Hotter than I like...

[BachOn]

I began with water cooling about 10 or 12 years ago. I went back to air cooling for awhile, but the truth is that water cooling gets into your blood. So when I built my latest computer, I decided to head back to the water.

I used a combination of older and newer components to build an external water cooling system. My mid-tower computer sits on top of this box. The tubes run in and out the rear of the computer case through one of the expansion slot panels. All of the other components for water cooling are in the external box.

I've tried pushing my Core I5 2500K CPU a little harder. I can get it up to 4.5 without changing any voltage, but it just doesn't seem to be stable. I was seeing idle temps running about 40 - 42 degrees C, which seemed too high for stock voltage on a water cooling rig. Running Prime was pushing the temps higher than I liked and I kept getting the occasional Blue Screen Of Death. And the water temps inside the reservoir in the box kept gradually climbing. I was routinely seeing temps of 48 degrees - which is just too hot to effectively cool a CPU. So I figured something in my water cooling setup still needed work. BTW, I have a digital meat thermometer with the probe inserted into the water reservoir.

My water block is a D-TEK Fusion II. I guess it isn't the absolute best block around, but it certainly should be able to adequately cool a single I5 CPU. The silicone tubing is 3/8 inch ID. I have a good sized Eheim aquarium pump that is more than equal to the task. I have some type of Arctic Silver paste. I favor a little more, rather than too little of the paste.

I reasoned that the two fans were doing a reasonably good job of cooling the radiator. The air exiting the radiator was quite warm to my hand. My radiator is an older converted auto heater core that was sold under the name "The Big Momma". It's 6x6 square and about 2.5 inches thick. It was always one of the top contenders from a few years ago. I have two 120mm fans within a shroud to push-pull air through that sucker. Each fan is about an inch and a half from the radiator surface. And they seem to move the air through it effectively.

My theory was that too much of that warm air coming from the radiator is being trapped within the wooden box. I did a temporary experiment using a 3 inch fan at the rear of the box to remove more of this inside air. The temps on my CPU went down by at least 4 degrees C at idle. And the temporary cardboard panel I used for this experimental fan was nowhere near tightly attached to the box. Even with only a rated 30 CFM air flow, the inside of the box was defintely cooler. So it looked like getting more of that hot air out of the box might improve the cooling of my CPU.

So I bought a third 120mm 120 volt fan. It is supposedly rated at 65 CFM. I'm now constructing a labyrinth air mechanism I favor to reduce sound while I try to get that excess heat out of the box.

These are just cheaply made Radio Shack 120 volt sleeve bearing fans. I say cheaply made instead of cheap. The Shack gets $26 per fan. I already have the two radiator fans on a controller. I use one of those rotary fan controllers like you use for a ceiling fan. I've decided to put this last exhaust fan on a separate controller. I already had one in my junk box from a previous project. This should allow me more options to regulate things. My logic is that having more controls is usually a good thing.

If anybody can think of other things for me to consider, I'm open to ideas.

Bach On!

 

[Conumdrum]

Have you opened the CPU block to check for gunk? It happens. And increase airflow, your heater core with the right fans is plenty, but you could have other issues stopping you from getting good temps.

 

[BachOn]

No, Big C. I didn't pry the water block off and check the paste on the CPU. It's only been on there about a week. I guess I should double-check it. Let me do the other project first. Then I will.

I'm curious: what kind and how much paste do you use on your CPU? Most of the reviews of the various types of pastes I've read are only showing 1 or 2 degrees of difference. Radio Shack sells Arctic Silver. But like Windows, A.S. has gone through several generations of their product. Are the later versions that much better? And I've never found the pads were as effective as a good heat compound/paste. Tell me the truth. Is my thinking wrong?

BO

 

[Modz 4 Higher]

There are several data tests on different thermal pastes. What he is talking about is gunk build up 'inside' the block. In the channels the water actually flows through

 

[BachOn]

Duh!!!

What was I thinking? In designing the layout of the parts for my outboard watercooling box, I tried to think carefully about air flow. Makes sense, right? Not thinking it carefully through, however, was a mistake.

