Make:

• Submersed Songs: Fish that make music
• Medical aesthetic footwear
• Whiteboard Pong!
• Mod your Famicom for better video output quality
• "Math powered" light
• Record your heatbeat for the neighborhood to see

That's inspired. :)

My Computer Science teacher had a stash of old parts he used in his lessons (this would have been mid 80s) including a stack of 8inch square cores. Probably got binned when he retired, would love to have them now...

Is anyone aware that IBM is going to do it again but now nanotechnology powered™?

This seems simple enough -- are there any small DIY magnetic core memory projects? Say, a few bytes? I found this image that gives a pretty good idea how it's made:
http://www.physik.uni-augsburg.de/exp6/imagegallery/samples/corememory.jpg

I googled around for DIY magnetic core memory, but didn't find anything.

The basic theory is still sound, remember the early shuttles used core memory (it's said the core recovered from the sea after the challenger disaster was still readable..)

@Ian: You may want to read the wikipedia article on Core Memory. That, combined with that picture, should be enough to get you headed in the right direction on a DIY solution. There are a few tricks to this tech that take a few read-throughs before it makes sense. See the "Description" section:

http://en.wikipedia.org/wiki/Core_memory

As to where you might find those little ferro-magnetic rings is beyond me. Good luck!

@Ian - In the wikipedia article that Pragma linked you to, there is a link to a July 1976 Byte magazine article that is the closest I've seen to a DIY project, but I don't know how hard that would be to build.

As to where to get them, they turn up on ebay. You can get ones that are already wired, or sometimes whole cans of unused ones turn up. But the cans have like 50K of the little buggers in them. :)

After some research, it seems the hardest part of building a single bit core memory is the analog detection circuit (called sense/inhibit).

Here is a description with two circuits, the one with a comparator looks easier for me (figure 6c):
http://www.cs.ubc.ca/~hilpert/e/coremem/index.html#sensing
The whole page is essentially a core memory how-to, IF you're an analog person. I'm a uC guy with no clue about the analog world.

Line drivers are no problem, I built a single pair with nixie parts, and added adjustable current by controlling the high-side switch (NPN-PNP pair) with a uC PWM. The only problem I see is the current required (400ma, it seems), but voltage is not important, so whatever we can get our hands on will work.

I have a sneaky source of ferrite rings that I don't want to share until I can test it.

@everybody - thanks for the comments and links!

Posted by: nodada on July 4, 2008 at 3:54 AM

@ian: I'm also trying to build an core memory. i'm not sure about the right ferrit core. as far as i know the permeability of the core should be very high. > 10000.
i'm currently using a h-bridge as driver. tested it with some cores with a permeability of 1000. i' messure some effects on the sense line with an oscilloscope. please let me know about you sneaky source...
best,
ralf

Subscribe today, save 42% and get web access to MAKE free. MAKE Digital Edition is available only to subscribers.

$34.95 / 1 year
(4 Quarterly Issues)

Subscribe now