Reverse Engineering the Macintosh SE PCB & Custom Chips for 1:1 reproduction

[Kai Robinson]

@Kai Robinson that's right, I've ordered the 4 layer PCB from China and parts.  However, I realized this attaches to the PLCC on the logic board, right?  If so, I'm not sure I quite understand the entire design.  I get that the 84pin PLCC is on top with the BBU attached, but what about the one for the logic board, is there an adapter?  Can you send a photo of a completed unit?  Or maybe a part number for that.  I'll de-solder the PLCC from an SE board and attach the break-out and send it to @asicsolutions.  I don't really want to burden him too much on this project as I know he is busy and also is focusing much effort on the Xceed project.  While I think this might help a bit with this project, if it looks like it will take too much time, I'll see who I can dig up in China to help. 

So the board you have - you desolder the BBU socket - solder either of the socket locations on the PCB into the board, plug the BBU back into a socket located on that PCB, and then the pin headers allow you to snoop the signals. 

 

[maceffects]

@Kai Robinson are the pins going to be long enough or will I need a spacer?  Sorry for asking so many questions, I just want to have everything in place so I can solder it once it comes and send it off.

 

[Kai Robinson]

@maceffects https://www.ebay.co.uk/itm/40-Pin-2-54mm-0-1-19mm-Long-single-row-Male-Breakable-Pin-Header-NEW-UK-Seller/202448089589?hash=item2f22d8b1f5:gmsAAOSwEqhbqS9P or

https://www.ebay.co.uk/itm/2-54mm-Single-Row-Male-Connector-PCB-Board-Pin-Header-Long/223980988737?hash=item34264ea941:g:XwwAAOSwROtemjgA

is what you'll want to use as a spacer - to clear the other components. Just use rows of strips. 

I've found a supplier on UTSource claiming to have BBU's (searching by 343S0603) and i've ordered 8 of them - i'll sacrifice 1 or 2 of them for the cause, if they actually DO turn up!

 

[maceffects]

@Kai Robinson OK I got it use 2.54 pitch headers I bought mainly for the headers on the side as spacers to clear components.  Perfect!  I just wanted to make sure since I'd hate to spent a long time soldering dozens of 84-pin connectors, haha.  Let us know of those parts turn out to be genuine. 

 

[Kai Robinson]

Mad idea - this weekend, who fancies a zoom call to brainstorm/collab?

 

[maceffects]

Mad idea - this weekend, who fancies a zoom call to brainstorm/collab?

I'd be in if we can get some others involved.  I believe @asicsolutions will be busy this weekend.  I am free except 5-9pm (CST) on Saturday. 

 

[quorten]

@Kai Robinson Honestly I think much of the new interest here needs to spend some time to review some of the existing documentation we have before we can get much progress on a zoom call.  Really, there's not much hidden info remaining after my analysis thus far, I just feel that no one has yet taken the time to look at and play around with my work-in-progress.

But yeah, after some prep-work to get us together into a more cohesive team, I'd be in for a call.

 

[maceffects]

@quorten I'm a late comer to the party so I'm not sure what existing information you have.  Can you repost so we can all have a fresh look?  I will admit right now that I'm an accountant not an electrical engineer so I'm best suited for grunt work rather than careful electrical analysis, but I'll do whatever I can.  That includes paying someone who might be able to do more complex analysis than we can do.

 

[asicsolutions]

I can hop on a call. I'm not sure what I can add so far. I'm happy enough to throw in my experiences with capturing PCI traffic on a board I developed using my trusty albeit aging logic analyzer and what I could try to do with it to try to capture stuff in the ASIC. I'm happy to help understanding the die shots if you get any, although, all the ASIC work I've done, I've had access to the GDSII files and sources to fabricate, so I'm not sure how much help I'll be in reverse engineering it that way. 

 

[quorten]

@maceffects Sure, my GitHub repository for my leg of the project seeks to be a useful reference repository that links to all the pertinent information.  A lot of the essential information should be copied in and covered there.  I'll be updating it with a few more links shortly but it is pretty comprehensive so far.

https://github.com/quorten/macsehw

@asicsolutions It would probably be a good idea for you to head on over to hardware/fpga/bbu/test_mac128pal.v in my GitHub repository and play around simulating what I have there.  That will give you a good glimpse of how the signals will be looking like on your logic analyzer with the BBU, and also it will be helpful for forming a plan on which hypotheses need to be tested/verified about signal generation.

@Kai Robinson Do you think you could post the updated BOM and board diagrams/photos in your Macintosh SE Reloaded repository?

 

[maceffects]

@quorten, thank you!  This is some useful information.  I think with the Verilog data combined with a logic probe, this shouldn't be much problem.  Looks like you've done tons of work.  Thanks for all your efforts.

 

[asicsolutions]

@quorten Looks like you've got quite a bit of stuff there. I've got a copy of VCS, so I'll get it up and running there. I owe Synopsys a SNUG paper for loaning it to me, so maybe this might end up there. I'll take a look and see what I can gather. I immediately focused on one of your points, in that if we chose to implement this in an FPGA, we should probably make a daughter card and put on memory. It would cost peanuts to put down all the memory the MAC could use, probably in a single chip  

This looks like it could be fun.

 

[asicsolutions]

@maceffects

These folks seem to have one. Anyone live close? Do you think they'd let someone borrow it?

https://www.freegeektwincities.org/

Just saw this was already picked up. 

 

[ScutBoy]

Just saw this was already picked up. 

That's me. I'll pick up the machine and get the logic board with the BBU and PDS card to Maceffects. I'll take some better pictures and share them here when I do.

 

[maceffects]

@ScutBoy Seems like you had everyone worried it would be lost, ha!  Thanks again for snatching that up.  I really only need some high-res photos of the PCB.  I really hope that it is through hole and without anything too advanced on the PDS card itself.  If it is through hole, reverse engineering will be easy.  If it has more than a few SMD and 4+ layers it will be beyond my skills and I'll have to send it out.

For completeness sake here is a discussion about the Mac SE Brainstorm.  I love that, unlike most accelerators, it actually speeds up the bus.  To me that makes it better than a 68020 or 68030 (at lower speeds) since there is still a bottleneck. 

 

[Kai Robinson]

Well... To close out 2020, I got some new simms to go with this project... I ordered a batch of 50 and got a stencil.

Tight fit in new sockets, but they do fit

 

[maceffects]

@Kai Robinson they look lovely.  Excuse my potential ignorance, but isn’t the SE limited to 1mb SIMMs?  And if I recall many Macs of this era didn’t like 3 chip memory SIMMs.  If you think we have a way to allow more memory, I’d be very keen on learning more. 

 

[Phipli]

@maceffects

Were there ever any SE memory upgrades that didn't use virtual memory (and so also need an MMU)?

 

[Kai Robinson]

@maceffects they'll never use all 4mb - they'll read as 1MB SIMM's (as the needed address lines aren't connected at all.) - You also don't have to install the 3rd parity chip either. Seem to be working great in other systems of the vintage. Can't see why they wouldn't work for that - but even if not...well i now have enough SE/30 boards to get going

 

[max1zzz]

@Kai Robinson Have you found a source of the DRAM ic's for those sticks? I looked at building some a while back but couldn't find the IC's for anything approaching a reasonable price....