Mac II PowerCache Adapter - divining its unique secrets . . . maybe.

[Trash80toHP_Mini]

A long time ago in a lifetime far removed, mcd and I were talking about the PowerCache Adapter Cloning project. On a hunch, I figured that teasing apart the signals on his Macintosh II adapter between discrete 68020 and MMU components and the IIci cache slot might be very useful in gaining an understanding of the much "simpler" 68030 PDS adapters. So he sent it along to me and I dug it out from under the heap yesterday:





*Solder side mirrored to match topside.*

Plan of attack is to de-solder all components and re-create all traces leading to the three ICs and then those heading out to the IIci Cache Slot Adaptation in Illustrator as usual. I'm going to add a new twist to the process this time around though. I'm planning to spread the ICs out a bit so I can label them like in a "real life" schematic diagram for those of you who are good at deciphering such things. That way someone might be able to help me without needing to decipher my visually oriented artwork schematics.

So, whatcha think,? Am I possibly right in guessing that this will turn out to be illuminating as to the processes of the PowerCache adaptations?

 

[Trash80toHP_Mini]

OOPSIE! Accidental post, but I might take this opportunity to give my thanks to jimjimx for donating his IIx adapter to the cause. It has the cutest little 20 pin GAL on board to translate directly from 68030 to IIci cache slot. Since DayStar went belly up back in 1999 cloning that Gate Array Logic IC won't be a problem.



Local comrade Techfury90 is already working on that angle of the project. [] ]'>

edit: does anybody know the package type of the GAL offhand? Gotta find an SMT Wire Wrap socket or SMT to Wire Wrap DIP pin prototyping adapter for it.

 

[Trash80toHP_Mini]

Bug report: I was editing the post above and the post below and the forums software kept swapping the attachments a/o not displaying the images when I posted the edited versions. Making a new post for the one below was the workaround. GREMLINS!

8-o Now that was really  unexpected! From the standard edit window the balance of this post reapeared!

View attachment 11261

 

[Trash80toHP_Mini]

First pass at mapping the traces between components. Gotta find the original, unscanned PDF for the MC68851 user's manual to snag the postscript data. Notice how clean the MC68020 pinout is by comparison, that rocks!



It's nice to be working with well documented stock parts instead of Apple's ASICs for a change. If needs be, there's always the Macintosh II schematic available for purchase. [] ]'>

 

[Trash80toHP_Mini]

IIsi PowerCache TwinSlot adapter:



I'm wondering if the 20 pin GAL pads on the right might be what it takes to convert the PDS of the SE/30? Nah! that would be just too sweet. :

Note the involvement of signals between FPU pads and 28 pin GAL. That may account for bumping the GAL's 8 input pins and 8 input/output (configurable) pins to ten bits(?) of I/O. Dunno, there's much food for thought along all these lines of research.

 

[jimjimx]

The one I sent is a bit different….

--jim

 

[jimjimx]

Sorry for the scratches & dust, but….

I bought that from the Salvation Army, for $30 in 1996, inside

of a Iix (509 N. Union, they haven't had Apple things in a decade….)

… And…  It was WITH the turbo 040… 

it was unseated from the 030 socket, just flopping around in there,

(if you haven't seen the inside of a II, or Iix, there are 6 NuBus slots,

which is a lot of room for things to "flop around")

but I bought it anyway, even though, at the time, my main computer

was a IIx. I knew it was something special….

After dialing up the "Innernet" on the telephone, I quickly learned that I had a quadra Iix.

And this is where we end up…..

Myself, and many more, …. WHY DID I GET RID OF THAT"

I had the turbo 040 in my GFs Iici for a while, and I knew it was over when she

gave it back to me because she had a PPC 7100...

--jim

 

[Trash80toHP_Mini]

Yeah, I'm afraid to even read the old collector's regrets thread. :-/

When your card arrives I'm going to break out another thread about reverse engineering it. Gotta keep the decks clear of as much tangential distraction as possible, HEH! Listen to me, the tangent  .  .  .

meanwhile, back at the pickle plantation: Found a very interesting read about the PMMU, apparently it's more capable than the on die version packed into the 68030.

http://blog.ehliar.se/post/58268464354/the-motorola-mc68851-pmmu

This has me wondering what might come out of this particular line of research.

 

[jimjimx]

You probably have this already, but I found the programming sheet with block diagram here;

http://mil.ufl.edu/4712/docs/introgal.pdf

 

[Trash80toHP_Mini]

Nope, didn't have it so thanks very much. Went over all this project with techfury this evening, this is more his kind of thing. I just do electron plumbing at the PCB level to do a visual flowchart kind of schematic development. He gets into the nitty gritty of addressing, interrupts and knows how all the different ICs do their thing. This willl be right up his alley.

Luckily, he has no problem reading the trace analysis of the connections between ICs that I do in Illustrator. He's really looking forward to seeing the interaction of the signals between the discrete 68020 CPU and its matched 68851 PMMU and how they break down between the PAL, GAL and ROM on their way to adapting the two sockets to a IIci Cache Slot.

