Slashdot Mirror
Cyrix Hotplate Howto
fimbulvetr writes "Looking to put those old processors to work? Tired of catching flack for having hardware, but no use for it? Worry no more! Doc from rabidhardware.com shows us how to employ 7 Cyrix processors to build a spectacular cooking device. Cooking instructions not included. Void where prohibited."
From the printable version: http://www.rabidhardware.net/index.php?id=44&print able=yes
Welcome to the 21st century. The age of conservation, renewable materials and Jolene Blalock. As we're urged to replace our gaming equipment on a weekly basis, many tonnes of silicon, lead, and copies of Daikatana make it into our planet's landfills.
At RabidHardware we strive to be environmentally sound (so says our lawyers). By re-using hardware we would have so hastily discarded in our youth, we can now give our dear Earth a new lease on life. Seeing how the Great White North is in the middle of one of the more colder winters as of late, and I'm on a budget (read: cheap bastard), I figure we could get two birds stoned at once with this latest project: A CPU-driven hotplate.
Enter the intrepid Cyrix(tm) Central Processing Unit. Instead of piling the landfills with these retired, non-biodegradable heathens (or donating them to NASA for shuttle heat shielding), we may as well put em to further use. So what do we do instead you ask? Well, there is only one thing a Cyrix CPU does well besides reflecting heat, and that is producing it.
To do this we'll be using 7 6x86 Cyrix CPUs ranging from 100mhz to 150mhz, dissipating an upwards of 20+ watts each. All chips will be supplied with 5v regardless of their original requirements, which I imagine will also improve the thermal output.
These be my materials:
- 7 Cyrix CPUs (1x PR120, 4x PR166s, 2x PR200s)
- Lexan sheet for CPU base
- aluminum/copper/cookie sheet for hotplate surface
- AT 250W power supply
- wire, solder, fixin's
- 7805 Voltage Regulator
After a quick look at this handy chart, I've decided to use pins A7 (core voltage) and B10 (ground) for our electrical connections.
Step 1 - Processors
First off, clean up the procs in question, as stuff like ancient heatsink compound or warranty stickers (that haven't already burned off of course) will impede heat transfer!
Most important thing we need to do is supply power to these little thermo-electric heaters of ours. All we need to do is run a 12v rail from the power supply into the voltage regulator (which will output +5v @ 1A) and connect the procs in parallel:
I realize there may be a better way to do this but we don't have time for rational thought. After all, my bacon expires tomorrow and I am VERY hungry! I also realize we could just use the 5v rail directly off the power supply. The reason for the VRE though is to regulate the amperage available to our hotplate while adding an extra stop-gap to keep our high quality power supply from exploding (prematurely?). SAFETY FIRST!
Feel free to remove the surrounding pins for easier soldering, sure as hell won't be needing them anymore!
Step 2 - Goop
Next off, we'll need to create our most excellent thermal interface. The Arctic Silver 3 which I've had in my toolkit for several years should do nicely, as it has a peak temperature of 180c. Feel free to don a piece of plastic or your favourite straight razor (preferably bloodless) to develop that sexy paper-thin layer of arctic goop, but I'm sort of in a rush.
You may have noticed we've also attached a heatsink to the regulator. Unfortunately, as Thermaltake or Alpha haven't made performance VRE heatsinks (yet), I had to go with a generic brand. Don't worry though, we'll make up for the performance loss later.
Once you've got your transfer medium installed, it's time to add the hotplate surface. I went with a generic piece of metal from a cookie sheet, but a aluminum or copper sheet would do better. I'm assuming you'll be lapping the side that the procs will be in contact with, right? Once it's ready, position your hotplate surface and press down to further spread around that silvery goodness.
Now is probably as good a time as any to mention: As with any of my projects, make sure you have a fire extinguisher and bomb squad nearby. Again, SAFETY FIRST!
At RabidHardware we're always in for any extra per
sorry about that formated the link badly http://www.rabidhardware.net.nyud.net:8090/index.p hp?id=44
The only things certain in war are Propaganda and Death. You can never be sure which is which though
Here is a mirror, as I have found the site to be /.'d already.
I'm not good in groups. It's difficult to work in a group when you're omnipotent. - Q
http://www.google.com/search?q=cache:N6NjjllZah8J: www.rabidhardware.net/index.php%3Fid%3D44+&hl=en
Google is your friend. http://64.233.183.104/search?hl=nl&q=cache%3Ahttp% 3A%2F%2Fwww.rabidhardware.net%2Findex.php%3Fid%3D4 4&btnG=Zoeken&lr=
My guess is that he didn't REALLY get 120 watts through, but enough current to warm up the chips nicely.
From a quick Google search on one of these...
USM 7805 is a 3-terminal positive voltage regulator designed with built in internal current limiting, thermal shutdown and safe-area compensation for maximum flexibility and safety . With adequate heat sinking provided, USM 7805 can deliver up to 1.5A output current.