Example: I figured the air coming off the push-pull fans at the radiator should blow onto the reservoir to help cool it. WRONG!!! These fans are pulling heat away from the radiator. So the air these fans are blowing at the water reservoir is only making it hotter. Duh!!!

Now I've got to either reorganize the placement of the parts in the box, or see if I can somehow place a deflector in there that will shield the reservoir from the hot air coming from the rad.

Am I the only one who does dumb stunts like this?

BO

 

[m0r7if3r]

having hot air in your res will affect exactly nothing. also, even if the air is hotter than the parts its moving over, it can have a cooling effect due to the motion of the air alone.

 

[BachOn]

mOr7if3r,

Water from my water block returns to the water reservoir of my system. From there, it goes into the pump, through the radiator, and then returns to the CPU.

My reservoir is a high-impact plastic underground electrical cable junction box. It is roughly 4 inches square on each side. It has a top with rubber gasket that is held on by four screws. Though it is probably water tight, I keep it in an upright position. I guess the plastic is maybe just over an eighth of an inch thick. The outside wall closest to the radiator is now considerably warmer than the other walls. I'd say the reservoir is about two or three inches from the fan output off the rad. The heat from this air seems to be heating that wall.

I don't claim to be a brain surgeon or rocket scientist, but it would appear to me that some of the heat on this hotter wall of the reservoir would be transferred within and heat up the water ever so slightly.

What I am hearing you say is that this air is going to be cooler simply because it is moving. Air blowing on your skin may feel 4 - 8 degrees cooler C than still air. But if the heat level from the rad is 48 degrees C, the moving air - at best - would still be 40-44 degrees C where is strikes the wall of the reservoir. I'm trying to get my water closer to 35 degrees or less. Cooler water can absorb more heat from the CPU.

If my thinking is wrong, please help me (and maybe others) better understand the physics. I'm figuring every single degree drop in the reservoir temperature counts towards cooling my CPU a little better. Any clarification will be appreciated.

BO

 

[m0r7if3r]

I can't explain the physics of why air hotter than the surface it is moving over can cool the surface it is moving over, as I do not understand it myself. I was told that by a member of this community whom I very much trust, especially in regards to that kind of thing...you can believe it or you can not, it's your perogative to do so, and I woudln't blame you for thinking it was BS (I would too, were it not for the source).

The real issue here is that the amount of surface area in contact with the water is not enough (as far as I can tell at least) to cause any sort of temperature variation detectable by any of the equipment that costs less than your rig...especially when you look at it in comparison to the surface area of a CPU block or radiator.

My advice for you at this point, since you seem rather convinced you're right (and you very well may be), is to alleviate what you think the problem is and see if it ACTUALLY fixes anything. If it does, I'll totally eat crow...but I'll also be all kinds of interested in your exact setup so that I know what to look for in case anyone else ever has that issue.

 

[manu2b]

Sorry Morti, I had to

 

[m0r7if3r]

Fortunately, admitting I'm wrong is something I've never had an issue with.

 

[manu2b]

Joke apart, I believe you're righht. It has something to do with thermodynamics and convection...
I've read that somewhere on the internet a couple of month ago. Can't remember where...

But still love the picture

 

[QBobWatson]

What was I thinking? In designing the layout of the parts for my outboard watercooling box, I tried to think carefully about air flow. Makes sense, right? Not thinking it carefully through, however, was a mistake.

If you have several fans pushing air into or out of a box without enough vents then you can lose a ton of airflow due to unnecessarily high or low static pressure. I think you're right about not exhausting the hot air effectively. Can you just take off one side of the box to check?

If the air coming off the radiator feels hot to you then not enough air is moving through the radiator. The water should only be a few degrees C hotter than ambient.

 

[Conumdrum]

Have you opened up the waterblock yet? How long have you had it and used it without opening the block up? Just gotta know........

The air blowing on the res is cooler than the water in the res because:

Your water temp is hotter than the air, you have cooler air blowing on the rad, so there is a increase in air temps, but still cooler than the water in the res.

So in theory, yep.


