[artrafael]

I like how it can be used as a drink mat.

Yes that's what I am doing with mine

Just be careful not to spill your drink! On the other hand, if you overclock your RPi, you might be able to use the top of the case as a coffee mug warmer.

[JasonQsine]

its got seating for the cpu but isn't there 3 area of concern for heat?

http://nwazet.com/Media/Default/Product/...ayback.png

Are you using over voltage? The Rpi we got doesn't heat up as the pictures.

[Milhouse]

its got seating for the cpu but isn't there 3 area of concern for heat?

http://nwazet.com/Media/Default/Product/...ayback.png

Are you using over voltage? The Rpi we got doesn't heat up as the pictures.

There was a design error in v1 Pis that caused the LAN9512 USB/Ethernet chip to run hotter than intended as it was acting as a 1.8v regulator. The lack of mounting holes suggests the above thermal image is of a v1 board with the LAN9512 chip using too much power and consequently running too hot. More discussion here, SMSC is the manufacturer of the LAN9512 part.

Here's a thermal image of a v2 board (note the mounting holes) - the LAN9512 now runs slightly cooler than the SoC:

[JasonQsine]

Yes the board we got is v2. We ran some test on it. Overclocking does heat up Ar1 chip quite a bit, but the power regulator and ethernet chip temp is much lower. When we tried over voltage there was warning it will cause error in the SD card, and it did. We got 30% higher performace but SD card got error and can not be used to reboot. The temp was stable with the new alum enclosure even the Rpi is over voltaged.

[The_Doc]

Form over function!

Firstly, I do like the look of the case... but if you're going to design a case for the Pi, and you're going to machine it from aluminium, then it should be machined in a manner that facilitates cooling of all the hot spots. Granted, your design provides a cooling solution for the CPU but the CPU isn't the only part that gets hot, especially if you're overclocking/overvolting your Pi, which I presume a lot of your potential customers will be doing. This case should also be in contact with the other 2 hot spots, and thermal pads should be fitted to the contact areas.

[awp0]

Form over function!

Firstly, I do like the look of the case... but if you're going to design a case for the Pi, and you're going to machine it from aluminium, then it should be machined in a manner that facilitates cooling of all the hot spots. Granted, your design provides a cooling solution for the CPU but the CPU isn't the only part that gets hot, especially if you're overclocking/overvolting your Pi, which I presume a lot of your potential customers will be doing. This case should also be in contact with the other 2 hot spots, and thermal pads should be fitted to the contact areas.

LOL. Or you could buy a heatsink kit for $2 shipped from China. A heatsink independent from the case might even be easier to manage. That is, if you think a heatsink is needed at all (debatable).

[The_Doc]

Form over function!

Firstly, I do like the look of the case... but if you're going to design a case for the Pi, and you're going to machine it from aluminium, then it should be machined in a manner that facilitates cooling of all the hot spots. Granted, your design provides a cooling solution for the CPU but the CPU isn't the only part that gets hot, especially if you're overclocking/overvolting your Pi, which I presume a lot of your potential customers will be doing. This case should also be in contact with the other 2 hot spots, and thermal pads should be fitted to the contact areas.

LOL. Or you could buy a heatsink kit for $2 shipped from China. A heatsink independent from the case might even be easier to manage. That is, if you think a heatsink is needed at all (debatable).

LOL... Or you could expect something that's advertised as "A solid aluminum enclosure that sinks heat for overclocker's " to actually be capable of sinking heat for overclockers.

Whether a heat sink is actually needed is only debatable if you don't overclock/overvolt your RasPi enough to run XBMC smoothly. I overclock mine enough to run XBMC smoothly and it needs heatsinks.

[awp0]

Post your component temps. I'm curious how far out of spec your USB/Ethernet chip is (for example) given that it really shouldn't be impacted by overclocking and overvolting.

[The_Doc]

The ethernet chip won't be affected by overclocking but it is affected by running a lot of peripherals and streaming 1080p movies, and earlier versions get very hot due to a design flaw...not everyone has the later version.

Have a look here for temperatures of a stock RasPi with no case

http://raspberrypi.stackexchange.com/que...components

Those temperatures are already quite high and they rise dramatically when you enclose the Pi in a case, and even a relatively light overclock will raise the temperatures to their maximum design limits.

I tested my temperatures with the Pi in a case using a Fluke multimeter. With no overclock/overvolt, the CPU/RAM and the USB/Ethernet chips (at idle) were 64 degC and 59 degC respectively. At a 1000/500/500/6 overclock, streaming HD video, the CPU temp rose to 95 deg before it crashed and the ethernet chip was at 80 degrees, which is at the limit of its design spec and at that temperature it should have a heatsink.

I now have heatsinks on the CPU, the Ethernet and the voltage regulator, and I have a small fan on the case. The CPU now idles at 34 deg and maxes out at 52.





