[voigts]

I have been toying aroung with the idea of using a car rad to go fanless. After much searching, I have found this rad, a GDI 431272 for a 1990 GEO Metro LSI (http://www.rockauto.com/ref/GoDan/De...tml?431272.jpg) . It is copper core with plastic tanks, is only 5/8" thick, and runs $140. I have also found another brand ( http://www.radiatorworld.com/radiato...px?carno=26640) for $109. I am wondering if this would work ok without fans for cooling, and if two 50z pumps or maybe 2 MCP655 pumps would do the trick? What do you folks think?

[MVC]

brint did this a while back. You might want to do a search and see if he ever posted temps or just pm him and ask how it's working....

[Diggrr]

THIS is what I use, and the store I bought mine from.
It's all coppercore/brass tanks like a heatercore, and the center is built almost exactly like a BlackIce rad.

The core's thicker than the one you found, so it may not like the no fan plan though.

There's links on that page to have them call you with a quote, or to request an auction for your specific find. Plus $15 shipping can't be beat!

[LabRat23]

You do not need an extra pump for the car radiator. I have a rad from an S-10 and the backpressure is almost zero.

[voigts]

what are the dimensions of that rad diggrr?

Labrat: You mention backpressure being almost zero. Can you elaborate on what exactly you mean and how you measure this?

I was looking at the Geo Metro rad because of its size. I figure I would find a wa y to attach it to the side of my case if I want to.

[Craptacualr]

It would probably keep you about 40-45 after 6 hours of load running if I had to guess. I ran my custom heatercore (think BIXIII) without fans for 6 hour load temps of about 55C way back in the day just to see if it could be done. that radiator is at least 3 times the size of a BIXIII so I'd imagine you could get low 40s on load... It would take you a LONG time to get there, but it'd happen eventually

[Skeen]

I've thought about doing the same for a passive setup but I wasn't sure what to do about the huge rad inlets. The one you linked, for instance, has barbs that are 1 1/8 in. Were you going to use huge hoses and really REALLY clamp them down to your block? Or were you going to put new barbs on?

Even with small barbs soldered on it seems like your flow rate would drop considerably once it hit the (relatively) large tank in the rad. I guess not, though, since people here have apparently done it. I was just wondering how.

[XeonStrikeForce]

rad inlets easy, cap em off and use 1/2" Barbs. FLow ? lol water can not be compressed there for flow will be the same as it was befor. I too was thinking of doing this but I was planning on mounting it in my window as I did with the tiny swift tech cheapy rad they sent me, That way just a light breeze will provide a larg impact due to the surface area !

[Skeen]

Oh ok I see how you'd do the barbs. I'm still not clear on "flow". Perhaps I'm thinking of a pressure drop.

I guess what I'm thinking of is the head pressure. I just didn't know how a pump's max head is figured when there's a large volume of water to push. I'll try to illustrate:

The first pic (head2 yeah my naming conventions are all f'ed up)shows a pump with 6ft of head pushing water up a tube with a similar diameter as that of the pump barb.

Now in pic two (head) this is the same pump pushing water up a tube and into a large reservoir (pond or whatever). Is the pump still going to be able to push all that water 6ft high?

[Diggrr]

That radiator is 16" x 20" fin area, and even has a handy drain plug.

To make the fittings fit, I actually flared the ends of the fittings on it until a 1" female thread by 1" male copper pipe fitting fit perfectly inside, then torch soldered it into place.
I use a pump with 1" threads on it, so it's screwed directly into the lower fitting.
I use 1" ID tubing to go up through the floor above to my computer where it serves a three-way copper manifold. Each of the 3 legs are 3/8" tubing for cpu, gpu, and a heatercore that cools my case air, running in parallel.
Click the further link in my sig to see the manifold/case setup.

Once you get the threaded fitting attached, you can pretty much adapt anything to it for using such a small pump.

Oh, and Skeen, other than conquering loop restriction, head only matters in an open loop system. With a normal closed loop watercooling system, velocity and suction allows a relatively small pump to do a good job as long as it's bled free of air.

