29 Jun
2007
29 Jun
'07
8:44 a.m.
On 29 Jun 2007 at 1:08, Brent Hilpert wrote:
Core logic is quite a different form of gate
implementation than what we
normally deal with, the static state of the network is represented by the
magnetic field of the cores (not by DC voltage levels at gate outputs) and it
requires AC activity into the cores to propagate changes through the network.
Bitsavers has the operations manual and complete schematics for the
SS-80 with a really good explanation of how these things work. If
you're used to semiconductor logic, core logic is *very* different.
If you consider other machines of the time and that the entire logic
power came from the clock signal, 6 4X150s (at what, 250W plate
dissipation per the each?) wasn't bad at all.
It'd be fun to see if the large cores for this type of logic could be
obtained to experiment with this logic type again.
And let's not forget the Packard-Bell 250, whose memory consisted of
magnetostrictive delay lines and a bit-serial ALU. I think it had a
word length of 22 bits.
Cheers,
Chuck