MCW5002-AT - Just 8-10 sub ambient without condensation worries?

[DDogg]

Edit/Add: I did the groundloop, see below for pics.

I'm looking for just a 8-10 degree sub ambient solution without any worries of condensation. I don't know if theses questions can be answered, or have been before, but hope some replies might help my learning curve. Most of the stuff turned up with search seems to all be about sub zero cooling. I only need 8 c less, not 50 . On to the problem and question.

I can comfortably run 2860 MHz at 2.125 bios VCore when my inlet temps are around 22-23c to my 6002-A. Unfortunately, this room stays about 27-28c which causes my daily inlet temps to a 6002-A to be around 30-31 after several hours of max CPU temp stabilization on my daily setting. SiSoft show 168 watts. ['comfortable' to me is =< 48c]

Since I know inlet temps of 22-23c will work with SiSoft 195-200 watt dissipation, I've been searching for a 8-10C sub-ambient solution to accomplish that, or it's equivalent.

Q> In your opinions, can a Swiftech MCW5002-AT, attached to the CPU with a variable voltage controlled by an electronic thermostat (reading the outlet temp), drop the CPU the approx. temp as would an inlet temp of 23c into my 6002-A without causing condensation worries at my Dallas, TX dew-point?

It may well be that dumping the rad and just using a ground loop will better accomplish what I want to do. Those temps seem to stay around 21 C, but before I start digging trenches in the flower beds, I had wanted to ask opinions from some of the experienced members of the forum about a variable voltage TEC solution.

PS - Anybody familiar with http://www.hydrocool.com/Hydrospin.html

/Add: Ideally I was most interested in an affordable "on demand" solution, triggered by inlet or outlet temp, that could be spliced into my existing water circuit. There does not seem to be one except something like an aquarium chiller which is bulky and I would think noisy?

[The Dark Hacker]

are you sure that the ground stays that cool during 90+ degree days. i dont really know. but you will most likley get condensation here because the humidity is uaually 50% +

[DDogg]

Originally Posted by The Dark Hacker

are you sure that the ground stays that cool during 90+ degree days. i dont really know. but you will most likley get condensation here because the humidity is uaually 50% +

My understanding is that 3 feet down the soil temp is fairly constant. We don't have to worry about a frost line so it might be doable. My PC is located such that it lends itself to doing it in a neat, clean way that would not be unsightly. Putting copper through the brick could cause serious domestic problems though. Have to think on it some more and weigh the risk/reward Presently I don't know how to calculate the length of copper required to exchange 200 watts with a 21 C thermal sink. Never was too good at that stuff.

As for condensation, I'm just guessing from looking at the weather maps, but I would think the humidity in my computer room would never be so high as to cause condensation with 8-10 below ambient room temp. I may be wrong.

While a tec block like the swiftech would be my simplest and affordable solution, I am realizing that it may be too difficult to control so as to not have condensation. Seems like my only solution is to chill the water somehow, or using a ground loop. I would prefer to get some of this heat out of the room and the ground loop would do that. My kid's, and my wife's computers are also in here so you can imagine how much heat gets dumped in here.

I know Swiftech makes that big monster twin TEC external block, but I also think I remember it takes 40 or 50 amps. That's over the top. Heck, I just need 8-10 C lower inlet temps. Seems like there ought to be a way that was not too expensive to buy or operate.

I expect I have a lot to learn and a rude awakening in store for me.

[The Dark Hacker]

well if you are going to dig down threee feet it should be ok, but you will need to go rent a trencher or something to dig a long hole that deep. you will break your back digging in texas soil with just a shovel and other hand tools

[bobkoure]

Originally Posted by DDogg

I'm looking for just a 8-10 degree sub ambient solution without any worries of condensation.

You could just check your weather historical data to see if the dew temperature is ever within 8 or 10 degrees of ambient. You're in central Texas? Looks like your dew point is about 60F/15C about now (from a Weather Underground dew point map - might show something different by the time you get to looking at it). So you can cool down to a little above 60F without condensation - outside, tonight. Humidity inside is likely higher, and so a higher dew point...

