[cctalk] Re: Core memory (was Re: Computer Museum uses GreaseWeazle to help exonerate Maryland Man)

[Paul Koning via cctalk]

> On Feb 2, 2023, at 11:03 AM, Sellam Abraham via cctalk 
> <cctalk@classiccmp.org> wrote:
> 
> This discussion on core memory got me thinking:
> 
> Is it possible to "read" core memory by examining each core using some kind
> of instrument that would sense its "charge" (or lack thereof)
> non-destructively?
> 
> Could a piece of paper be placed over a core plane and fine particles of
> iron sprinkled onto the paper to literally "see" the bits?
> 
> Please excuse me for any ignorant assumptions made here on my part.
> 
> Sellam

That's a very good question.  The difficulty with doing this isn't that the 
core states are "magnetized" and "not magnetized" but rather they are 
"magnetized in one direction" and "magnetized in the other direction".  So 
flux-visualizing won't help because 1 and 0 will look the same.  

The other problem is that a toroidal core, by design, constrains nearly all the 
magnetic flux inside the core material, with very little "leakage".

That said, it would be an interesting experiment to take a core memory module 
and hold a small Hall effect probe up against the cores, especially inside the 
core opening.  I don't know if those sensors can be sensitive enough to pick up 
the leakage flux.  If yes, it may well depend on the core diameter; the later 
generations tended to have smaller cores for both density and speed.  CDC 
mainframe cores are probably small for the time, I don't remember seeing the 
numbers.

        paul