![]() | ||
|
|
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. |
![]() |
Thread Tools |
![]() |
#1 |
Cooling Savant
Join Date: Nov 2003
Location: Indiana
Posts: 127
|
![]()
First off I know a rad is aheat exchanger but in the world of racing we speak of rads as rads and heat exchangers as a rad with another rad, or "heat exchanger", built into the rad. The main part of the rad cools the water and the second is for the oil which is then cooled by the fact that the water passages and oil passages are side by side. This is much more effective on the oil cooling than just a plain oil cooler, or just a rad for the oil only. There are other benefits to running heat exchangers but that's another story.
Here's some examples of what I'm talking about- Fluidyne Heat Exchangers What I have been thinking of trying for some time now is using a small version (one for a shifter kart which would be about 2 1/2 time the size of a large HC) to use in water cooling. What I want to do is use the main part of the rad to cool the water going to the blocks. The second part I would run a seperate loop to, from the pump and back to the res. A seperate pump and completely seperate loop altogether would work best but space is a factor. Since I do not have all the formulas necessary to do the appropriate math here I'm wondering if you guys think this would be worth the effort to try this. What about flow rate to run for each loop? I'm thinking that maybe each loop should run at differing rates. for pumps I have a Mag3 as well as a MCP600. I also have MCP 300 I'm considering using to make that second loop seperate if I can find space for it. Thanks for any advice or smack in the head I may get ![]()
__________________
"The more I learn the less I know, but I never knew as much as when I was sixteen" |
![]() |
![]() |
![]() |
#2 |
Cooling Savant
Join Date: Jun 2003
Location: North Billerica, MA, USA
Posts: 451
|
![]()
I don't get what the point is? What are you trying to cool with that seperate loop?
Gooserider
__________________
Designing system, will have Tyan S2468UGN Dual Athlon MOBO, SCSI HDDS, other goodies. Will run LINUX only. Want to have silent running, minimal fans, and water cooled. Probably not OC'c |
![]() |
![]() |
![]() |
#3 |
Cooling Savant
Join Date: Mar 2002
Location: FL
Posts: 787
|
![]()
You could run a colder liquid in the second loop, and that would cool the water, but that would be sorta ineficient, since you could just run the colder liquid through the blocks.
__________________
When you do things right, people won't be sure youv'e done anything at all. Looking to buy/trade for used Deep Fryer and Vacume Pack Sealer. |
![]() |
![]() |
![]() |
#4 |
Cooling Savant
Join Date: Oct 2002
Location: Louisiana
Posts: 174
|
![]()
I understand what you are attempting to do (Imagine that, me understanding anything ) The real flaw in your plan is not the set up but the fact that water temp only varies by a few degrees at best.
In order for the exchanger to be of any real benifit the water would have to be (a guess) 30 degrees or better different. NO PC or combination of blocks as of now are capable of introducing that much heat into the water. By building separate loops you might get .5 degree differance in CPU temp ( again a guess ) and the real question is do you want to spend another 200 bucks to get that .5 degree temp diff? PC rads work on a low temp diff. once you get within a couple degrees of ambiant, it would take a rad the size of a wall and a few hundred fans to get temp to drop any lower. What I am trying to say is it dont take much to reach the point of diminishing returns with straight water cooling. I know I didnt explan it very well but maybe the higher educated around here can put it into correct context. Aardil |
![]() |
![]() |
![]() |
#5 |
Cooling Savant
Join Date: Nov 2003
Location: Indiana
Posts: 127
|
![]()
You got it Aardil
![]() And that's about what I figured. Not much change if any. The second loop should be powered with a second pump on a completely seperate loop so the water isn't going through any blocks. The other idea to do with the second loop is to help cool the pump. Of course a second rad and pump is what I'd do here. Since my MCP600 appears to be completely sealed I am going to try and figure out a way to liquid cool the whole pump. With all the talk about pump heat lately, which after being steered to Cathars chart on pump heat I'm amazed at the heat produced, I've been thinking it's about time to cool the pumps as well. At the very least the impeller housing. Encasing a whole DC pump with something to liquid cool it can't be terribly tough. And this will at least negate some of the pump heat issue which should eliminate some of that problem. We'll see when/if I get it done ![]() There's a third idea as well but I'm just gonne try that one and se how it works. I have tons of ideas but I'm not willing to do the math like most guys here as I'm more into the ideas and experimenting. Trial and error's more fun to me ![]() The 200 bills isn't an issue since I have a spare at my race shop that I no longer use. It's an older copper Modine unit that's no longer made. They're all aluminum now.