[Damian]

Just a random guess here, but did anyone ever try using a bike rad? Seems like it would be a nice compromise between a heatercore and car/truck rad.

[Diggrr]

Most I've seen are aluminum to go with the aluminum engine...that would be a bad mix with a copper waterblock.

[Damian]

Eh, I thought heatercores were aluminum too.

[Moto7451]

Heater cores are made of copper or aluminum. The ones that people buy for watercooling are almost always copper because of out copper water blocks.

Radiators have traditionally been made of Aluminum to cut costs & to lower weight.

[Craptacualr]

Originally Posted by Skeen

Oh ok I see how you'd do the barbs. I'm still not clear on "flow". Perhaps I'm thinking of a pressure drop.

I guess what I'm thinking of is the head pressure. I just didn't know how a pump's max head is figured when there's a large volume of water to push. I'll try to illustrate:

The first pic (head2 yeah my naming conventions are all f'ed up)shows a pump with 6ft of head pushing water up a tube with a similar diameter as that of the pump barb.

Now in pic two (head) this is the same pump pushing water up a tube and into a large reservoir (pond or whatever). Is the pump still going to be able to push all that water 6ft high?

Sort of... The first one is true. A pump with 6 feet of head will not be able to push water out of the top of a 6 foot pipe with the same ID as the outlet fitting for that pump. In the second case you are actually imposing a pressure drop with that rapid expansion of diameter. Now its albeit a VERY SMALL pressure drop, so you'd really only see like an inch or so head difference less. That second pump would prolly only have 5'11" of head pressure in that case. Remember, every time you add a component, you add pressure drop, and lower the pump's ability to push water. Pressure (pressure drop) can be converted to head just like meters can be converted to yards. So when you add a component, you add negative pressure to the system and to balance that out the pump head decreases. Make sense?

[Skeen]

Ok thanks Diggrr and Craptacualr (funny, I aways thought your name was Craptacular). I suppose it makes sense since you can have a big 'ol res and it doesn't really make any difference. But there must be some flow rate drop since just having alot of extra tubing one doesn't need ends up being an impedance. I suppose it just not enough to matter.

[voigts]

This has really stirred up some good information. I am not really concerned about the inlets being 1 1/8. I haven't measured, but I should be able to get 1" female copper adapters to fit inside easily and then solder or if all else fails JB Weld them in place. It is then just a matter of getting adapters to get down to 1/2" to fit the pump.

I am wondering about the discussion on flow rates, etc. when adding a larger rad like this. If you are adding extra tubing to flow through (in the rad) thereby adding more restriction, how can it not slow down the water? That is why I asked about using two pumps. Even in a closed system there is still gravity, friction, and forcing water through things to slow it down, correct? A pump can only do so much.

I have my fans running off of a 25w rheostat with several diodes soldered together to lower the voltage range to about 5v to 10v. I was having a problem with a loose connection, and came back to the computer one day to find that the rad fans were not running at all. The CPU temp was reading 50c with light use of the computer. Not bad considering no fans running. Of course I went back and fixed the connector. If a BIXIII can do that, I figure a rad three times that size should be capable of passively cooling ok.

If the rad is mounted vertically, would it have to have airflow to cool effectively? Or would it need some kind of airflow.

What kind of pump are you running Diggrr?

[Craptacualr]

Originally Posted by voigts

I am wondering about the discussion on flow rates, etc. when adding a larger rad like this. If you are adding extra tubing to flow through (in the rad) thereby adding more restriction, how can it not slow down the water? That is why I asked about using two pumps. Even in a closed system there is still gravity, friction, and forcing water through things to slow it down, correct? A pump can only do so much.

Well, ok this delves somewhat deeper into fluid dynamics for a clear answer. A radiator or heatercore of the variety we use is a parallel setup in and of itself. Water enters a manifold at one end and is diverted into many thinner channels to another end where a manifold picks up the water and passes it out of the radiator. By nature, a manifold of this type is very restrictive and creates high pressure drop (remember, more pressure drop = less flow). It has many rapid expansions and conractions in flowpath diameter and changes in shape which are always sources of high pressure drop. If you think about it, the water path goes from circular at the barb, to square at the manifold, to rectangular in the channel, then back to square, and back to circular. At the end of all those changes a large pressure drop has been created.