[ferdb]

Sounds like a lot of effort when you could just crank your clock speed back about 1% and be done with it. However there is always the entertainment value of building new things.
Good luck with your project

[Kobuchi]

If your home is serviced by an underground water main, just run a tap a minute to learn what the temp is at that depth.

Originally Posted by DDogg

interested in an affordable "on demand" solution, triggered by inlet or outlet temp, that could be spliced

Periodically cycle water from a buried reservoir through a case reservoir? A relay could switch a utility pump located in the pit. The (big cheap noisy) pump would only stay on long enough to refresh the loop. This solution will cost your household the least kilowatt hours but may break your back to set up. Personally, I think the long-term energy savings is worth it. A welcome bonus: this powerful secondary pump could run through a very restrictive filter.

[Seyeklopz]

You can find a database of ground temps for different regions, but you don't need to worry about that, it's usually pretty cool, especially in places with frost heavng in the ground. I've seen a couple cooling systems like this. It's best if your soil is more like clay or compressed sand. Fluffy soil isn't very conductive. Try to go down 6' and put some sort of heat spreader on or around the pipe like some copper roof flashing.

[davekusa]

I had the mcw 5000-pt. The block are well insulated and condensation is not an issue. I was running anywhere from 15C to 20C below ambient and never had any condensation problems. Stick with swiftek for a TEC block. Its a quality product and you might want to wait and see if they come out with a MCW-6002-AT????????????

[Seyeklopz]

I will be building my own MCW-6002-AT. I'll post it as I work on it. The block should get here in a week, and the pelt in 1-3 weeks. I'm also trying to find my own copper for a cold plate locally.

I won't be able to test it much until the other WC mods are finished and the whole circuit it tested in the new case.

[DDogg]

Thanks for the replies. I did the groundloop - Got some pics for y'all.

I got industrious today, but my back is just killing me. Found 60 ft of 3/8 copper on sale at Home Depot for 28.00. 1/2 was 60.00 bucks, so I went with the 3/8.

Basically I butterflied it into two 30 foot coils, and then ran 3/4 rigid copper supply lines with the coils in parallel. My little fountain pump didn't like trying to push though 60 feet, but it was OK with this setup.

Also, this allowed a 50 year old guy that smokes too much to dig the hole which was only 5x2x4 feet deep. I stopped when I hit muck. It only took me a little over an hour, but this was a flower bed and most of it was good sandy loam. Digging was pretty easy until I hit the muck. There was no way I was going to try to dig a 50 foot trench, even with a trencher. I put in an 18 zone Sprinkler system 10 years ago and I have 1 in pipe all over this place, except I kinda forgot exactly where it all is :-).

Plus I thought it might be useful for other folks to see you can do this fairly easily if you have diggable ground. I did the whole job in a little over 5 hours from machine off (to drain) to reboot on the groundloop. The costs were low. I think I dropped about 75-80 bucks all told including a new roll of solder, flux and a propane bottle.

Anyway, it seems to works great. No rads and no fans, except the normal case fans. Man the silence is nice. The room is so much cooler.

Measured incoming using another thermocouple than the probe I used outside [which showed 21C - See pic - squint a little] so I don't know if they jibe. It shows 25C inlet temp, which doesn't agree with the exterior reading of 21C I took. /Add: [I have near 15 feet of clear 5/8 od tubing from the ground loop to the point I am measuring inlet temp - I'm wondering if that could be sucking up some room heat?]

The good part is that the temp does not seem move...period. I've been throwing 2.150 VCore at it for a couple of hours now and it just stays at 25C.

SiSoft show 200 watts as the heat being dissipated. Don't know if that is accurate, but it should be portable reference number to anybody reading this for comparison.

It is too early to completely bless it. I need to throw a lot of heat at it for several days to make sure those two 30 foot butterfly coils can dissipate it OK. So far, it seems very encouraging.