__________________
"The more I learn the less I know, but I never knew as much as when I was sixteen" Last edited by rocketmanx; 01-30-2004 at 11:48 PM. |
![]() |
![]() |
![]() |
#6 |
Cooling Savant
Join Date: Apr 2003
Location: Blackburn / Dundee
Posts: 451
|
![]()
OK.. if you arn't cooling that second loop with a separate radiator... you are basically just going to see the loop as if you had added X litres of water to the loop and second pump... i.e. it isn't going to make much difference in the overall picture.
This is what I would recomend: (providing the ID of the tubing is the same mind) - mod it so the water flows through the radiator through both circuits (making that a quad pass rad?!?! OR - use it two cool separte computers, as a spacesaving idea. Unless you get these sort for free/on the cheap it would be better for anyone just to get a normal large HC... or two HC depending what takes their fancy.
__________________
1800+ @ 2247 (214x10.5) - STABLE, 512MB PC3700 TwinX Cosair RAM, NF7-S v2.0, GeForce3 Ti200 Parallel BIM, 120.1 Thermochill, Eheim 1048, Maze 3, Maze4 GPU, "Z" chipset, 1/2" tubing, PC-70: 5x120mm & 9x80mm fans. Internet Server & second machine (folding 24/7): 512MB DDR RAM, XP2000+ |
![]() |
![]() |
![]() |
#7 |
Cooling Savant
Join Date: Jun 2003
Location: North Billerica, MA, USA
Posts: 451
|
![]()
Rocketmanx; I think you have a confusion about what 'pump heat' is about... Pump heat doesn't have anything directly to do with the heat put out by the pump motor, indeed for many pumps there will be very little heat transferred from the motor to the pumped fluid. Pump heat is the heat generated in the fluid by the friction of making the fluid move, friction between the fluid and the pipes it flows through, etc. Much of what is labled 'pump heat' doesn't actually originate in the pump itself, but rather gets generated by the flow through the system.
Thus cooling the pump motor might get rid of an additional heat source in your case, but won't have direct impact on coolant temp. (unless you use the same loop, in which case you'll raise the temp by whatever you pick up from the pump motor that would have otherwise been dissipated in the air.) Cooling the pump head with the same loop will have no major effect other than to possibly raise the temps by a small amount due to the extra friction of the added lines. Cooling the head with a different loop might lower the temps, but only by the same amount that adding that much cooling anywhere else in the loop would. In short, there is no magic advantage to be gained by cooling the pump. Gooserider
__________________
Designing system, will have Tyan S2468UGN Dual Athlon MOBO, SCSI HDDS, other goodies. Will run LINUX only. Want to have silent running, minimal fans, and water cooled. Probably not OC'c |
![]() |
![]() |
![]() |
#8 |
Cooling Savant
Join Date: Mar 2003
Location: southeast asia
Posts: 164
|
![]()
I'm no expert here. If you easily can do what ur planing , why not try it to let the results speak for itself?
![]() |
![]() |
![]() |
![]() |
#9 |
Cooling Savant
Join Date: Mar 2003
Location: St. Paul, Minnesota, USA
Posts: 123
|
![]()
Gooserider laid out all the major issues in plain english. I agree with what he said, but I'll add some info that nobody brought up anyways.
