On 14 Jun 2012 at 0:58, Alexander Schreiber wrote:
Well, from what I'm currently reading in
"Dealers of Lightning", very
early semiconductor memory was quite a bit less reliable than core,
coupled with miserable yields. Presumably both improved reasonably
quickly. It had the advantage of being smaller, cheaper and faster,
though.
Wasn't the commodity DRAM chip back then a 22-pin 0.400" wide 3-
supply device--2107? I think that almost everyone made them.
4K bits IIRC. Wasn't there a TMS4096 that was much the same thing?
Not a nice beast, IIRC, slow and power-hungry, with an inconvenient
+12 chip enable.
I suspect it was regarded as a huch improvement over the 1103 IK*! DRAM.
That thing was PMOS and has 16V logic levels on all pins (address, data,
seelct, etc). Intel did make some level shifter ICs to go with it (3207
and 3208 I think), but from what I revall, they came out significantly
later than the RAM itself.
Oh, and the 1103 needed +16V and +19V power rails, not particularly
useful for anything else in the machine.
Mildly off-subject, but there was the 1103 and the 1103A. IIRC, the
former needed a 'precharge' signal before every operation (after the
address was stable), the latter didn't, but the Precharge pin on the 1103
became a no-conenct on the 1103A so you could substitute the foremr with
the latter with no problems.
-tony