Journey to geothermal

Thick8 · Jun 3, 2015

EarthDog said:
Odd. That counter intuitive to all the testing people have seen... I wonder since you do not have a radiator in the loop that it needs more time in the loop to cool off...

I know. There is 40ft of flattened copper tubing in the aquifer so maybe by slowing the flow down I'm removing more heat from the CPU. Maybe it doesn't matter how fast the water is moving in the cooling manifold as all of the heat is removed regardless of the CPU temp. I'll have to do some more tests with everything in the loop.

Thick8 · Jun 3, 2015

So I ran the system full out. 100% pump speed, 5Ghz (200x25)@1.55v. Water temperature stayed at 21.7c and the CPU leveled off at 54c. I only ran it till it went stable (10 min). sitting here at idle for about 15 minutes the water temp is staying at 21.4c. It doesn't seem like the heat load has any effect on the inlet water temperature.

EarthDog · Jun 3, 2015

Typically it takes around 30 minutes or so (depending on the size of the loop) for temperatures to 'normalize.' Your loop is huge comparatively so it may take a lot longer than that to normalize. I would give it AT LEAST an hour or so before you see saturation in such a loop. Remember a typical loop has around 1L of water, yours has a lot more I assume?

Thick8 · Jun 3, 2015

After 1 hour of prime; CPU went up to 55c. Water temp to 22.4c. After shutting down the computer I ran the pump by itself for about 30 minutes. The temp stayed at 22.1c. I think the afternoon sun and heat load is effecting the readings. So I've pulled the PEX back through the window and insulated it 12' down into the hole, capped it off, and sealed it with silicone. Then I insulated the rest of it and ran it under the house (mostly). I need the wife's help to get it the rest of the way through. I should be able to run another CPU test tomorrow with some of the variables removed.

Mandrake4565 · Jun 3, 2015

Thick8, I love it. A Fx 8350 at 1.52-1.54V staying at roughly 55c is not shabby at all. I'd bet you're drawing close to 300W at that voltage. What are the package temp readings BTW? Shame you don't have colder ground temperature!

GTXJackBauer · Jun 3, 2015

Mandrake4565 said:
Shame you don't have colder ground temperature!

My guess is he'd run into some condensation so he might be just at the right temps.

Thick8 · Jun 3, 2015

Mandrake4565 said:
Thick8, I love it. A Fx 8350 at 1.52-1.54V staying at roughly 55c is not shabby at all. I'd bet you're drawing close to 300W at that voltage. What are the package temp readings BTW? Shame you don't have colder ground temperature!

Extreme power supply calculator says 225W at 100% load.

Mandrake4565 · Jun 3, 2015

GTXJackBauer said:
My guess is he'd run into some condensation so he might be just at the right temps.

Likely but we can defend against condensation.

Thick8 said:
Extreme power supply calculator says 225W at 100% load.

LOL Imagine that, here's mine running Prime 95 at 4.7 with 1.46875 V. Taking into account I do have 12 120 mm fans but they were not at 100%. This is power draw from the wall, the Kill O watt isn't 100% accurate and you need to take into account the Psu efficiency, which is a Seasonic X 650 Gold unit. That said 225w when Oced, is laughable with these chips. Here is a real eye opener, running Prime 95 and Heaven at the same time, no it's not realistic but it just gives you an idea of how much power these things draw, at 5.2 with 1.5375v and my GTX 580 Lightning Oced, on a 1000w Gold Psu 827w at the wall.

t1nm4n · Jun 3, 2015

I have a question about all the water you had in the pictures when you were dropping one or the pipes, had it been raining there, or was that water that came up from the water table under your house? I mean depending on where in SC you live, it wouldn't be shocking to hit water at 25ft or less, so you might never actually saturate your system. I would be interested in how it does on an over-night stress test though. You got both your cards on the loop I take it, have you added them into the stress test yet if so?

Regardless, I'm really looking forward to hearing how bad-*** this is :clap:

, I think I'd have to go down about 50ft where I live in Texas, and the thought of digging in this rock infused black-clay is just not a happy thought, I envy you.

Thick8 · Jun 4, 2015

t1nm4n said:
I have a question about all the water you had in the pictures when you were dropping one or the pipes, had it been raining there, or was that water that came up from the water table under your house?

