On 2007-11-26 18:32, Roy J. Tellason wrote:
I've been
thinking about how one could make a reasonable memory, using small
reed relays. If one puts a magnet close to one end, it will cause the reed to
close. Move it back some and it will hold until the field is too weak.
If one set the magnet someplace in the middle of this band, one could use the
direction of current through the coil to set and reset the contacts.
One could create an array of these for the memory. Placing small magnets in
opposite directions for adjacent bits would keep from building too high a
field buildup, in the array.
I was thinking that one could use small round magnets glued to the end of a
threaded rod. This could then be adjusted to optimize the memory.
Now all I need is a pile of reed relays.
Dunno if they still sell them or not, but Radio Shack used to sell a pack of
20 reed switches for only a couple of bucks...
There's Chinese guys on eBay selling packs of fifty print relays for under
ten dollars plus shipping. They're only single throw, so a bunch would be
needed for some functions, but at those prices... An 8x8 RAM card would take
about 150 relays (eight latches per byte, eight enablers, one hold, one
select, and a couple extra for three bit address decoding). But really, how
much RAM would you need on a relay computer? Half the fun is in building a
somewhat limited machine to see what you can do, right? Plus, replace some
of the relays with switches, and you have ROM to hold a program.
Hmm... Time to fire up EagleCAD! :-)
,xtG
.tsooJ
--
who | grep -i blonde | date
cd $HOME; unzip; touch; strip; finger
mount; gasp; yes; uptime; umount
sleep
--
Joost van de Griek
<http://www.jvdg.net/>