingham, UK
Posts: 385
Default

The first PC where you have to walk upstairs to insert a floppy... Still, with that pumping power perhaps the res could double as a Whirlpool/Jacuzzi...
__________________
"There is a thin line between magic and madness"
nexxo is offline   Reply With Quote
Reply


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

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -5. The time now is 07:25 AM.


Powered by vBulletin® Version 3.7.4
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
(C) 2005 ProCooling.com
If we in some way offend you, insult you or your people, screw your mom, beat up your dad, or poop on your porch... we're sorry... we were probably really drunk...
Oh and dont steal our content bitches! Don't give us a reason to pee in your open car window this summer...