|Re: Memory Upgrades:42s &48G|
Message #2 Posted by Paul Brogger on 17 June 2002, 11:26 a.m.,
in response to message #1 by Ron Ross
I did two 256K upgrades to my 48G's. I piggy-backed one 128K chip on top of the other. (Most of the corresponding pins need to be connected above & below, with just a few unique chip-select signals.)
It worked O.K., until the piggy-backed chips started coming loose. As it turned out, simply bending the pins of the upper surface-mount chip down and soldering them to those below didn't last -- there was a slight gap, and a simple solder joint didn't seem to hold up very long. (And soldering individual tiny wires to connect each pin pair is for me nearly impossible to accomplish.)
I ended up backing 'em each down to 128K.
If I were to try again, I'd try some very fine bare copper wire, and "lash" the chips together, with figure-eight connections pin-for-pin, using one continuous strand (only for those pins that need to be connected, of course). Then, soldering the wire to the pins should be easy, and I would finish by cutting the crossover wires running over the top of the upper chip. (I haven't done this yet, but that's the idea.)
If I remember right, chip-select for 128K or 256K is fairly straightforward, but beyond that, there is extra logic required to generate the signals needed. Pay attention to chip-select logic, and look up all the references you can find. For my purposes, I decided 256K was as complicated as I wanted to go.