if the pinout is the same it is possible, though there's the chance that timing issues might appear due to the original pcb design.

weedeewee wrote on 2021-03-27, 12:38:

if the pinout is the same it is possible, though there's the chance that timing issues might appear due to the original pcb design.

Why would these timing issues occur? Are the memory chips that sensitive to the physical length of the traces, or is there some other reason?
Not trying to argue or anything, I'm just really curious. I'm by no means an expert on PCB design, so it's always interesting to hear what more experienced people have to say 😀

LunarG wrote on 2021-03-27, 13:14:

Why would these timing issues occur? Are the memory chips that sensitive to the physical length of the traces, or is there some other reason?
Not trying to argue or anything, I'm just really curious. I'm by no means an expert on PCB design, so it's always interesting to hear what more experienced people have to say 😀

yes, differences in trace length might mess up the signals.
there's probably other reasons as well, like pcb material impedance etc... but I ain't no expert nor have I experience in pcb design 😁

weedeewee wrote on 2021-03-27, 13:28:

LunarG wrote on 2021-03-27, 13:14:

Why would these timing issues occur? Are the memory chips that sensitive to the physical length of the traces, or is there some other reason?
Not trying to argue or anything, I'm just really curious. I'm by no means an expert on PCB design, so it's always interesting to hear what more experienced people have to say 😀

yes, differences in trace length might mess up the signals.
there's probably other reasons as well, like pcb material impedance etc... but I ain't no expert nor have I experience in pcb design 😁

It might be worth looking into at some point. I'm kind of looking to optimise the performance of my 486. I've got some 10ns cache chips that I'm going to install. I doubt it will really make any difference, but it couldn't hurt. I'd also enjoy experimenting with faster memory modules. The combination of fast cache and fast ram might be good for running 50MHz bus with tight timings.

LunarG wrote on 2021-03-27, 13:48:

weedeewee wrote on 2021-03-27, 13:28:

LunarG wrote on 2021-03-27, 13:14:

Why would these timing issues occur? Are the memory chips that sensitive to the physical length of the traces, or is there some other reason?
Not trying to argue or anything, I'm just really curious. I'm by no means an expert on PCB design, so it's always interesting to hear what more experienced people have to say 😀

yes, differences in trace length might mess up the signals.
there's probably other reasons as well, like pcb material impedance etc... but I ain't no expert nor have I experience in pcb design 😁

It might be worth looking into at some point. I'm kind of looking to optimise the performance of my 486. I've got some 10ns cache chips that I'm going to install. I doubt it will really make any difference, but it couldn't hurt. I'd also enjoy experimenting with faster memory modules. The combination of fast cache and fast ram might be good for running 50MHz bus with tight timings.

Actually there mighty be a slight chance the 10ns sram does hurt, as faster parts tend to draw more current. Even at the same speed.
Personally I haven't had any issues yet and most likely it's well in spec 😀