AMDMB's Watercooling article

[UnaClocker]

http://www.amdmb.com/article-display.php?ArticleID=179
Go read that, they hit the nail on the head with this article. Alot of you could learn a thing or two from that article. Especially anyone that thinks a bigger pump is a guarantee for lower temps.

[myv65]

I appreciate the mentioning of the article along with your kind words. Hopefully you'll find the remaining articles every bit as accurate. The next one will cover fluids and will go up in about ten days, give or take.

As noted in the introduction to this series, I'm really looking for feedback from people with water cooling experience. Nothing beats hands-on experience mixed with a good background education. I've got a little of the former and a lot of the latter. If anyone's got suggestions, I'm all ears.

My goal is to have a complete series of articles that covers as much material as possible while making the facts as plain as I can. Any assistance reaching this goal is great and will be credited to the contributor.

Thanks again,

Dave

[Brad]

I dunno una, I didn't see a single piece of hard evidence that a huge pump won't help.


myv: great article, could you do a real world test series, finding out temps with certain rads, blocks, pumps, tubing, etc

[bigben2k]

You're right Brad, but I don't think that was the point. It seems like this approach is mostly theoretical, so I wouldn't expect real data anytime soon.

But that's ok, because there's a lot of theory to cover. I'd like to see an article that covers flow restriction calculations. I found an article (I don't remember where) that converts typically found items in plumbing, to a useful figure: head.

We all know that a pump can only lift/pull water to a certain height, but few of us have actually caculated the whole resistance of flow as a measure of that height. With that info, we'd know what to expect from a typical rig, and have a better idea about what pump to choose.

The article does mention some figures about flow rate versus effective temps, and so it seems that a rig should be flowing at least 30 gph. Anything less becomes exponentially less efficient. Anything more becomes exponentially irrelevant.

Using 30 GPH (which in itself is still debatable), and that 5 psi pressure is just about ideal for max flow rate through our tubing, we can then calculate the actual needed size of that tubing. As Una found out, bigger is not necessarily better.

Knowing the size of the tubing, and the overall flow resistance, we can then select an appropriate pump.

(and for the kiddies out there running an Eheim 1250, no, your flow rate is NOT actually 317 gph).

[Cova]

One thing I always see everyone mentioning around here is that a pump will only lift water so high. Assuming that the pump is in a closed-loop system and all the seals are good, height shouldn't be a problem. The weight of any water that the pump is lifting should be offset by the weight of any water that is going down the other side of the system. Since it is a closed loop the two forces should even out fairly close to each other.

The only think I can see that would force the pump to actually lift a lot of water would be if the volume of water on the "up side" was significantly more than the volume of the water on the "down side". But it should be easy enough to design your system so that most of the high-volume pieces (rad, block, res if you have one) have the water flowing relatively downwards through them, or at least angled so that if that component was filled and then disconnected that gravity wouldn't pull all the water out of it towards the output of the pump. Should be able to get gravity to work for your system instead of against it.

[deeznuts]

i think most people are equating head with restrictions. the promotional graphs usually show a dropoff in flow with height (head) while in a closed loop system, yes it shouldn't matter, you can analogize all the restrictions (fittings, turns, block, radiator, etc.) into head, and that this will reduce the flow.

[bigben2k]

Yes, the gravity of the water should compensate for itself on the backside except if the length is different, as it will most certainly be, but that's mostly irrelevant, in our application. The mass of the water in the system is.

The point that I was trying to make is that Head can be equated with flow resistance: all the elbows, twists and turns, the rad and, by far the most restrictive, the waterblock.

The other important point is that if you try to cram a high flow rate in a small tube size, the flow resistance actually increases. keeping the pressure at or under 5 psi will PROBABLY lower the resistance, hence the bigger tubing. But that again, is nothing compared to the waterblock.

In short, if you can figure that your rig has a flow resistance that can be calculated to 3 ft of head, then that 2 ft head pump won't do a thing for you, but churn water.

We need a plumber!

[pHaestus]

Yes it was a very well thought out article. I think that the take home message (for me) was that you can probably be doing more with the smaller pump you already have. Moving to a true 1/2" system (where the fittings are 5/8"OD and 1/2" ID) where possible and removing elbows to try and limit the pressure drops everywhere except the block will let you often get by with a smaller pump and still have optimum performance. Not sure what 11x6 radiator was used though; I wouldn't expect that large a difference if they were both similar heatercores.

Comments about head are good; People who went to larger Danner pumps in hopes of higher flow rates often were disappointed by temperatures. That was because the larger Danners still don't have that much PSI production and so all you are basically doing is forcing an even larger pump to produce more heat and dump it into the sstem (pumps generate more heat when throttled back with either a ball valve or with a lot of 3/8" barbs).

[myv65]

Brad,

That data is yet to come. There's still a few more articles left to cover radiators, blocks, fluids, you get the idea. After all of that there will be a wrap up explaining the stuff to think about when you put it all together and what to expect.


Bigben2k,

There's plenty of data on fittings available. Any engineering fluids text has them. Since those tend to be pricey, I often recommend an alternative source for the data. All you need to do is find a website that has the right tables. Most (industrial) pump manufacturers include engineering data in their catalogs and often online. Names along the likes of Goulds, Gorman-Rupp, etc., give you a pretty good idea of pressure drop vs line size and fitting types. These are all empirical numbers, but give us guys in the field a good start for estimating purposes.

In the world of small pumps, these charts lose some significance. The line lengths and absolute pressure changes are very small. In my experience and opinion, the only way to know your system flow is to perform the experiments yourself. This means measuring flow versus time with inlets and outlets at the proper locations to equate to the "real" system.

