• Updated 2023-07-12: Hello, Guest! Welcome back, and be sure to check out this follow-up post about our outage a week or so ago.

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

cheesestraws

Well-known member
Ah, one always misses one.  Cheers.  I'm following you at a distance: got all the resistors and diodes on now, capacitors next...

 

Kai Robinson

Well-known member
Yeah i just ordered all the axials - amazed that they're almost all unavailable now from most, if not all, suppliers. Might have to change the new design to use radials instead, and from a quick glance at sprint - that should be easily doable. Hell, i might even go SMT for a giggle, assuming the prototype works! Even if all i get is a beep on power on and a raster, i'll be happy, even if nothing else works!

 

quorten

Well-known member
Wow, I feel like I've come late to seeing this development.  1/2 what I was wishing would happen.

The other half, modern schematics for the Macintosh SE.  Well, I've seen what @elemenoh has done for the SE/30 schematics and these two separate projects together really makes me want to go and do likewise for the Macintosh SE, if no one else has yet stepped up to the task.

Interesting that I thought the sequence of events would happen the other way around, but well, some things just turn out a bit more chaotic than ideal.

 

CircuitBored

Well-known member
What an astonishing project. It's dragged me out of lurker mode... I would absolutely love to be the owner of a "brand new" SE someday!

For now I am sadly not really smart enough to contribute much brainpower to this project but I am very good with soldering and have an enormous amount of experience working on/with old Macs.

If at some point you need someone UK-based to help with assembling these things in bigger batches then I am your man. Congrats again on a stunning show of ingenuity here.

 

max1zzz

Well-known member
On the subject of the axial capacitors, Personally I would just leave the board as is, If needed the axials can be substituted for normal radial lead ones without much trouble. This is what I ended up doing on all my Mac II series machines due to the difficulty of getting the axials

 

quorten

Well-known member
Personally I would just leave the board as is, If needed the axials can be substituted for normal radial lead ones without much trouble.
+1, I would also agree.  I'm guessing probably the main reason for doing otherwise would be as part of the first stage of Tindie sales to recoup the costs.  But otherwise, for the publicly available designs, definitely I would prefer 1:1 so that we can emulate the good old feeling of retrofitting unoriginal capacitors.

 

Kai Robinson

Well-known member
+1, I would also agree.  I'm guessing probably the main reason for doing otherwise would be as part of the first stage of Tindie sales to recoup the costs.  But otherwise, for the publicly available designs, definitely I would prefer 1:1 so that we can emulate the good old feeling of retrofitting unoriginal capacitors.
Actually the reason is actually because the axials are getting hard to find, and twice as expensive as the radials.

Like i said, ultimately, i'm doing this project to do something for the community as a whole.

Obviously it'd be nice to put these boards on tindie and recoup the costs of development, but even if i don't - ehh, i've still learned new skills and tried to help. 

 

quorten

Well-known member
Looking at the fuzzy Macintosh SE schematic right now, a first try at a grayscale conversion and gamma-adjusted image rescaling seems to make at least some parts easier to read, albeit at reduced resolution.  I'm looking to try my hand at redrawing this if I think I can see sufficient detail from the originals.

 

Kai Robinson

Well-known member
One thing you can do, is take the sprint layout of each layer, and import that into KiCad - then work backwards to a schematic. If you'd like me to take a screenie of each individual layer in sprint, you're welcome to them. 

 

Kai Robinson

Well-known member
Ok - i've done a final revision of the silkscreen, and added the original 1977 Apple logo as the original Mac SE board used (despite the fact that by then, Apple used the Apple Garamond font...which was...wierd.)

I've also changed the size of the via's to 0.8mm with an internal diameter of 0.4mm, and tented them by excluding them from the solder mask - should look much more professional compared to the prototypes. 

macseboard-final-rev12a.JPG

macseboard-final-rev12a-rear.JPG

 

quorten

Well-known member
So I decided to split up modernizing the Macintosh SE main logic board schematics into two stages.  The first stage is for my own proof that I can actually read anything meaningful off of the original scans: trace over vector graphics and typed text on the original schematic image.  It's almost finished, but not quite, it has a few text transcription errors (due to the difficulty of reading), but this is what I have so far.  Fortunately, there is a certain degree of redundancy in the text labels, and seeing the big picture clearly really helps too.