The most significant source of heat transferred into the water by a pump is from the warmth of the motor. There is theoretically a certain amount of heat generated just by the motion and friction of the water being turned and thrust around by the pump, but this is negligible for our purposes. Different types and even models of pumps transfer different amounts of heat from the motor to the water. Some transfer almost none, and some can really heat up the water passing by. Cooling the pump's motor to channel away some of this heat is the obvious solution, and there are several ways to do it. An easy way is to make sure the pump's motor housing is in the flow path of a fan. A lot of setups I have seen have the pump mounted in front of the heat exchanger's fan to carry away any excess heat the pump produces. This will probably be sufficient to prolong the life of the pump by making the motor a little cooler, but it will in no way take away a lot of the heat. Another method uses a submergable pump (many pumps, like the commonly-used Eheims and Danners are capable of being used both in-line or submerged). Build a reservoir that has the whole pump right inside of it. The pump's inlet can be simply exposed, but the power cord and outlet plumbing will need to be run out through some type of watertight seal. This way, any heat the pump makes has to go into the water. Your water temperature will be increased, but running the outlet directly to the heat exchanger will make this a moot point. Unlike the fan option, this does not aid in keeping your water cooler at all. It increases it. The main advantage of this setup is to avoid potential pump leaks and cool the pump with hopes of a longer life. The only other thing I can think of doing with the extra passages in the radiator you describe is to run some kind of cold liquid through it. Don't bother making a water chiller go through, because you would get much better efficiency by simply running this chilled water directly through the blocks. The main forseeable advantages the second loop might facilitate is running either liquids you want to keep out of your waterblocks for corrosion, contamination, sediment, etc. reasons, or to get by with a slower flow in the secondary loop and a fast flow on the block loop. You could try building some kind of in-ground loop system to exchange heat with cold sand, or maybe run a long loop of tubing through your attic (if it's wintertime) to chill the water below ambient and hopefully cool the clean, fast-flowing block loop's water further than a normal radiator could with room temperature air. The next issue that comes up is that the double-loop radiator is not really designed to transfer heat from one loop to the other. It is designed to transfer heat out of both loops and into the air. If you really wanted to try to chill one circuit of clean/fast water with another separate circuit of less-ideal water (slower, dirtier, not-water, etc.), then a parallel flow exchanger would be the best choice. (someone please correct me if I'm using the wrong term for this) A parallel flow exchanger is designed to transfer heat from one liquid pass directly into another liquid pass without having to go through air. The easiest style to make is just a copper tube inside of a larger tube. The small tube has the warm water in it, and the bigger outer tube has the cooled water in it. The longer the nested tube setup is, the more room you have to play with for heat exchangure. Parallel exchangers are most efficient when the two loops are run in opposite flow directions. Think about it: when the warmest water comes into the end, it first encounters the chilled water that is about to leave the exchanger, and is already as warm as it can get. It will still be cooler than the warm water, so some heat is transferred out. As the warm loop's water moves down the exchanger, it is jacketed by progressibely cooler and cooler water, until the warm circuit's flow reaches the far end of the exchanger, where the chilled flow's water is at it's coldest, allowing the temperature differential to stay up. If you ran the flows in the same direction, the warmest water in the hot loop would immediately see the coldest water in the cold loop, which would have a great differential, but by the time you got to the far end of the exchanger, the warm would have been cooled down, and the cool would have been warmed up. The differential will be worse.
__________________
Abit NF7-S ... xp1800+ ... Abit GF4ti4200 ... Maxtor 120gb SATA150 Currently gathering parts to water cool everything! (read about the project) |
![]() |
![]() |
![]() |
#10 |
Cooling Savant
Join Date: Jun 2003
Location: North Billerica, MA, USA
Posts: 451
|
![]()
Excellent discussion Krazy. I think you added some good points that hadn't previously come up, particularly the comparison between this type of rad and the parallel flow type exchangers.
About the only point I might mildly argue is the question of how much impact on water temps switching from an in-line to a submerged pump installation. Most pumps seem to be in the 30-50 watt range, and assuming that submerging the pump put all that heat into the water instead of the air, you are talking about adding an amount of heat to the loop which is nearly in the neighboorhood of an extra CPU! This might or might not be significant depending on just how good the rad in your loop is, and if it has the extra heat dissipation capacity to spare. A system with a big core like a 2-342 probably wouldn't even notice, but I'm not sure about a setup with a Chevette core, or one of the smaller specialty rads. Gooserider
__________________
Designing system, will have Tyan S2468UGN Dual Athlon MOBO, SCSI HDDS, other goodies. Will run LINUX only. Want to have silent running, minimal fans, and water cooled. Probably not OC'c |
![]() |
![]() |
![]() |
#11 |
Thermophile
Join Date: Oct 2002
Location: U.S.A = Michigan
Posts: 1,243
|
![]()
Rather than have a rad on the second, cooling loop why not bong? Would increase the temp differance between the two loops and is much lower cost to run than a chiller.
|
![]() |
![]() |
![]() |
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
Thread Tools | |
|
|