That said, the poor thermal transfer property of the plastic res, the very low surface area of the res and the very small difference of the res exterior and the air coming off the rad means it will reduce your water temps by exactly .00000 Wait, I messed up. .0000000001C is closer to the actual number.

 

[BachOn]

Guys, I really appreciate the feedback. FYI: since I've been married over 36 years, crow is a regular staple of my diet. That's one of the first things a wife teaches her husband to eat.

I guess I was somewhat playing devil's advocate on this water reservoir issue.

1. The water block has only been in about a month. I've got to do remove it when I install the new Indigo Xtreme Thermal System. I'll perform surgery and do the examination on the interior of the block then - FedX says sometime this week. I had cleaned my radiator well before I put this thing together. I use distilled water with no additives. I don't see particles floating in the water in the reservoir. The pump is the big Eheim 1250. It has a pretty high flowrate. I was able to cushion it and position it so it doesn't buzz very loudly.

2.As I've mentioned, I use a homebrew labyrinth system to reduce fan noise out to the room. I'm pretty sure this can restrict air flow. I'm now working on a more permanent fan system for the rear of the box to remove excess heat. While I'm tinkering with that, I expect to take off the front panel (that's the location for the air intake) and check out what kind of impact this has on temps, both at the CPU and in the reservoir. If temps are significantly better, then a redisign or elimination of the labyrinth may be called for.

3. OK. You guys have convinced me that air blowing on the walls of the reservoir will have an almost insignificant impact on water temps. (Yum! Yum! This is great crow! I think it's the garlic and black pepper that makes the difference. But it does need a little more salt.) ;-)

The two 120mm fans are arranged in a push-pull pattern on the radiator. Since the rad surface area is more like 140mm, I used a shroud to increase flow throughout the fins of the entire surface of the radiator. Each fan is perhaps an inch or two away from the radiator fins. Each of these fans is rated at something like 60 CFM. But I've grown to distrust such manufacturers ratings. They're just four blade Radio Shack sleeve bearing products. Each is rated at something like 38 db. I have a fan controller that allows me to cut the RPMs down. But this is at the expense of cooling. Right now, they are running wide-open.

I'm not down in the dumps about this. Everybody needs a project and/or a hobby.

Bach On


One thought would be to use 140mm fans. But I can't seem to find those in a 120volt variety. I was trying to avoid having to run a separate 12 volt line inside the box.

 

[m0r7if3r]

Guys, I really appreciate the feedback. FYI: since I've been married over 36 years, crow is a regular staple of my diet. That's one of the first things a wife teaches her husband to eat.

Congratulations man! Not many people last that long!

I'll be interested to see what happens when you teardown, clean, and rebuild.

Someone else said this already, but they didn't quite explain it, so I'll try to rehash it, if you want to check whether or not it's the hot air that's an issue, open up the side of the case and point a box fan (or similar high airflow fan) into the case to move air through. If you see temps drop then you've found your issue (it should be pretty quick, maybe a minute or 2)

 

[BachOn]

Took off the front cover on my box.. The temps at the CPU and in in the reservoir went down significantly. So my labyrinth seems to be choking off the incoming air.

Looks like a redesign is in order. Back to the infamous drawing board.

BO

 

[Conumdrum]

Yahoo! Uber find! Looks like we have another satisfied winner!

 

[wagex]

Just distilled no additives, sounds familiar you probably have algae if its been running for a month like that trust me

 

[BachOn]

What additives are worth the effort and the money, wagex? From what I've seen here, some folks use them and others say they just aren't needed. What do you use - and why?

Big C, I'd hardly call it a "winning" find. But you are correct that it is at least progress. I'm more and more convinced that my radiator just isn't up to the task at hand. Evidence, I cut off the computer overnight. The water temp came down to about 27 degrees. The problem is that the room is more like 24 degrees. Doesn't that seem odd? I'd think the water should go down to room temp. Am I wrong in my thinking?

I'm now exploring options for a newer radiator. Any sage advice is always appreciated?

BO

 

[Conumdrum]

Water temp will go down to room temp, but it takes a long time. I'd say overnight, a good 10 hours. Water holds heat very well. Possible that your water measurement device could be whacked.