As it turns out, the reverse engineering on the PCB layout will indeed be helpful in correcting minor errors.  Like, resistor identifiers and values, having zero ambiguity on that.  Also, other things can easily be solved by consulting other tech specs like datasheets, the PDS slot documentation, and so on.

View attachment retrace_se_mlb_schem.pdf

 

quorten

Well-known member
Also, you'll have to make the second schematic page from scratch, seeing as there seems to be none on the internet.
Judging from the 1/3 in the bottom right corner and the tiny text above it that appears to say "SHT" (sheet) in scans of other pages, I'd assume page 3 of the main logic board schematic is also missing, unless you've turned up something for that.  At least we have some hints about which components go on which pages based off of the numbers in parenthesis on the arrows to connections on other pages.

I'm feeling pretty confident now about my retrace of page 1, I'll be able to move to drawing up full MLB schematics in KiCad soon.  I still need to check the PDS pinout with a second source since I had trouble finding it.  Could someone provide a link?  My work is up here on GitHub.

https://github.com/quorten/macsehw

Link straight to the redrawn PDF.

The BBU pinout is definitely helpful!  Where did you get it?  It uses a slightly different terminology than the original deciphered schematic.  Also there are a few technical differences that I've spotted.  Possibly some of these might be a (deliberately) introduced error on the original schematic?

  • RDO0, RDO1, RDO2, ... (instead of RD0, RD1, ...)
  • 11 *ENTD1K mystery symbol (indicate inverted signal, replace EXTPDS placeholder)
  • 12 *EN245 (indicate inverted signal)
  • 13 ROW2 (RAM configuration pin, instead of generic VCC, though maybe it's really the same thing)
  • 17 MBRAM (RAM configuration pin, instead of *EOP for 53C80 SCSI controller)
  • 32 *IRQ (indicate inverted signal)
  • 44 C16MRSF2 (16 MHz clock descriptive label instead of FCLK)
  • 45 *SELSCC (indicate inverted signal)
  • 50 SNDRST (SNDRES) (replace PB7 placeholder)
  • 53 VIDPG2 (replace PA6 placeholder)
  • 52 *EAREN (replace EXTPDS placeholder)
  • 82 C2M (2 MHz clock for 74F257 chip select, as opposed to EN257 label)
  • 60 *HSYNC (indicate inverted signal)
  • 61 *VSYNC (indicate inverted signal)

An important point, the Reddit you've linked for a replacement RTC does indeed have a minor error, the time in seconds should be separate from the 256 bytes of PRAM.  It's worth cross-checking this source code with the corresponding source in a few different software emulators before sticking one of those chips in a real hardware system, not to mention that it still does need a bit of work as originally mentioned.  I started making a few changes in the copy in my GitHub repo with the schematics.

Really, both the RTC and the BBU, neither of these are hard to re-implement, it's just that... it requires a bit more understanding on the software side of affairs.

@Kai Robinson One important thing missing from the printed circuit board reproduction thus far: looks like you forgot the B1 reference designator.  Sure, here on 68kMLA we all know it by its much more affectionate name, the PRAM battery, but nothing quite completes the experience of removing and replacing those beasties without the unapologetic reference designator to assign an ID number to yet another nameless battery.

 
Last edited by a moderator:

Kai Robinson

Well-known member
Mostly a combination of the Macintosh Classic schematic (a cost reduced SE that also uses the same BBU), and backtracing pins on the board. Sprint has this wonderful 'test' feature - click on any pin of the BBU, and you can see exactly where it goes. I'll send you the sprint layout file if you like, you can take a look yourself.

 

cheesestraws

Well-known member
Nice.   I am waiting for the last value of capacitor to arrive, then I'll be able to make progress again.  Those axials are a right pain...

 
Top