I'm with you on your conclusion. The chip at current limiting gives you about 7.5 Watts, not 120 Watts or anything close. I have a night light for the kids that put out the same heat he could have gotten.
The truth shall set you free!
The main reason gas is preferred is because of heat control. With gas, you can quickly adjust the amount of heat going to your pan. Electric takes time to both heat up and cool down.
Well, let me backpedal a bit on my flat denial.
I can imagine designing a system that used a 20W heater to heat a very large mass of metal to a suitable temperature to fry an egg, with sufficient heat capacity to keep that temperature throughout the cooking of the egg
Now, assume a mass of copper (Specific Heat of 0.385 Joule/g/C)
I just fried an egg, using my little hotplate. I used a deep-fry thermometer to measure the temperature of the oil at 135C near the middle of the cooking process (just before I turned the egg). From raw egg to breakfast was 3 minutes 30 seconds (plus or minus 15 ... It's difficult to juggle an egg a hotplate a fry pan and a stopwatch without setting the kitchen on fire!)
So, I soaked 850 (power output of hot plate) watts into my breakfast for 210 seconds, or a total energy input of 178.5 kJ. So, how much copper do I need to heat to a 135, such that after sucking out 175,000 joules it will be about 120. Fifteen degree drop , 175000 joules, comes to about 30 kg of copper.
To heat 30 kg of copper to that 135 in the first place (from an ambient of, let's say 25) will take 110 * 30000 * 38.5 equals 1.27 MJ.
At twenty watts, a mere 17.6 hours, assuming your heater and the block of copper are in a perfectly insulated space. Putting it in the real world will make it take longer (in fact, probably an infinite amount of time because of radiation loss).
Learn the difference between heat and temperature
How does the Slashdot Effect happen given that no slashdotters ever RTFA?
I personally preffer to fry my eggs on an Athlon XP. Just make sure it's an original Athlon as new ones run too damn cool.
Also, this is 2005. The era for electronics projects is 20-30 years IN THE PAST.
Save ALL your money, and just BUY whatever it is you need. Chances are, whatever you need, it's already been done, and it's cheaper and better than anything you can throw together.
Also, a scope is not a substitute for a DMM. You're an idiot for even thinking that, and suggesting it to an amateur should get you a jail term.
If you want confirmation, just go read his article on lapping and you'll see that yes, it is tongue-firmly-in-cheek. The silly staged 'cooking' pictures also could have clued you in ;)
(I'm not an electrical engineer and I caught this one...)
Repetition does not transform a lie into the truth. - FDR
Get a decent soldering station with sponge and adjustable wattage (or if you really want to spend money adjustable temperature.) They are suprisingly affordable, also for a while you won't need a soldering ironas much as you will need a breadboard and lots of wire.
Just about any multimeter will do, even a $5 analog one, most of the time it is more a question of presence and magnitude rather then particular readings. (Though the nicer ones do have some cool features).
Look at http://www.sparkfun.com/ for your starting out. Lots of tutorials, a forum, a well stocked supportive store, and unlike digikey it has a limited enough selection that you won't feel like you are paging through a large city's phone book.
Do eventually sign up for digikey/mouser/future etc. catalogs though, just don't let them intimidate you.
The real hard part is finding a project.
I'd do something interesting, but my server can't handle a slashdotting.
I have a hard time believing this. What are the processors actually doing? If you simply power up processors without being connected to chipset or memory they won't do a whole lot. Most likely it would get through reset (assuming reset doesn't do any chipset queries, etc.), then put out a Code Fetch for the reset vector, which would never come back, then it would go into some sort of shutdown state. So I find it hard to believe that just powering up processors would produce that much heat. There are specific workloads which could be used to maximize power throughput, but obviously they are not at work here.
The 7805 itself may have an internal 1A limit but 1A is plenty for driving a current-boost transistor, at which point the limiting factor becomes pass transistor heatsinking.
1- connect ~5 ohms resistor between reg-In and power source
2- connect PNP base to the regulator-resistor node
3- connect PNP collector to regulator output
4- connect PNP emitter to power source
At very low loads, the transistor's base remains reverse-biased so the regulator provides all load current but beyond 200mA, the load's bulk migrates towards the pass transistor.
Without this patch though, the 7805 would either shutdown or burn out... from the looks of the only picture that loaded before the site died on me, they used a tiny heatsink and probably no pass transistor so...
TO220: Tja = maybe 20C/W with the tiny heatsink
Pd = (Vin-Vout) * Iout = (12-2)*1 = 10W (the CPUs are practically shorting the output so 2V seems reasonable.)
Tj = Pd * Tja + Ta = 25C + 10W * 20C/W = 225C
Since most silicon ICs do not like junction temperatures above 125C, I declare their 7805 most likely deep-fried.