There's no point trying to convince me that (my) RasPi doesn't need heat sinks, because I've checked its temperatures and to keep the CPU within design temps, it needs a heatsink, and without a heatsink, the ethernet chip is at the very limit of its spec, and those temperatures were measured with a relatively low ambient temperature and will no doubt rise when the ambient rises.

[Milhouse]

Have a look here for temperatures of a stock RasPi with no case

http://raspberrypi.stackexchange.com/que...components

Just to point out those are thermal images of a v1 Raspberry Pi (note the lack of mounting holes introduced with the v2). The temperatures on a v2 Raspberry Pi could be quite different, particularly in the region of the LAN chip as the v2 fixes the design flaw that drove up the temperature of that chip.

At a 1000/500/500/6 overclock, streaming HD video, the CPU temp rose to 95 deg before it crashed and the ethernet chip was at 80 degrees, which is at the limit of its design spec and at that temperature it should have a heatsink.

That's odd, as the CPU will throttle once it hits 85C (v1 and v2).

[awp0]

Yeah, those temperatures seem unusual to me. If you read about this topic over on the raspberry pi forum you'll find a general consensus that heatsinks are not needed, especially for the V2, even when overclocking. This post seems to sum it up nicely: http://www.raspberrypi.org/phpBB3/viewto...43#p347643

It's also been demonstrated that the heatsink alone (without fan) only drops the raspi component temps by a few degrees after the system has reached steady state. That's with a set of small standard finned heatsinks. And it's hard to know what effect a solid aluminum enclosure like this will have when compared to a thin plastic enclosure. The aluminum is a much better conductor of heat and will likely draw heat away from both the air inside the case as well as the CPU. So the air temp inside the case might be lower in this case compared to a cheap plastic case. Your "80 degrees edge of design limit" temp on the USB/ethernet chip would likely improve. And I believe that is a non-issue for V2 owners.

Anyway, the_doc, I suppose this isn't worth discussing since you are admittedly unwavering on this topic. For everyone else considering this enclosure, at least they have the information at hand to make a decision. Getting back to your original objection, I do not think this enclosure is misrepresented because it does have a heatsink for the CPU and even if when considering your unusual temperature readings I think it's worth noting that they still haven't strictly demonstrated the requirement for cooling on anything other than the CPU (which is indeed handled by this case). I believe even your raspi would run without issue inside this enclosure.

[ghart999]

Looks completely bad-ass. Good luck in the kickstarter.

[The_Doc]

Yeah, those temperatures seem unusual to me. If you read about this topic over on the raspberry pi forum you'll find a general consensus that heatsinks are not needed, especially for the V2, even when overclocking. This post seems to sum it up nicely: http://www.raspberrypi.org/phpBB3/viewto...43#p347643

Have any of those people actually measured the temperatures difference between running a Pi in an enclosed case and running it in the open air?

Anyway, the_doc, I suppose this isn't worth discussing since you are admittedly unwavering on this topic.

It isn't worth discussing because you're comparing temperatures from people who are running their Raspberry Pi in open air to those running it in a sealed enclosure, and you're expecting those figures to be somewhat similar, when nothing could be further from the truth.

I've been building and overclocking computers for about 25 years, and I do know a little bit about cooling at this stage, and my personal experience is that a RasPi shouldn't be enclosed in a case if you plan on overclocking it. Heatsinks should be added and at least allowed passive cooling from ambient air, because electronic components work better and last longer when they're cooler.

e2a... I'm running at 1.2 gig now with SDRAM @ 680 a GPU freq of 355 with a 710 core freq and an overvolt of 8... and it's @ 46 deg streaming a 1080p 20Mbit/s video.

[awp0]

If or when someone is able to more thoroughly test an overclocked Pi inside this case, I'll be interested to see the results. As someone with a distant background in heat transfer, I won't be surprised if the large aluminum block enclosure performs well at not only drawing heat away from the CPU, but also drawing the heat out of the air and keeping the internal air temperature lower than that of a plastic case.

And no, the comments prevalent through the raspi forums include people with overclocking and inside cases.

[The_Doc]

As someone with a close background in heat transfer (my company designs and builds radiators and heat exchangers), I can categorically state that yes, aluminium would be a much better conductor of heat than a plastic Pi case. I can also tell you that aluminium conducts heat approximately 10,000 better than air does (you did of course know that air is an insulator, not a conductor?... of course you did, you have a 'distant background in heat transfer') So like I said in my first post... If I was designing an aluminium enclosure for a Pi, I'd machine it so all of the hot chips were in contact with the case. You obviously don't think this would be beneficial, so we'll simply have to agree to disagree on that one.

And on that note, I'll wish the guy every bit of luck with his kickstarter campaign and bid you a good night.