That's the water that I'm pumping into the pipe. The water goes down the pipe and the soil/sand/clay comes back up with the water on the outside of the pipe. If you look at the original pictures you can see that I cut a saw-tooth pattern in the pipe to aid in the cutting of material. The trick is to take it slow in an up, down, twisting fashion. Dig down about a foot and work the pipe until the water coming back up flows clear and then cut down the next foot. Took about 3 hours to get the 3" pipe to 35'. Grated I dug a water well a couple of years ago that pumps 12.5 GPM so I used that and water coming in from the street so i was shooting about 15 GPM down the pipe. You'll notice three hoses coming from the water head. 2 are from the well pump and 1 from the house.
Note: Never leave the pipe in the ground without the water flowing unless you plan on leaving it there forever.

t1nm4n · Jun 4, 2015

Ok, that makes sense, I sort of thought you just hammered away (literally) until you got it to the desired depth, was wondering how PVC could survive that much pressure.

Thick8 · Jun 6, 2015

EarthDog said:
Typically it takes around 30 minutes or so (depending on the size of the loop) for temperatures to 'normalize.' Your loop is huge comparatively so it may take a lot longer than that to normalize. I would give it AT LEAST an hour or so before you see saturation in such a loop. Remember a typical loop has around 1L of water, yours has a lot more I assume?

I just did the calculation and it's just over a gallon. 120' long with a 1/2" diameter. I thought it would have been more than that.

EarthDog · Jun 6, 2015

I was thinking a bit more myself.

A gallon compared to a liter is a big difference. Not sure if it would scale linearly, so I cant pin point it, but would imagine a couple hours and it would saturate the loop.

GTXJackBauer · Jun 7, 2015

That's really impressive to see the Apogee II (MCP-35x) pump do all that work.

Thick8 · Jul 9, 2015

So it's been a busy few weeks. The honey-do list is actually a binder.
I've mounted the GPU coolers. unfortunately I ordered the wrong crossfire connector so that's on order now. I've misplaced my MOSFET cooler so I will have to dig that up tomorrow. I was looking at the GPU cooler ports and I was thinking that it might be better to hook them up in parallel rather than in series as it doesn't look like they can flow 1 GPM. Thoughts?
Here's a pick of the mounted blocks. Wow these things are heavy.

Thick8 · Jul 25, 2015

IT'S ALIVE!.
Just leak testing it. Should have the PS wired in tomorrow and then I'll post some numbers. I sure hope this works...

Silver_Pharaoh · Jul 25, 2015

PEX FTW!

Hope it doesn't leak!!! :cheers:

Can't wait to see temps under load for a few hours

Mandrake4565 · Jul 25, 2015

Thick8 said:
IT'S ALIVE!.
Just leak testing it. Should have the PS wired in tomorrow and then I'll post some numbers. I sure hope this works...

Love it Thick8 can't wait till it's up and running.

trents · Jul 25, 2015

Johan45 said:
Ground temperature varies by location E_D so someone in a more northern climate will have a better result. Here's something I found for an illustration it's in Fahrenheit.

View attachment 164414

But we aren't talking about surface ground temps, Johann45. We are talking about ground temps at what, 25 ft. down? At that depth, you should get a constant year around temp of about 50-55F year around anywhere in the world. Of course, that would only apply to the far end of the water line and the sections of it nearer the surface would be warmer. Insulating the part of the loop that enters the ground for the first 5' or so would lower the temps, I would think. And when he puts the section of the loop that is now running up the side of the house under the house that should also help a lot. Even at night when the sun is not beating on the tubing the air temperature in that part of the country is still pretty warm this time of year.

Thick8 · Jul 26, 2015

I'll have some pics and updates tomorrow.
CPU: 5007Mhz (217.7x23) 1.5250v
GPUs: 1175/1475 MSI afterburner voltage @ +88% and Power @ +50%
Pump speed: 20%-50%. Raising above 50% (3500 rpm) doesn't make any difference in temps; it just gets louder (4400 rpm).

So far the CPU never goes over 56c and the GPUs never get over 46c. Furmark with both the CPU and GPU burn-in tests running at the same time with all the heat generating stuff turned on. I ran it like that for about hour. Water temperature maxed at 26.4c. My only temperature concern is the vcore-1. It's getting in the mid 70s range. I've done some research and this is a common concern with this board. I may pull the vram block to change out the thermal pad with TIM. Or I may cut a hole in the bench case to put a BeQuiet fan under it. Or I may just leave it as the temp only ramps up when everything is maxed out.

Oh yea. Did I mention I can barely hear it running and the room doesn't get warmer.

Journey to geothermal

Member

Member

Gulper Nozzle Co-Owner

Member

Mr. Clean Senior Member

Water Cooling Senior Member, #TEAMH20HNO

Member

Mr. Clean Senior Member

Member

Member

Member

Member

Gulper Nozzle Co-Owner

Water Cooling Senior Member, #TEAMH20HNO

Member

Member

Attachments

Likes the big ones n00b Member

Mr. Clean Senior Member

Senior Member

Member

Similar threads