What is not encouraging is I seem to have caused a slow leak when I was covering up about half the hole. Don't know what I can do now that it is down in the muck. There is no way I can R&R the thing. Maybe I can dig down enough to try. Big sigh on that one. Got in a hurry. I should know better by now.

/EDIT: JOY, YEAH! - I found the leak which was in my res. I had twisted a fitting O ring when I was replumbing this thing. Wow, you can't imagine how bummed I was thinking nearly all this work had been wasted. Man, like the best christmas present ever in my head to find that leak was inside the house, instead of 4 feet down in black muck.

[The Dark Hacker]

where in sisoft does it estimate your heat dissappation

[DDogg]

Originally Posted by The Dark Hacker

where in sisoft does it estimate your heat dissappation

CPU and BIOS information icon. Scroll down to the cpu temp and true voltage.

[Kobuchi]

Geothermal cooling. I love it.

Now mind the wife doesn't stomp a retaliatory shovel blade through it.

The runs to and fro will benefit from insulation. This may protect them from accidental damage too if not in the wall.

[Seyeklopz]

I hope it doesn't hurt your heating bill...

[DDogg]

Originally Posted by Kobuchi

Geothermal cooling. I love it.

Now mind the wife doesn't stomp a retaliatory shovel blade through it.

The runs to and fro will benefit from insulation. This may protect them from accidental damage too if not in the wall.

Kobuchi, I hope she never even notices it or she might take the shovel after me. Um, she just doesn't understand about these things, ya know (chuckle). Amazing how a 6 feet 220 lb man can be so scared of a 5-5 125 lbs pissed off women wondering what the new pipes through the brick is :-). Ha, I'll tell here the cable guy did it. She might buy it.

Yep, I think I'll get some of that black split foam stuff and put it on the risers outside. Hmm, maybe inside also. There is about a 6-7 foot run along the back wall before the tubes come in to my desk. I may stick one on the incoming. Thanks.

====

Seyeklopz, heating is not too much of a concern in Central Texas, except for a couple of months of the year and then many times only for a day or two here and there. My A/C has been running all day and it is October. The heat has never been on yet and it may not be for another month or two. Hmm, I don't even think the pilot light is lit.

[DDogg]

Update:

As I was retiring around 2 AM, my inlet temp showed as 27C. This was after throwing what SiSoft estimated as 180-200 watts into the ground loop for around 9 hours. I don't know if the thermistor I'm using is actually reporting the correct temp, but for comparison purposes that was a rise of 2C from my freshly filled starting point 9 hours previous.

Before going to bed I setup the machine at 2.00 actual VCore as reported by 8rdavcore. This was a bios setting of 2.1 volts VCore. SiSoft showed 180 watts estimated wattage. I set it running Prime95. I has some worry that the temps might continue to rise above 27C.

This morning, 6 hours later, I was pleased to see Prime95 still running and the inlet temp was still at 27C.

Given that the temp did not rise in 6 hours after a 9 hour previous input of the approximate same SiSoft estimated wattage, I think one could then start to conjecture that this small butterfly ground loop can dissipate the estimated 180 watts reported by SiSoft.

As I said before, the SiSoft number may not be the real wattage, but I would think it would serve as a portable comparison reference for others. Comments on that, or suggestions for properly measuring the wattage input to the ground loop?

I would like some help on how to pin down the loops max dissipation potential in case I use it to sink the heat from a TEC in-line chiller.

[bobkoure]

Originally Posted by DDogg

...heating is not too much of a concern in Central Texas, except for a couple of months of the year and then many times only for a day or two here and there. My A/C has been running all day and it is October....

So your major power usage is for AC? Sounds like maybe you could present this to the wife as a means of getting PC heat outside in some way that might be cheaper than paying for AC to pump it out. She might totally approve.
FWIW, there are heat pumps designed specifically to work with water pumped up from a well (so geothermal heating/cooling). I looked into these for the last house I built. For someone in the northeast, they almost make sense (slightly better on energy, but higher equipment costs to buy/maintain) but I'm a lot north of you. Your wife might get very interested in the energy/money/ecology part of this - and at which point you're spending your valuable free time building a "pilot project").