It's gonna be an interesting week or two coming up! Most especially so after your adapter arrives for comparison of how the 68030's unified CPU/PMMU conversion is done in that one little GAL vs. the Macintosh II adapter's five component minuet.

I work best in collaboration, so we got a lot covered and I had tons of fun!

Had a lot of fun with carboy before tf got here and then when we tried (unsuccessfully) to get OS9 installed on the SATA drive from his QS. Oh well, not everything needs to be just so to be perfect. [] ]'>

 

[Trash80toHP_Mini]

Got tired of looking for a "clean" .PDF with the postscript objects. Turns out the 68020 art had a lot of unnecessary boxes in the layout, so cleaned that up and built the PMMU pinouts. If any traces go to the pins with gobbledegook names under a line, I'll edit that in if and when.



Next up is depopulating the PCB and scanning in the traces on both sides, much easier than usual!

edit: gotta fix that pinout's name to MC68851, right now it thinks it's an FPU! :-/

 

[jimjimx]

edit: does anybody know the package type of the GAL offhand? Gotta find an SMT Wire Wrap socket or SMT to Wire Wrap DIP pin prototyping adapter for it.

It's a 20 pin PLCC. I'm not sure if it'll fit into a socket without a lot of cleaning up, and re dressing of the pins.  Although not hard

I googled images for "20 pin PLCC socket" and found tonnes of 'em.

 

[Trash80toHP_Mini]

Thx!

I got started on the de-soldering project. I'm getting used to the new gun, but I'm wondering if the hot air gun might be faster/less damaging? It would likely just open up a new can of worms when it comes to cleanup. Probably not a good idea for thruhole work.

 

[unity]

If you need a working card for further testing at some point, to barrow, I have one I can pull out of my II.

 

[jimjimx]

Thx!

I got started on the de-soldering project. I'm getting used to the new gun, but I'm wondering if the hot air gun might be faster/less damaging? It would likely just open up a new can of worms when it comes to cleanup. Probably not a good idea for thruhole work.

To me, it looks like it's surface mounted. I would try to get it off by fluxing all around it, and wicking off as much solder as you can, repeat, than try to remove it with a re work station.  

edit.  I just took a look at the last pictures I took,  and it should be easy with a rework station, after fluxing & wicking, 

 

[Trash80toHP_Mini]

Yep, for that surface mount GAL, I'll definitely be firing up the hot air rework station. For the time being, I'm practicing with the vacuum de-soldering gun on the connectors of the two adapters. I want to be able to take really good pics of both on the copy stand before I remove the active components. I'm taking breaks to get things set up in Illustrator for three or four interlocked projects at once as the inspirations strike me.

Spent last night with the passive adapter/riser card for the Project30 feasibility tests. It also just happens to be the prototype for an almost identical unit to be used in the SE/30 for PowerCache adaptation trials. Things could hardly be working out better on that front.

I planned for the riser to be as large and as tight a fit as possible.The IIsi NuBus adapter is smack up against the upper level chassis return and the riser is smack up against the front of the floor plate. Next step is to file a pair of very shallow V notches in the SE/30 chassis for lateral stability of riser and adapter.

The Plexi prototype doesn't have any fit or stability issues as it's designed to be an SE/30 expansion card from the get go. I can't bolt the two Project30 cards to the mounting tabs as in that case due to extreme space limitations.

 

[Trash80toHP_Mini]

If you need a working card for further testing at some point, to barrow, I have one I can pull out of my II.

Thanks, I'll keep that in mind. What I really need and have been all but begging for would be any and all pics I can get of actual SE/30 and IIsi PowerCache adapters with a single IC adaptation. What would be a dream come true would be for someone to buzz the connections to go along with pics of the traces on the adapter PCB. As yet, nobody has responded. :-/

edit: when it comes to dreams come true, if anyone would put together a shopping list for me so I can order IIsi caps from DigiKey, that would be greatly appreciated. I need to get at least two of the four up to snuff for this PowerCache adapter project and more urgently, for baseline testing the passive riser/adapter for the NuBus in SE/30 project. It makes no sense to start testing the passive riser/NuBus card combo in the SE/30 before verification that it works in the IIsi!

 

[herpderp]

If this is still an active project, I may be able to provide some help (I was a Mac nerd a long time ago, used to have all these chips memorized)  I also have a II Daystar adaptor like the one above if you need it for testing or parts.  The diagram is labeled wrong, that's a MC68851 not a 68881.  The chip is a PMMU and basically maps logical memory address to physical ones (LA = logical address, PA = physical address, etc)  FC is function code, iirc.  One of the first books I picked up as a young nerd was my dad's MC68020 programmer's manual.  I think it's still kicking around somewhere, along with the MC68851.  Let me know, I can dig all this up.  I miss this stuff

 

[herpderp]

Also, this may be useful:

https://html.alldatasheet.com/html-pdf/4166/MOTOROLA/MC68851/258/1/MC68851.html

Oddly enough, all the signals are not accounted for on the PDS adaptor, some are missing, most notably the bus control signals (DS/RW/DSACK, etc)