The 5101L is a low power CMOS memory which is used in some MPU units as non volatile (battery powered) memory for storing high scores, etc. While there doesn't seem to be any problem sourcing them in the States, I had trouble trying to get hold of one in the UK. What follows is a description of how to make a simple adapter which will allow you to replace a dead 5101 with a readily available alternative, without modifying your MPU board, and save you money at the same time (in the UK, at least).
Just to clear up any confusion: If your machine is working OK, then:
*there's no reason to do this*
(It won't make it better, stronger, or faster)
The 5101 is a low power CMOS stores 1024 bits as 256, 4bit 'nibbles' About the smallest capacity, widely available chip which could replace it is the 6116 which stores 16384 bits stored as 2048, 8 bit bytes. It's low power CMOS, with 100ns access time, standby current 10na, and data retention guaranteed down to 2v. Because the 6116 has more data and address lines than the 5101, we can play a few games to make life simple when making an adapter.
One thing about the 5101 - It has separate data in and data out pins. The 6116 hasn't. This conversion only works where these inputs and outputs are connected together - Check your board, pin 9 should be linked to pin 10, 11 to 12, 13 to 14 and 15 to 16.
The idea is to make up a 'piggy-back' board to adapt a 24 pin, 0.6" pitch 6116 chip to a 22 pin, 0.4" pitch 5101 socket:
VIEW WITH STRIPS ON UNDERSIDE
______________________________________
| | strips run this way
| . . . . . . . . . . . . | <===================
| 1 1 |
| . . O . [o] x O . . . [o] + | . hole in strip-board
| I I I I | |
| . . O . [o] x O . . . [o] | | x break in strip on
| I I I I | | underside
| . . O . [o] x O . + x [o] + |
| I I \ I I | [o] IC socket pin
| . . O . [o] x O + x + [o] . | soldered to board
| I I \ I I |
| . . O . [o] x O . \. . [o] . | [ ] IC socket pin cut off
| I I \ I I |
| . . O . [o] x O . .\ . [o] . | O PCB pin protruding on
| I I \ I I | *underside* of board
| . . O . [o] x O + x + [o] . |
| I I | I I | + - + Wire link (between
| + . O . [o] x O + . . [o] . | holes marked '+')
| | I I I I |
| | + O . [o] x O + . . [ ] . |
| | | I I | I I |
| | + O . [ ] x O + . . [o] . | Note: no break between pins
| | I I I I | 11 & 12 of the 5101
| | . O . [ ] . O . . . [o] . |
| | I I I I |
| + . . . [o] x . . . . [o] . |
| |
| . . . . . . . . . . . . |
| |
|______________________________________|
END VIEW (to make things clearer)
__________
/-(XXXXXXXXXX)-\
<------- 6116, happy at home
| |
[X]==============[X] ><------ modified 24 pin socket
======================================== ><-- strip board
I I
I I ><--- Pins to fit into 5101 socket
I I
>
The most awkward part is arranging some PCB pins, lengths of thick-ish tinned wire, or whatever to stick out of the bottom of the board, to plug into the 5101 socket. If you can get hold of 0.7 mm dia PCB pins, then push these in from the component side of the board, and solder in place. Alternatively, get the optional 22 pin socket, push 1/2" lengths of tinned copper wire into all its holes, and feed these through the stripboard. Trim off flush with the top of the board after soldering. The socket with the wires in it can then be plugged into the existing 5101 socket, without having to align 22 bent bits of wire. (I used wire wrap pins, and had to force them into a cheapy 22 pin socket)
The official order of construction would be:
1 [o]===___===[o] 24
[o] [o]
[o] [o]
[o] [o]
[o] [o] view is from *above* (pins down), as is normal
[o] [o]
[o]=========[o]
[o] [o]
[o] -[ ]<-
->[ ]<- [o]
->[ ]
<- [o]
[o]=========[o]
>
Unofficial order of construction:
Modify socket, solder it in, saw out 'spreaders' so you can get at the board to install pins and links, use screwdriver to make track breaks because you've got a blob of solder next to where you want a break so the rotating cutter won't work, plug in, unplug, check board and find short, pray you didn't break anything, plug in again, and go.
(A friend told me this ;-)
The CS1 (active low) OE and WE signals are fully implemented. The CS2 (active high) line is patched to one of the address lines of the 6116, so it is theoretically possible that the chip will be active when it shouldn't be, this doesn't matter, as any data written during this time will be diverted to an unused area of memory. Data cannot read unless the OE line is low. (In the case of my machine, CS2 was tied to the reset line, so was always high, anyway).
It might be an idea to fit a 100nf capacitor between pins 12 and 24 of the 6116 socket (good practice), also I get a vague feeling that the unused data lines should be connected (via 1meg resistor, say) to either ground or +5v, although this would make the wiring more complicated.
It worked for me, if anyone manufactures them for profit, I want my fee.....
Andy
Copyright Andy Gray, 1997 / Revised 9th October 1999 /
andy@misterg.fsnet.co.uk