Oh, and BTW, the Romans used to cool buildings using geothermal - they'd put long clay pipes underground, leading into the building (source of cooler air) and a cupola-like-thing on the roof, which heated up and drew air through via thermosyphon. Of course, we both live in harsher climates than the Romans did...

[lolito_fr]

Interesting experiment
The question is not "could it dissipate that amount of heat", but rather how hot would it get?
At what point do you consider it to be no longer "dissipating"?

Predicting a temp is actually fairly easy. If you measured that the inlet is rising by 6°C for 200W (27°C-21°C), that gives you 6/200=0.03°C rise per W. So if you feed 400W in, you get 400*0.03=12°C rise (above the sink temp of 21°C). Which gives an inlet at 33°C.
All assuming that your system has actually reached equilibrium, ie. temps aren't going to go up any further(?)

edit: bob, my room is cooled with geothermal (13°C well water + large rad + nice big blower - best thing I ever made)

[bobkoure]

Originally Posted by lolito_fr

edit: bob, my room is cooled with geothermal (13°C well water + large rad + nice big blower - best thing I ever made)

Very cool - although it's not cooling season in France right now, is it? Even in southern France, I'd have thought this was mistral season...

[redleader]

Will you eventually saturate the gound around the loop? I mean you're basically tapping a stored resiviour of colder material. Eventually you should use it up right?

Or is the ground conductive enough that it doesn't heat up much around the coil?

[Kobuchi]

Originally Posted by DDogg

she just doesn't understand about these things, ya know

That's why we do it.

***

You stopped digging when you hit hard "muck". That's saturated clay, I guess, great stuff for wicking up deeper water and staying cool. If your coils aren't actually buried in it, changes in soil moisture will show up in MBM. Water the flowerbed in dry months. Wife take note.

[lolito_fr]

I'd have thought this was mistral season...

We don't have that in the southwest (luckily) You're right though, it shouldn't normally be 24°C this time of year

redleader - guess that's the main flaw in my calcs. (assuming that the "sink" temperature will remain at 21°C).

[DDogg]

Inlet temp is still at 27C. I'm starting to feel real good about this.

Originally Posted by bobkoure

... She might totally approve.
...

She would look at me and say, "Good try, dear, that was really some lovely BS - Quite inventive" :-D

Originally Posted by lolito_fr

... The question is not "could it dissipate that amount of heat", but rather how hot would it get?
At what point do you consider it to be no longer "dissipating"? <snip>

Yep, I take your point. Perhaps I should have phased it more along the lines of - How much heat can be added until the ability to sink that wattage is inadequate and the temp can no longer be held at 27C. Something like that.

Originally Posted by redleader

Will you eventually saturate the gound around the loop? I mean you're basically tapping a stored resiviour of colder material. Eventually you should use it up right?

Or is the ground conductive enough that it doesn't heat up much around the coil?

As far as using it up, seems to me you would have to heat up the whole planet. Its all one contiguous ball of heat sink material ;-) Technically, I guess, my heat is being sinked into the polar caps as they are the coolest points on the earth (I think).

I'm sure I'll eventually find a point where the surrounding ground can't wick the heat away from the coils fast enough to hold a constant temp as it is doing now. Still, that would only cause my temps to move up very slowly and re-stabilize at a slightly higher constant with that specific load. [lolito_fr's point, I think]

The coils lay in a 2x5 feet square. 10 square feet (sorta-kinda 'cause they are coils) hooked up to the planet is one big mama heat sink. I don't think my piddly 200 watts is going to bother it too much.

Originally Posted by Kobuchi

You stopped digging when you hit hard "muck". That's saturated clay, I guess, great stuff for wicking up deeper water and staying cool. If your coils aren't actually buried in it, changes in soil moisture will show up in MBM. Water the flowerbed in dry months. Wife take note.

Around here when you say muck, you are talking about black stuff that squishes up between your toes, and it sure ain't oil :-) We have a low water table this time of the year.

[killernoodle]

DDogg melts polar ice caps to cool computer. He is sorry.