HOWEVER, running all those paths in parallel eases the magnitude of that pressure drop across each individual path. Remember, the main presure drop is going from a 1/2" ID barb to a manifold wich is in some cases is 6" to 1' wide. Then again going from that large manifold into a smaller 1/16" square channel further increases the pressure drop. The NUMBER of channels does not increase the pressure drop with anywhere near as much magnitude as the manifold itself does... So a larger radiator is nowhere near as restrictive as having 2 radiators whose manifold and fin area sizes add up to that of the one large one. Does that make sense?

yes, a pump can only do so much, which is characterized by its pressure/flow curve. It is possible to hook up a pump to a setup with so may blocks and fittings, and radiators that it can no longer push water through that system. I've never heard of anyone doing that with a WC system, but in THEORY its possible.

Also, one thing I have to clear up, in a CLOSED loop system, tube height of the heights WE use creates a negligable pressure drop. For example, in my setup, my pump is 2 feet below my radiator, but because my system is closed (not open to ambient anywhere). Those two feet of tubing are NOT 2 feet of head. They're really only a couple inches because of the bends, tube lengths, and fittings they have to go through to get to those heights. Head pressure is needed to get water to the top, but then head pressure is added to the system when the water falls back to the level of the pump, so there is little to no pressure drop in theory from that 2' increase. Once the system is open to ambient of a certain size all bets are off with that rule, so if you have a big wide open resevior everything changes. Few of us do, and a T-line which is open at the top is not the right type of ambient exposure to change the head of your system.

I'm sure I probably confused you more than answered your questions, but lemme know if you want more info, I can prolly provide it since fluid dynamics is part of my job

Edit: for a fanless radiator, the orientation should have no appreciable differences on CPU core temp provided there are no small air currents in the location of the radiator. If there are very slow air currents, that changes the heat transfer by adding forced convection in addition to stanard convection/radiation which COULD have a difference on your temps. Calculating that in theory would be way more difficult than just trying it both vertically and horizontally in the same space though

[XeonStrikeForce]

think nuclear cooling towers ! A carfully built unit can accelerate the convection air currents cuasing it to fource even more air though the rad.

[lbbo2002]

Craptacualr, great explanation!!! I was wondering about height placement for the RAD (vs pressure drop) and you explanation makes perfect sense. I am very new to the water cooling thing and I understand that explanation.

I appreciate the thought that went into you post.

[Diggrr]

Originally Posted by voigts

If the rad is mounted vertically, would it have to have airflow to cool effectively? Or would it need some kind of airflow.

What kind of pump are you running Diggrr?

I run a Rainbow Lifeguard QuietOne 3000 <link> 780 Gallons per hour, 10.5 feet of head height...I love overkill, but it's powering 3 parallel loops.

In all this, don't forget that fast moving water transfers heat best. It happens in the waterblock and in the heat exchanger too. So an underpowered pump on a bigger radiator might just dissapoint, because the slower water passing through it would lessen it's effectiveness. It's not just a game of bigger radiator=better.

I don't know what it would do without fans, or mounted verticaly doing so. It would do better than a heatercore just because of it's size, but there's no way to give you a number.
I can say that if you're dissapointed with it's performance, an easy to build cardboard shroud and a single 120mm fan would improve it without much noise.

[Craptacualr]

Originally Posted by Diggrr

In all this, don't forget that fast moving water transfers heat best. It happens in the waterblock and in the heat exchanger too. So an underpowered pump on a bigger radiator might just dissapoint, because the slower water passing through it would lessen it's effectiveness. It's not just a game of bigger radiator=better.

