Modified Rad Causing Setup Dilema

psychofunk · 05-11-2004, 03:05 PM

Okay, so I got this caprice heatercore and sweated on a barb in the lower tank. And on top I am splitting my flow into the rad between the two inlets (what used to be in and out are now both in). My idea is that one, it is now a less restrictive single pass rad and the second is that the split inlets will slow down flow through the rad allowing for better cooling. The single outlet at the bottom will then raise flow back up.

So first off, in splitting and then rejoining flow do I regain 100%? 90%? 80%? How much flow. And with this in mind how should I set my system up. Pump > CPU > GPU > NB > Rad > Pump or is most flow regained that I should make it Pump > Rad > CPU > GPU > NB > Pump so that the cpu gets the cooler Rad water as opposed to the slightly heated pump water?

zer0signal667 · 05-11-2004, 04:16 PM

Keep in mind that slowing flow does not help cool the water. Just like faster flow is better to transfer heat from a surface, it is also better to transfer heat to a surface. However, the faster flow due to less restriction may help you out. I think heatercores are pretty low resistance to begin with though, so I'm not sure why you would even want to do this...
Got any pics? SOunds interesting, in any case.

psychofunk · 05-11-2004, 11:15 PM

Anyone? Any idea about the flow loss from splitting and rejoining?

Groth · 05-11-2004, 11:50 PM

I don't think it will make an appreciable difference. You have a little less flow resistance through the two parallel barbs, but add the flow resistance of a Y fitting. Going from double pass to single pass is good for performance, but the 2 vs. 1 inlets is more a convenience thing.

I assume the top tank is still split?

psychofunk · 05-12-2004, 12:42 AM

I think I was a little unclear. It is two in and one out. What my concern is that if I run the setup, pump outlet > Y > Dual inlets on Rad > WW CPU block > Y > Maze4 GPU > Z chip NB > Pump inlet, that I will be hurting flow because of the fact that the high velocity water out of the pump outlet will be split by the Y and then rejoined at the bottom of the modified rad. So I am trying to find out if in seperating and then rejoining is there any performance loss and if possibly there is a better setup to use, such as going pump outlet to blocks.

Normally it would be a choice of higher velocity (pump outlet to cpu block) or cooler water (Rad to cpu block). And normally I would opt for the cooler water but my concern is that the process of seperating and rejoining would rob me of enough flow that choosing pump outlet to cpu block may be the better choice. Now if the seperating and rejoining (2 in 1 out) is no different then a normal one in one out then I will go with rad to cpu block.

psychofunk · 05-12-2004, 12:56 AM

Okay, here is a pic of the rad

Groth · 05-12-2004, 01:05 AM

When you have everything in series they all get the same flow - the first thing after the pump does not get a higher velocity. If the first one was getting more water than the second, where would the extra water go?

The water temperature across a component changes very little - on the order of a couple tenths of a degree. Will you be able to notice the tenth degree difference between having the rad before or after the CPU block?

Since you are using a waterblock with dual outlets, you should save yourself the flow restriction of the Ys: Pump->GPU->NB->CPU=>Rad->Pump

psychofunk · 05-12-2004, 01:12 AM

Quote:

Originally Posted by Groth

When you have everything in series they all get the same flow - the first thing after the pump does not get a higher velocity. If the first one was getting more water than the second, where would the extra water go?

The water temperature across a component changes very little - on the order of a couple tenths of a degree. Will you be able to notice the tenth degree difference between having the rad before or after the CPU block?

Since you are using a waterblock with dual outlets, you should save yourself the flow restriction of the Ys: Pump->GPU->NB->CPU=>Rad->Pump

So what you are saying is that I will see very little to no diff in temps regardless of block order so ditch the Y's and put the dual outs of the block to the dual in's of the rad? Hmmmmmmmm!

Does this mean that flow, with all the restictions in the loop, is the same throughout the loop. So my fight is to reduce the number of restrictions. I was under the misconception that in different parts of the loop that there could be different flow rates. You have to forgive me but I am a laymen and I am trying to wrap by brain around this.

Groth · 05-12-2004, 01:28 AM

Quote:

Originally Posted by psychofunk

So what you are saying is that I will see very little to no diff in temps regardless of block order so ditch the Y's and put the dual outs of the block to the dual in's of the rad?

Exactly.

Quote:

Does this mean that flow, with all the restictions in the loop, is the same throughout the loop.

Yes, the flow is the same. Think about it: if a liter of water flows into a block, doesn't the same liter of water have to flow out, and into the next bit?

Quote:

So my fight is to reduce the number of restrictions.

Yep. Less restriction equals more flow, more flow gives better performance.

bigben2k · 05-12-2004, 03:00 PM

The flow resistance due to splitting would be small, relative to the heatercore flow resistance that you're getting back. You should gain a little bit of flow overall

Your configuration is an excellent idea, I think I'll do the same thing.

sandman · 05-12-2004, 05:03 PM

why not have the two outlets of the WW go into the top of the heatercore?

psychofunk · 05-12-2004, 05:54 PM

Quote:

Originally Posted by sandman

why not have the two outlets of the WW go into the top of the heatercore?

That is what groth suggested. He actually said that there would be little to no difference in temps no matter what config I run the block in so I am considering that exactly.

Blackeagle · 05-21-2004, 07:34 PM

The converted to single pass rad does alter water VELOCITY, as you now have double the number of tubes carrying the flow. For RATE is however increased slightly due to the 4X lower resistence the rad should now give.

