• 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.

Macintosh SE/30 Schematics (modernization effort)

Bolle

Well-known member
I have started with matching the footprints on most of the Apple ICs on the SE/30 board to pinouts.

If you send me the Kicad files I can merge them what I got and start working on arranging stuff on a draft board.

I would however go for a few changes... as jt already mentioned four 72pin SIMMs with 32MB are cheaper to come by than eight 16MB 30pin ones.

Also implementing space for four 29F040 ROM chips as a replacement for the ROM SIMM is a good thing I think. The ROM SIMM and socket often get hit by the battery so it just makes sense to go for a cheap chips-go-right-onto-the-board kind fo replacement.

 

maceffects

Well-known member
Where's @maceffects and his engineer friend when you need him... ;-D

With my limited understanding of this from a technical level, I can acknowledge this would be a pretty hefty undertaking. Creating a board like this was much easier for Apple because they'd done it before with other machines, and had lots of skill and knowledge already in the company, as well as access to other resources and money.

It would be interesting to see what could be done with a partial redesign. The Classic II logic board fits a 68030 into a board half the size, though with other limitations.

As with all these projects, scope is important but so is having the resources, namely money, time and knowledge.

I for one, would happily contribute to a kickstarter for such a project - unfortunately I wouldn't be able to assist much more than that.
I thinking making a reversed engineered PCB isn't hard at all, however, the real trick to making a new board would be those VSLI chips and their proprietary information.  If someone ever cracks those (the SE/30 has several) then making new boards isn't hard, minus sourcing the 68030.  I can even make the metal brackets that attach fairly easily. 

 

elemenoh

Well-known member
I'm going through everything again and creating bus lines for address lines and other grouped, but not connected things. This should help with readability and linking between sheets. I've got a couple done already and should have the others done in a few days.

Screen Shot 2020-05-21 at 4.09.39 PM.png

 

LaPorta

Well-known member
This will be unbelievably instrumental in helping me diagnose my SE/30 spare board’s insane ADB issue.

 

elemenoh

Well-known member
OK I finished fussing with the address lines and the KiCAD files have been cleaned up, archived and added to the shared folder. The topmost sheet just links to the other sheets at the moment. I didn't import any of the global pins onto it, but that should be quick and easy if it's helpful to you.

Go forth and be awesome, fellow nerds. If you find these helpful in your projects, please do me a favor and check my post history and help me out with my various unsolved mysteries :)  

https://drive.google.com/drive/folders/10VO6n72vsn52xqMyOgSZpi11tbyv7_Yx

 

Trash80toHP_Mini

NIGHT STALKER
I thinking making a reversed engineered PCB isn't hard at all, however, the real trick to making a new board would be those VSLI chips and their proprietary information.
I wonder if the formulas are in the donated Apple archive and if so, the Woz might be approached in terms of licensing or release into the public domain for educational purposes?

If someone ever cracks those (the SE/30 has several)
As @techknight confirmed, there are a lot of damaged boards out there just begging to be donors. Breaking 30 pin SIMM sockets is becoming a real problem. Upgrading the design to more readily available(new?) 72pin Sockets/SIMMS would be a no-brainer for me.

.  .  .  making new boards isn't hard, minus sourcing the 68030. 
I've done research on embedded 68030s for my accelerator projects. They're out there, Bolle said the last batch out of Freescale is available. Checked it out and no indication of incompatible instruction set was noted for it as was the case for others on the list. It's SMT, so you'd want to move it to the side and provide cutout circuitry and PGA socket for the inevitable reproduction of the SE/30's custom accelerator. Clock rate's fine for SE/30 and the Performer, but not high enough for the likes of Bolle's accelerator reproductions.

Slavishly recreating the Apple PCB design makes no sense to me at all.

 
Last edited by a moderator:

Trash80toHP_Mini

NIGHT STALKER
I would however go for a few changes... as jt already mentioned four 72pin SIMMs with 32MB are cheaper to come by than eight 16MB 30pin ones.