The guys have said it here, in a closed loop gravity has no real effect. Actually it has a large effect, but not on pressure drops. It's still responsible for keeping the fluid in a liquid state. Without gravity, all our fluids boil into vapor. Anyway, if you keep the inlet and outlet at the same heights they would be in the real system, then you can measure into a bucket what the flow really is. This is what Rich did for the data I used in the article.

Glad to see that most folks enjoyed the article. There's still a lot of ground to cover, but I'll get it all there eventually.

To all, I'm still open to any experience/data you'd like to share. Thanks for reading.

Dave

[Cyco-Dude]

Originally posted by myv65
Without gravity, all our fluids boil into vapor.
??? please tell me where you heard this! water boils away in a vacuum; gravity has nothing to do with it. when you lower the pressure, you lower the boiling point.

[UnaClocker]

I put this thread in the Liquid Cooling section for a reason. I didn't want it moved. There was already a thread covering this in the News Section, nobody was commenting on it there.

[SCompRacer]

Originally posted by pHaestus
Not sure what 11x6 radiator was used though; I wouldn't expect that large a difference if they were both similar heatercores.




It was a thick radiator style aluminum trans cooler (http://www.enteract.com/~richol/Water/transcooler.jpg). It had 5/16" OD inlet fittings when tested. After determining there was sufficient material in the tank ends, I cut them off and drilled/tapped it for 1/4" NPT pipe thread. Have not retested it for flow yet.

[myv65]

Originally posted by Cyco-Dude

??? please tell me where you heard this! water boils away in a vacuum; gravity has nothing to do with it. when you lower the pressure, you lower the boiling point.

I have a question in response to your question. Why is there an atmosphere on Earth?
.
.
.
.
.
.
.
Answer: Without gravity there is no atmosphere on earth. The air pressure that we have at any given elevation is nothing more than the gravitational force (weight) of the air column between that particular elevation and the "edge" of our own atmosphere.

It isn't something you'd normally think about, but is true all the same.

[Cyco-Dude]

well, theres no gravity in the space station; youre telling me water boils away in a space station? :p

[myv65]

Actually, there is gravity on the space station, but that isn't why water doesn't boil away there. I'm sure you also know there's gravity on the moon, yet no atmosphere. The space station merely falls at a rate equal to its movement away from the earth, netting no change in altitude. As I'm also sure you know, stuff doesn't boil away there because of the (artificially) pressurized capsule. The moon has insufficient gravity to hold onto any gases.

Maybe I'll have to toss in a little data from water's phase-change diagram in the fluids article. . .

[bigben2k]

CD, I think that myv65 is trying to say is that without gravity, there would be no atmosphere, the result of which is space vacuum, which will boil away any liquid.

The moon, even though it has gravity, doesn't have enough pull to retain an atmosphere. I've never understood that, but I do know that it's a fact.

[Cyco-Dude]

lol, i know what hes saying, and i think he knows what im saying (you can have pressure w/o gravity). this has gone way off topic i think hehe :p

[Avatar28]

Originally posted by bigben2k
CD, I think that myv65 is trying to say is that without gravity, there would be no atmosphere, the result of which is space vacuum, which will boil away any liquid.

The moon, even though it has gravity, doesn't have enough pull to retain an atmosphere. I've never understood that, but I do know that it's a fact.

Actually, the moon DOES have an atmosphere. It's just very very VERY thin. Enough so that for all practical purposes you can say it doesn't have one.

[pack rat]

Please be advised that I own PATTENT for GRAVITY US.PAT# 5,167,329 I also own all the gravities in the galaxy ZQ-153 from which I have emigrated.

[DodgeViper]

Through my experience in my setup, Eheim 1250, Gemini High Volume W/B, Chevy Chevette heater core, and 1/2” silicon tubing. The Eheim pump delivered 190 GPH at a 12” rise. Once the block was placed inline the GPH dropped to 64 gallons per hour at 12” rise. Next was to place the heater core inline. Now the GPH has dropped to 60 gallons per hour.

The inlet side of the pump I placed a 90 degree barb. This has caused a restriction to the flow, but the biggest restriction lies in the water block. With only 60 gallons per hour of water moving through the system I would have to agree that bigger is not better. Heat is being added using a bigger pump. I wish I had a Ehiem 1048 that consumes 10 watts to test against the 28 watts that the 1250 puts out. I have found that leaving the pump run 24/7 without any fans cooling the heater core and the computer turned off, the water temp will climb to over 140F degrees after about 7 hours. This causes stress on the CPU as it will take 4-5 minutes for the CPU temps to stabilize to the normal operating temps. If your overclocking you may encounter that your computer may not even post because of the higher CPU temps at start up. This is the reason I use a relay to turn the pump on and off. It has been said turning the pump on and off damages the pump. This is B/S. A magnetic drive pump has one moving part.

Currently my system is totally enclosed and my delta temps are 8c. I am using a push/pull rad/fan setup with the fans blowing onto the pump. Just behind the pump I have 2-80mm fans pulling the air from the case.

[pack rat]

the best result for pumping capacity is to have water reservor level even with CPU. I have seen mazes that are actually heat exchangers for rising temperature of water. It is simple , the higher temp difference for water, the higher the CPU temp. For radiators (all things the same) higher water flow lower temp. difference. Greater temp difference between air and water better performance by the radiator, that is why radiators in parallel perform better than in series. There are self-priming pumps that will "pull " water from under its suction pipe, it does that by pumping air thereby creating a vacuum that pulls water. Ehiem pumps do not have that ability. Any addition to the water such as anti-freeze, water-wetter will alter total head on the pump and performance, also temp of the water has miniscule influence. Also performance curve of the pump is not linear in respect to changes in discharge pressure offered by the piping on both discharge and SUCTION especially.