Early Macintosh home brew 4MB memory upgrade board development

[pezter22]

I did not expect success with my first try.

First image is with stock Mac 512k 64mb ROM - No chime and steady screen distortion.

Second image is with ST M27C512 ROM's with Mac 512k flash. - No chime. Distorted Growl, then screen would flash, then go blank. This would repeat.

Any suggestions would be appreciated. The ROM-inator PCB's arrive this afternoon. I will put together one and check again later this afternoon.

 

[Builder68]

All wired up. EDO RAM in place. Switches are set.

I'm going to attempt to boot later today. ROM I think will be my issue with booting.

I currently have stock ROM installed. I have some ST M27C512 ROM's programed for the 512k. I also have a BMOW ROM-inator that I am setting up later today. I also downloaded the updated ROM file placed in this thread for the ROM-inator.

I also have some WinBond W27C512-45Z ROM IC's coming in a few days.

I don't necessarily need to use a ROM-inator installed because I have a HD20 drive to boot from.

I did not expect success with my first try.

First image is with stock Mac 512k 64mb ROM - No chime and steady screen distortion.

Second image is with ST M27C512 ROM's with Mac 512k flash. - No chime. Distorted Growl, then screen would flash, then go blank. This would repeat.

Any suggestions would be appreciated. The ROM-inator PCB's arrive this afternoon. I will put together one and check again later this afternoon.

Yeah, I've never had success on the first try either!

A couple of things:

Could you confirm if you first tried to boot the Mac with the Mac Plus ROM set and no RAM card installed? (Reinstalling RP2 & RP3). I would definitely do this before attaching the RAM Card, just to be absolutely sure the new ROMs work as expected.

Once you can boot the Mac with the new ROMs, attach the RAM Card and use the DIP switches to disable it. Try it this way and let me know what happens.

 

[pezter22]

Could you confirm if you first tried to boot the Mac with the Mac Plus ROM set and no RAM card installed? (Reinstalling RP2 & RP3). I would definitely do this before attaching the RAM Card, just to be absolutely sure the new ROMs work as expected.

I did try the ST M27C512 with the Plus flashed into them, but had a similar result.

I'm going to have to wait until the WinBond W27C512 ROM chips arrive. Turns out the ST M27C512 ROM chips have data on them and I don't have a strong enough UV light to erase them. I just read the chips and there's a mess of data on them.

I do have the SST 39SF040 ROM chips recommended for the ROM-inator, but my PCB's are finally arriving tomorrow (delayed) so I can make a ROM-inator.

I have to say the board fits perfectly inside the case. The motherboard slid in without any issues.

 

[Builder68]

Today, after some delay, I received the second PCB run that I published a month ago. It was supposed to be error-free, but I found a couple of mistakes in the traces. These were errors that I thought I had fixed while testing the first PCB run.

The first PCB run is working, but it had a few errors on the outputs of the 74139 in the schematic. I fixed these with bodge wires. However, it seems the design rule checker in my PCB editor didn't catch these changes made in the schematics, and the second PCB run still has the misplaced traces despite having associated the corrected schematics. I suspect something is misconfigured in my KiCad settings, as this is the second time this has happened to me.







On the bright side, the fix is easy and doesn't require cutting any traces:

• Desolder R5 and R7.
• Solder one side of R5 to TP2.
• Solder one side of R7 to TP4.
• Solder a bodge wire from the other side of R7 to pin 10 of U3.
• Solder a bodge wire from the other side of R5 to pin 6 of U3.

That's it!

I can now confirm that the Mac boots with the second PCB run and reports the expected memory size.

 

[zigzagjoe]

Today, after some delay, I received the second PCB run that I published a month ago. It was supposed to be error-free, but I found a couple of mistakes in the traces. These were errors that I thought I had fixed while testing the first PCB run.

The first PCB run is working, but it had a few errors on the outputs of the 74139 in the schematic. I fixed these with bodge wires. However, it seems the design rule checker in my PCB editor didn't catch these changes made in the schematics, and the second PCB run still has the misplaced traces despite having associated the corrected schematics. I suspect something is misconfigured in my KiCad settings, as this is the second time this has happened to me.

View attachment 86453

View attachment 86454



On the bright side, the fix is easy and doesn't require cutting any traces:

• Desolder R5 and R7.
• Solder one side of R5 to TP2.
• Solder one side of R7 to TP4.
• Solder a bodge wire from the other side of R7 to pin 10 of U3.
• Solder a bodge wire from the other side of R5 to pin 6 of U3.

That's it!

I can now confirm that the Mac boots with the second PCB run and reports the expected memory size.



View attachment 86456

Are you always opening the pcb from the kicad project view, and never by double clicking it in explorer/finder? The latter opens the pcb without the project and won't check that the schematic matches the board. Also loses some constraints and other settings that are stored in the project.

 

[Builder68]

Are you always opening the pcb from the kicad project view, and never by double clicking it in explorer/finder? The latter opens the pcb without the project and won't check that the schematic matches the board. Also loses some constraints and other settings that are stored in the project.

That's really good to know; thanks for pointing that out! I definitely didn't realize opening and saving the PCB file from the explorer would cause issues later.

I thought that just because the "Update PCB from Schematic" function appeared active (meaning the PCB file was opened from the KiCad project view), it would mean the association with the schematics was fine.

It's interesting that you mention losing constraints and other project settings that way. The only workaround I had found for the "Update PCB from Schematic" function not updating the new pad/trace assignments was to delete the footprint of the 74139.

I'll definitely make sure to always edit the PCB by opening it from the KiCad project view from now on. Thanks again for the valuable tip!

 

[Joopmac]

I have another question
Would this fit a 128k logic board 630-0101?
Second question: I have a 128k board with a defective RAM IC, would this board bypass the built in RAM?

 

[Builder68]

I have another question
Would this fit a 128k logic board 630-0101?

The PCB revision I have is 820-0141-A / 630-0118 (the green sticker was removed). This PCB revision was designed to be either a 128K (630-0101) board or a 512K board (630-0118). So basically, my machine was originally a 'Fat Mac' and was later upgraded to a Mac 512Ke (Mac Plus ROM & 800KB floppy drive).

The original Macintosh 128K came with PCB revisions 820-0086-C/F (630-0101).

The early PCB revision 820-0086-C lacks a resistor network on the 6522 VIA. This could lead to system lockups, and the installation of this resistor network was often part of early upgrade procedures. The later sub-revision 820-0086-F incorporated this fix.

PCB revisions 820-0086-C/F differ from 820-0141-A in that the latter has an additional 74253 IC to generate the RA8 signal when 256Kb ICs are installed (512KB of RAM). Jumper W1 on the 820-0141-A board should be bridged if the logic board revision is 630-0101 (Mac 128K) and open if it's a 630-0118 (Mac 512K).

Although I haven't tested physically, this RAM expansion board should fit on either PCB revision since both are the exact same size, and the RP2, RP3, U4F, U12E, and U13E footprints match the same positions (coordinates) on both PCBs.

What PCB revisions do you have?

 

[Builder68]

Second question: I have a 128k board with a defective RAM IC, would this board bypass the built in RAM?

This latest expansion board provides the full 4MB of RAM and disables the onboard memory via its DIP switches.

Regarding the defective RAM IC, it really depends on the extent of the defect. If it's causing a short circuit or can't maintain a high impedance state when pin #15 is pulled high, then you should definitely remove it.