Topic: [DESIGN] Custom-Built Lexan / Aluminum Mining Case - page 3. (Read 7424 times)

On my Sapphire 5830s it blows quite a bit stronger out the back end of the card than the front (with the grill). So I would have though pulling from that end would be better than blowing into it as wouldn't that go directly against the direction the cards fans already blow? I'm curious if you tried it with the back end fans reversed and saw what difference it makes?

Wouldn't you want to get cool air into the center of each card where it intakes rather than directly against the hot air coming out? Just curious how much you experiments with this.

I thought about this too.  I was considering a very similar setup but with isolated GPU 'slots' - basically a box around each GPU.

You can cut a slot in the lexan/acrylic/whatever for the PCIe connector to poke through, but otherwise it keeps each GPU from sucking down it's neighbor's heat, while efficiently exhausting it's own.

Such a thing would work pretty well even in a 'traditional' closed case as long as you can push enough air through the GPU channels.

That's badass. Might get better airflow around the gpu if you sealed below the gpus as well

That's awesome. I'd love to know more about the script and if you plan to share it. I'm running Ubuntu with very similar setup but my rig is mounted in a wire shelf unit that cost $8 here. Maybe next phase would be something like this.

thank you . the 6 fans blow inwards, onto the cards. raw materials were roughly $80, more or less. I think ~$25 for the lexan, ~$30 for the aluminum, and the rest for hardware (nuts, bolts, screws, velcro, etc). I wasn't trying to be as cheap as possible. I wanted to use quality materials and have an extremely solid result, yet not waste money or materials on anything unnecessary. you can pick the case up and shake it, or turn it upside down; nothing moves.

thanks! a script runs once per second to measure the temps and fan speeds. if a card's temp is 68-72C then its fan stays put, otherwise the fan is increased or decreased by 1%. fans range 40% (night) to 60% (day), occasionally reaching a peak of 70-80% on especially hot days. this means the temps (first column, green) remain reasonably constant, while the fans (second column, blue/purple) vary as the ambient temp changes.

here's the console output of the script:

looks awesome! do those 6 fans blow at the cards, or do they exhaust the heat? how much in material did this build cost you?

looks awesome!!! what are the temps like?

hi!

I've had a mining machine running for a few months, and I thought I'd share some images. I've benefited from many ideas, designs, guides, and photos from this forum! so, this is the third iteration: v1 was 2x5830s in an Antec 300 case, v2 was 4x5830s in an improvised open frame, and v3 (below) is 4x5830s in an aluminum and lexan (polycarbonate) case which I actually planned out before building it. in a nutshell: v1) ran fine but had no room for 4 cards and v2) ran too hot, required external fans, and was huge. my goals for v3 were: a) no external cooling and b) smaller than 12" x 12" x 18". plans were made in google sketchup first.





the frame is 1/16" x 3/4" aluminum L-shaped strips, and the siding is 3/32" sanded lexan. the frame is held together with #8 machine screws and a combination of hex nuts (internal) and cap nuts (external). the panels have 1/2" holes which can fit around a hex nut, yet be secured with a fender washer, so they can be removed without taking apart the frame or needing to get inside the case; they're also secured with sections of velcro.

here are a few more pictures; the last one was taken with the top panel off.









after much (too much) trial and error, I'm running debian wheezy (testing) with catalyst 11.6 and app sdk 2.5. through a combination of mostly python and rrdtool, I've built a visual monitoring system for the machine:



anyway, I'm happy to go into more detail about any of this stuff; feel free to ask!

take care, -dan