On Hackaday:
http://hackaday.com/2013/08/18/l...
Says it can be overclocked to 65.3 MHz.
Arduino on liquid nitrogen!
Author
Message
So then, I'm wondering if even an older IBM CPU heat sink bonded to an AVR in say, a TQFP package with a fan, if that could possibly allow a push of the AVR CLK speed up a bit? Or even using one of those water cooled heat sinks?
Does this mean that Atmel support the use of their products in this manner?
Level: 10k+ Postman
Joined: Tue. Nov 11, 2003
Posts: 10289 View posts
Location: Bartlesville Oklahoma USA
I remember in the early 80's, experts saying that we would soon have cryogenic computers on our desks. Seemed dangerous, expensive and a little impractical to me, and I was a big fan and early user of microcomputers.
Pretty cool experiment. In more than 1 way. Who doesn't like dunking things in liquid nitrogen?
If you don't know my whole story, keep your mouth shut.
If you know my whole story, you're an accomplice. Keep your mouth shut. 
Kartman wrote:
Does this mean that Atmel support the use of their products in this manner?
Why? Because I posted a link to someone else who has done this?
EW wrote:
Kartman wrote:Does this mean that Atmel support the use of their products in this manner?Why? Because I posted a link to someone else who has done this?
Not that I'm saying that it was, or even should have been but... maybe because of your affiliation with Atmel, an assumption of endorsement such things can be, itself, assumed.
It's an incorrect mindset, I realize. But people do have a tendency to read into things that simply aren't fact.
It's plain ridiculous.
Should I pull off my signature line and go stealth mode like other Atmel people here? And then have AVR Freaks users complain that nobody from Atmel comes here and listens?
Or should I put in an annoying disclaimer in my signature line?
As it stands, it should be pretty clear that any disclaimer is in the device datasheet, for recommended usage of said device.
But, as Torby says: Who doesn't like dunking things in liquid nitrogen? ;)
EW wrote:
I agree!
It's plain ridiculous.
EW wrote:
No. I don't think that will solve anything.
Should I pull off my signature line and go stealth mode like other Atmel people here? And then have AVR Freaks users complain that nobody from Atmel comes here and listens?
EW wrote:
I don't think that would help either as, people will simply look past the disclaimer and continue making assumptions, just as is done with various other posts made here on AVRFreaks.net, and with the datasheets.
Or should I put in an annoying disclaimer in my signature line?As it stands, it should be pretty clear that any disclaimer is in the device datasheet, for recommended usage of said device.
EW wrote:
I thought it was cool, too! I even started wondering if there was a way I could design a very small controller board, design an Aluminum case, machining the thing on my table-top mill and placing the controller in the case, totally flooding the inside space with Nitrogen and bring the necessary wiring out for access. But then... logic & reason got the better of me. So I shifted my thinking over to the PC style liquid cooled heat sink scheme I mentioned earlier.
But, as Torby says: Who doesn't like dunking things in liquid nitrogen? ;)
Eric, I for one thought it was fascinating (thank you), but have no intention of using it in my designs :lol:. No need for a revised tagline.
Eric, of course it is ridiculous, but for the reason that Carl points out one can't be too careful. If I put my engineer's hat on my thoughts would be " i'd always wondered what would happen - now someone's actually done it - cool!"
Vs
Politically correct/lawyer's hat - " hey Cletus! Someone from Atmel is sayin we can freeze their stuff and it works four times as fast"
Seems my subtle attempt at humour went astray. Tough crowd!
I was going to take a picture of my Arduino hanging over-a-clock, but I thought better. Eric is not happy right now. :(
I thought it was pretty 'cool'(no pun) to check out. Impressive that the thing actually worked at that level of cold at all is impressive.
So, what's the conductivity of liquid nitrogen?
Obviously not very conductive, as it didn't trash the ext osc on the PCB, or the rest of the circuit.
Wait until AtomicZombie sees this Thread. He was over clocking the Xmegas to 60+ MHz at room temperature.
I think I've recommended to Eric previously that adding the Atmel by-line to his signature is a good thing.
Agreed, interesting project, and follow-on comments (in the link) about the effect of the liquid nitrogen when poured in the mouth or on one's fingers. I don't intend to test those comments.
JC
jgmdesign wrote:
I was going to take a picture of my Arduino hanging over-a-clock, but I thought better. Eric is not happy right now. :(
S'alright! I'm good! :wink:
jgmdesign wrote:
I thought it was pretty 'cool'(no pun) to check out. Impressive that the thing actually worked at that level of cold at all is impressive.
Agreed. I was pretty surprised myself.
But components/epoxy do not have the same thermic dilatation coefficient? What becomes of the arduino, if it is often dipped into liquid N2? Do solders remain intact?
I dare say most things wouldn't appreciate such thermal cycling.
I'd rather a dewar flask than a flask of dewar's!
dbrion0606 wrote:
But components/epoxy do not have the same thermic dilatation coefficient? What becomes of the arduino, if it is often dipped into liquid N2? Do solders remain intact?
I was thinking something similar. That's why I was surprised it even worked. I would have thought that the solder would've given way.
But, hey! I'm a software engineer by training. :wink:
I'd like to protest (mildly) re this formulation:
Quote:
It turns out that you can run an Arduino at 65.3MHz when it is cooled with liquid nitrogen!
Not "an Arduino", but "this specific Arduino".
Apart from that it was a fun read!
Notes on LN2 use
1) Perhaps the processor worked, but other components may have problems with cryo-temperatures, especially tantalum and electrolytic capacitors. As JohanEkdahl said, "this specific Arduino" worked. I'm surprised that the processor worked, since semiconductor's properties vary considerably by temperature.
2) The main issue is with thermal cycling. Adhesives especially will have issues with repeated cycling between room temperature and cryo, any surface-mount component that doesn't have J-leads (to take up the stresses) may have problems.
Couldn't it run even faster if he upped the core voltage little(like the pc overclocking people do)?
Since the µc is cooled , it shouldn't be a problem to up the voltage, I guess...
Level: Raving Lunatic
Joined: Fri. Jan 23, 2004
Posts: 9856 View posts
Location: Melbourne, Australia
AVRs are not limited by thermal constraints; the flash and other memories will give out instead and mis-read.
- Dean :twisted:
Level: Moderator
Joined: Wed. Mar 28, 2001
Posts: 29341 View posts
Location: Sydney, Australia (Gum trees, Koalas and Kangaroos, No Edelweis, but plenty of Corona beer)
Maybe the user was running under the influence of "liquid amber"??