What's this???

From: Pete Turnbull <pete_at_dunnington.u-net.com>
Date: Sun Dec 19 05:32:07 1999

On Dec 18, 21:39, Richard Erlacher wrote:
> I've got a mystery before me . . . there's a component, probably a 64Kx8
> SRAM but, who knows? I can't find a lead on the manufacturer or
whatever.
> The part's numbered GLT751208-15 and four of them comprised the data
cache
> on an old '486 motherboard climing to have 256KB of cache. That,
combined
> with the number of the part, 512 for the number of k-bits and 08 for the
> data width . . . looks right, but I haven't seen a data sheet for a
64kByte
> Sram in a 32-pin package. All the ones I've seen are 128kByte parts.

That's my experience, too. Looking through my pile of early-90's memory
data books, I couldn't find that number listed. However, I found a couple
of 32-pin SOP devices made by Samsung that are 64K x 8: the KM68512 is
32-pin 64K x 8. It has the same pinout as the more common 32K x 8 28-pin
cache SRAM devices, except that pin 30 is an active-high CS (that would be
pin 28, Vcc, on most 28-pin 32K x 8 devices), pin 31 is A15 (the extra
address line), pin 32 is Vcc, and pins 1, 2 are N.C. (pin would be A16 on a
128k x 8 device). Apart from the NC on pin 2, that would make it the same
pinout as the common 128k x 8 devices. Incidentally, although the Samsung
book lists lots of other manufaturer's equivalents for nearly all their
devices, there are none listed for the KM68512. It's obviously not a
common configuration!

FWIW, several motherboards did use 4 chips rather than 8 for certain cache
sizes (mine uses 4 off KM681001 for 512K x 8 cache).

There are a few other devices that might fit the description, except that
they are BiCMOS centre-power devices (Vss and Vcc are in the middle of the
sides, not at the corners). It's very unlikely that any motherboard used
such chips for cache, I think.

Check where the 5V connection goes, and if it's on pin 32 (and possibly pin
28), I'd assume it's the "normal" pinout, and similar to the KM68512 I
found.

-- 
Pete						Peter Turnbull
						Dept. of Computer Science
						University of York
Received on Sun Dec 19 1999 - 05:32:07 GMT

This archive was generated by hypermail 2.3.0 : Fri Oct 10 2014 - 23:31:55 BST