How scarce (valuable) is core for the PDP-8?

From: Ethan Dicks <ethan_dicks_at_yahoo.com>
Date: Mon Apr 19 17:53:38 1999

--- Sellam Ismail <dastar_at_ncal.verio.com> wrote:
> On Mon, 19 Apr 1999, Ethan Dicks wrote:
>
> > I wish I had one of the stereo inspection microscopes...
>
> I have a similar device that was used for inspecting ICs. Its basically a
> microscope with a fine adjustment platform...

That's it. The one I'm thinking about has optional air bearings that
can be clamped down with a foot pedal. You clamp the board into the
frame, drag it around like a microfische, then lock it into place with
the pedal. The field of view is adjustable, and can zoom in to entirely
encompass a 1/4 W resistor or out to encompass several SMT ICs.

> I assume you are thinking to replace the wire strand completely, and not
> attempt to solder the two broken ends together?

I need to get some pictures of this...


Imagine a core plane...

   |||||||||||||||
 --///////////////--
 --///////////////--
 --//XXX//////////--
 --///XXX/////////--
 --///XXX/////////--
 --////XX/////////--
 --///////////////--
 --///////////////--

With the X's representing physically broken and/or absent cores. On my
particular board, each bit of 4096 cores is a square approx 2" on a side
(about 1/32" per core site), with the damage to two bits on the same edge
of the PCB.

I was originally planning on lifting the X and Y wires from one corner of
the bit to be repaired, unthreading the cores only where necessary, and
making any splices to the sense/inhibit wires at the edge of the core
(as I belive there already are). If I scavenge wire and cores from the
parity plane, I have more than enough raw material. If I attempt to
sacrifice one bad bit for the other, I don't have the surplus wire (the
X and Y wires must be preserved from edge to edge of the core PCB)

I have just thought of another, more devious method, but on second thought,
it would have to be clever indeed... Change the diode board such that
the damaged bits are not both accessed at the same time (i.e. X32, Y16 on
plane D4 is broken, but X32, Y16 on plane D3 is _not_ broken) and convert
the two broken bits into two fields of 2048 cores each (and use the parity
bit intact). It turns a 13-bit memory into a memory of 11 intact planes
and two half-planes. I'll have to study the geometry to see if this is
possible. Alternately, if the individual columns are intact, I could do
sort of a "one from column A, two from column B" approach. As long as the
sense amps don't balk at the extra resistance of doubling the length of
of the sense wire, it could work.

-ethan
 
_________________________________________________________
Do You Yahoo!?
Get your free _at_yahoo.com address at http://mail.yahoo.com
Received on Mon Apr 19 1999 - 17:53:38 BST

This archive was generated by hypermail 2.3.0 : Fri Oct 10 2014 - 23:31:44 BST