For my Apple-1 build, I used the 2519 replacement board designed by Szillat and P-LAB. This worked great, but I had the issue that the small-package 14557 shift registers are not available from major electronics components stores (or only in very large quantities). I had to buy those on eBay, and they are not reliably available there either. However, the SOIC-16 Wide chips are easily available. So I recently I redesigned the board to handle the larger chips. I managed to do this without making the board larger, partly by putting all the decoupling capacitors on the opposite side.
I have now put the files in this GitHub repository. As I mentioned elsewhere, I am just starting to learn tools like EasyEDA, so this can probably be improved, but the board works fine for me so far.
Congrats to this work, it's a valuable contribution to the Apple-1 builders scene - as not everybody is willing to pay the usurious prices now being asked for real 2519.
How about "hiding" the 2519 replacement card below the motherboard ? It looks "flat" enough to do that.
The 2519 socket would stay empty but could take some "dummy", and empty DIL-16 IC package. Or some DIL-16 resistor network which is high ohmic enough to not disturb the signals in any way. If you avoid the yellow ones, and find grey or black ones, it would look - almost - right.
Here is one experiment which might be worthwhile to do: are the 14557 ICs on the substitute strong enough to overwhelm the weak output drivers of a real 2519 ? If so, this would give us the option to plug in a real but defunct 2519 into the otherwise empty socket. Just for the correct looks !
And if this does not work, there is always the possibility to manipulate the socket itself to disconnect the 2519 outputs from the PCB.
This would allow the use of original but bad Signetics 2519 to get the right looks. Nobody would see the substitute PCB hidden below the motherboard.
Comments invited ! (Assume I have a bag of bad 2519 which did not make it through my burn-in procedure).
- Uncle Bernie