Also implementing space for four 29F040 ROM chips as a replacement for the ROM SIMM is a good thing I think. The ROM SIMM and socket often get hit by the battery so it just makes sense to go for a cheap chips-go-right-onto-the-board kind fo replacement.
Missed your post, glad to see you're on board already! On board(socketed?) ROM is a great idea. What do you think about provision for embedded CPU/cutout injection with PGA socket combo?

 

maceffects

Well-known member
I wonder if the formulas are in the donated Apple archive and if so, the Woz might be approached in terms of licensing or release into the public domain for educational purposes?
I don't ever see that happening.  Honestly, knowing Apple I'd be surprised if they still have them (or at least know where they are).  With a recreation effort for a board, something to consider is that we are blurring the line between original and emulation and if the effort/cost is high, emulation might be better. 

 

Trash80toHP_Mini

NIGHT STALKER
Don't know if it's in the treasure trove Apple donated or not, but it wouldn't hurt to ask. We'd be running period hardware, released formulas or no in the case of "repaired" junker boards of which I have one myself. That a few tweaks would be made to utilize a more readily available, just a few years later RAM config is no biggie, nor is convenience-grading the ROM. Emulation would be sticking a tiny current thing on a fake board with no possibility of using the expansion card toys we all love a/o covet and make our beloved accelerators pointless.

BTW, building an entirely socketed board would be fabulous for component testing in debugging and preservation efforts. Doing a hybrid SMT pad/thruhole component capable design might be worth investigating?

 

joshc

Well-known member
Finding the knowledge will be hard, I would be surprised if Apple kept much of this stuff. When you mention getting Woz involved, do you mean for the IWM knowledge? I don't think Woz had anything to do with the SE/30, he left Apple a few years before.

I found the names of the engineers of the SE/30 (it's in the ROM) but haven't been able to make contact with any of them yet.

 

Trash80toHP_Mini

NIGHT STALKER
When you mention getting Woz involved, do you mean for the IWM knowledge?
Nah, nothing like that, I just wish he'd do an end run around Apple's reflexive NO mentality and legions of lawyer drones. I seriously doubt the curators of that institution know much about Apple's archival donations in this kind of technical sense. The Woz and his Steveness both were long gone by the time of the SE/30.

 

trag

Well-known member
I would however go for a few changes... as jt already mentioned four 72pin SIMMs with 32MB are cheaper to come by than eight 16MB 30pin ones.
I think that if  you did this, the 72 pin SIMMs would have to be installed as identical pairs.   The SE/30 supports 2 independent banks of RAM, not 4, and so probably lacks independent RAS lines for four independent 72 pin sockets.

Even with identical pairs, it would be the equivalent of "composite" SIMMs.   A little messy logically.

 

Trash80toHP_Mini

NIGHT STALKER
Even with identical pairs, it would be the equivalent of "composite" SIMMs.   A little messy logically.
That's one reason why a pair of 64MB SIMMs, one per bank have replaced pairs of 32MB SIMMS for each bank in my related project. Combining a pair of 32MB or a quartet of 16MB SIMMs to fill a single 72pin slot was a thing, that's for yet another research project. Those boards really don't seem all that logically messy  .  .  .  but what do I know. ::)

 

johnklos

Well-known member
That's one reason why a pair of 64MB SIMMs, one per bank have replaced pairs of 32MB SIMMS for each bank in my related project. Combining a pair of 32MB or a quartet of 16MB SIMMs to fill a single 72pin slot was a thing, that's for yet another research project. Those boards really don't seem all that logically messy  .  .  .  but what do I know. ::) 
One 128 meg SIMM would give you two banks of 64 megs, with less fuss.

 

Trash80toHP_Mini

NIGHT STALKER
Haven't looked at 128MB as my target machines are IIsi and Quadra 950 which have discrete 64MB banks along with the SE/30's pair of such banks. IIRC, the last I looked I could get a pair of 64MB SIMMs for less than a single 128MB SIMM, has that changed?

 

balrog

Member
I would also recommend redesigning the battery circuit to use a safer chemistry. The chemistry that the 3.6 volt lithium batteries use is extremely toxic in addition to being corrosive.

 
Top