The question as to which is better in routing this set up is best answered with a question..

Which way results in the lowest number of fittings and the shortest lines with fewest tight bends in the lines? The answer of this question is also the answer as to which way is best.

I'd also strongly suggest you forget the NB block, they are a waste of flow rate. Not only does the block it's self impact flow rates, but the several tight bends required to route the lines to it will have a pretty large impact on flow rates. All these negatives for a block that will do NOTHING more than a Microcool or Swiftech active NB air cooler. So dropping the NB block will be the best thing I can see to do to insure best flow.

Best of luck!

05-11-2004, 03:05 PM	#1
psychofunk Cooling Savant Join Date: Jan 2003 Location: Los Angeles Posts: 365	Modified Rad Causing Setup Dilema Okay, so I got this caprice heatercore and sweated on a barb in the lower tank. And on top I am splitting my flow into the rad between the two inlets (what used to be in and out are now both in). My idea is that one, it is now a less restrictive single pass rad and the second is that the split inlets will slow down flow through the rad allowing for better cooling. The single outlet at the bottom will then raise flow back up. So first off, in splitting and then rejoining flow do I regain 100%? 90%? 80%? How much flow. And with this in mind how should I set my system up. Pump > CPU > GPU > NB > Rad > Pump or is most flow regained that I should make it Pump > Rad > CPU > GPU > NB > Pump so that the cpu gets the cooler Rad water as opposed to the slightly heated pump water?

05-12-2004, 03:00 PM	#10
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	The flow resistance due to splitting would be small, relative to the heatercore flow resistance that you're getting back. You should gain a little bit of flow overall Your configuration is an excellent idea, I think I'll do the same thing. __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

05-11-2004, 04:16 PM	#2
zer0signal667 Cooling Neophyte Join Date: Apr 2004 Location: Columbus, OH Posts: 25	Keep in mind that slowing flow does not help cool the water. Just like faster flow is better to transfer heat from a surface, it is also better to transfer heat to a surface. However, the faster flow due to less restriction may help you out. I think heatercores are pretty low resistance to begin with though, so I'm not sure why you would even want to do this... Got any pics? SOunds interesting, in any case.

05-11-2004, 11:15 PM	#3
psychofunk Cooling Savant Join Date: Jan 2003 Location: Los Angeles Posts: 365	Anyone? Any idea about the flow loss from splitting and rejoining?

05-11-2004, 11:50 PM	#4
Groth Cooling Savant Join Date: Mar 2003 Location: MO Posts: 781	I don't think it will make an appreciable difference. You have a little less flow resistance through the two parallel barbs, but add the flow resistance of a Y fitting. Going from double pass to single pass is good for performance, but the 2 vs. 1 inlets is more a convenience thing. I assume the top tank is still split?

05-12-2004, 12:42 AM	#5
psychofunk Cooling Savant Join Date: Jan 2003 Location: Los Angeles Posts: 365	I think I was a little unclear. It is two in and one out. What my concern is that if I run the setup, pump outlet > Y > Dual inlets on Rad > WW CPU block > Y > Maze4 GPU > Z chip NB > Pump inlet, that I will be hurting flow because of the fact that the high velocity water out of the pump outlet will be split by the Y and then rejoined at the bottom of the modified rad. So I am trying to find out if in seperating and then rejoining is there any performance loss and if possibly there is a better setup to use, such as going pump outlet to blocks. Normally it would be a choice of higher velocity (pump outlet to cpu block) or cooler water (Rad to cpu block). And normally I would opt for the cooler water but my concern is that the process of seperating and rejoining would rob me of enough flow that choosing pump outlet to cpu block may be the better choice. Now if the seperating and rejoining (2 in 1 out) is no different then a normal one in one out then I will go with rad to cpu block.

05-12-2004, 01:05 AM	#7
Groth Cooling Savant Join Date: Mar 2003 Location: MO Posts: 781	When you have everything in series they all get the same flow - the first thing after the pump does not get a higher velocity. If the first one was getting more water than the second, where would the extra water go? The water temperature across a component changes very little - on the order of a couple tenths of a degree. Will you be able to notice the tenth degree difference between having the rad before or after the CPU block? Since you are using a waterblock with dual outlets, you should save yourself the flow restriction of the Ys: Pump->GPU->NB->CPU=>Rad->Pump

05-12-2004, 05:03 PM	#11
sandman Cooling Neophyte Join Date: Aug 2003 Location: Ohio Posts: 50	why not have the two outlets of the WW go into the top of the heatercore?

05-21-2004, 07:34 PM	#13
Blackeagle Thermophile Join Date: Oct 2002 Location: U.S.A = Michigan Posts: 1,243	The converted to single pass rad does alter water VELOCITY, as you now have double the number of tubes carrying the flow. For RATE is however increased slightly due to the 4X lower resistence the rad should now give. The question as to which is better in routing this set up is best answered with a question.. Which way results in the lowest number of fittings and the shortest lines with fewest tight bends in the lines? The answer of this question is also the answer as to which way is best. I'd also strongly suggest you forget the NB block, they are a waste of flow rate. Not only does the block it's self impact flow rates, but the several tight bends required to route the lines to it will have a pretty large impact on flow rates. All these negatives for a block that will do NOTHING more than a Microcool or Swiftech active NB air cooler. So dropping the NB block will be the best thing I can see to do to insure best flow. Best of luck!

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