Diggrr is absoloutely right. Higher mass flowrates and higher velocities of water improve forced convective heat transfer and lower temps. the thin channels of a heatecore style radiator create a really high velocity through the channel and keep the water moving right at the edge of the metal, thats why they're so good at transferring heat. If you ever really look into the way Cathar's blocks work, especially the cascade and G4/G5, they use lots of small jets of water that moves at a really high velocity to increase convective heat transfer right at the point of contact with the CPU. Really a brilliant design, just wish he could make more of them

[voigts]

That was a very informative explanation Craptacualr. If (or probably when) I get into this project (I can already see my wife shaking her head at me about ANOTHER project!), it might be worth it to use my two flow meters that I got recently to evaluate what kind of flow hit a large rad would pose. I obviously don't want to buy a second $80 pump if I don't have to. It seems to me that with the pressure drop associated with a larger rad, better flow might help cooling efficiency.

Your explanation is also helpful regarding the fact that X number of feet of tubing does not equal X number of feet of head pressure lost.

After some more thinking on this, it seems to me with the explanations that you present that in most cases running two pumps should not make much of a difference over running one decent pump like a D4 or 50z. Even with a large rad there may or may not be much benefit to adding a second pump, correct?

And having a little airflow might make a niceable difference in temps, right?

[Craptacualr]

oh yeah, even a tiny bit of airflow makes a BIG difference in temperatures. forced convection is a very powerful method of heat transfer. Two pumps will increase your flowrate, however they're not 100% efficient in the real world so its not like you'll get double the pressure and double the flowrate. And yes, with a large rad there will probably not be that much difference in temperatures because the sheer size of the radiator creates a large enough surface area that adding more flowrate might be in a far less proprtional magnitude than that of the heat transferred over such a large surface area. Remember, Q=MCpDeltaT so if the large fin size creates a very large DeltaT, small changes in flowrate (M) have small changes in heat transferred (Q) because DeltaT>>>M. Furthermore, as you mentioned, 2 pumps is not a cost effective solution for that very reason when going with really big radiators.

A 50z is allready a pretty strong pump, especially for its physical size, so my suggestion if you did try out a really big radiator would be to get the rad and try it with your current pump. If the larger radiator really kills your flowrate, I'd look into getting a bigger A/C pump if you can fit one in your case. Like a Danner Mag 7 or perhaps an Iwaki MDZ20. Those things put out some hefty head pressures and would likely do the trick better than 2 50z's. Certainly one of either of those pumps is much more cost effective than 2 50z...

BTW, what kind of flowmeters do you have?

[Skeen]

So I was right?

I suppose the surface area that a big rad provides will counteract any flow rate loss.

No?

/edit:

Voigts

I'd like to see you get 4 6' long copper tubes and connect them with 5 "U" joints and lay that sucker on your floor. I wonder if just a ceiling fan in your computer room would be enough?

[Craptacualr]

Most likely you're right skeen. If he looses 1/2 of his flow, he'd need twice the fin size to make up for it, and vice versa. Its very unlikely that he'll loose 1/2 of his flow, so he should be good. without knowing the pressure drops of the two heatercores that are in question, I cant say for certain, but I'd say there's a 95% chance that the larger passive radiator will be successful

[voigts]

Sorry for not replying over the weekend Craptacualr. I had to go out of town for the Army Reserves this weekend. I have two Dwyer vfb-86-ssv flow meters that are brand new. I have not even had the chance yet to mess around with them.

Although an iwaki 20z might be more cost effective if I did not have a pump, I figure that if I do end up going with a second pump, it makes more sense just to buy another 50z since i already have one. I am hoping to do this sometime in the next month or so. I will keep you folks posted as to how it turns out if I decide to do it.

[Craptacualr]

Awesome, I'd love to find out what your results are. Just checked that out, those dwer's are the "thermometer" type of meter right? I've always wanted to put one of our high-tech pressure-drop flowmeters we have at work on my system just for curiousity's sake Maybe someday I'll get the chance

[voigts]

I would like to see some results from the high-tech digital type professional flow/pressure sensors. I am just glad to get something to work with. Another forum member sent them to me for nothing! I of course at least covered the cost of shipping and his purchase price, but I thought it was VERY generous of this person to do so as although he told me he got them on ebay for only about $10, those meters are like $80 each. I want to set up and do some experimenting in a month or two with different things to see how they effect flow.