>Date: Thu, 10 Oct 2002 14:21:42 +0100
>From: Mark Benson <[EMAIL PROTECTED]>

>OK, consider me significantly confused an somewhat nose-out-of-jointed.
>I guess it;s a little harder than i thought ;).

It would be challenging, but not necessarily impossible.  You really 
need to open up an AEHD+ if you want to know how hard it would be.

Phillip covered the task excellently.   Here's basically the same 
thing in list form, which might make it seem a bit less daunting:

1)  Identify the parts in the original.  Hopefully they're all 
marked.  They probably are.  One or more of the parts may be obsolete 
and difficult to find.   You can often get help in the news group 
sci.electronics.components.   Some of the denizens of that group are 
dealers who sell obsolete parts--but the prices.  Ouch.

2)  If any of the parts are PLDs (programmable logic devices) you'll 
need to figure out their contents.  Examples of PLDs are GALs, PALs 
and FPGAs.   You'd likely find one or more PALs or GALs.   Internally 
they contain logic gates (the stuff computers are made of) and a 
telephone exchange of possible connections.  When you program a PLD, 
most of the extra connections are "blown" and the remaining 
connections create the custom circuit the user wanted using the gates 
in the PLD.

The same circuit which is built with a PAL or GAL can be built with 
logic chips but it's much more compact to get the equivalent of eight 
or ten (or a thousand) logic chips in a PAL or GAL and then program 
it, rather than trying to squeeze a bunch of chips on a board.

PLDs can be read so that you'll know their programming, but there is 
usually a "security fuse" which, when blown, prevents reading the 
contents.  At that point, you must input data and read what comes out 
of the chip to figure out the contents.  This can be very straight 
forward if the chip was programmed for simple combinational logic (no 
feedback loops) but most PALs and GALs have registers that can be 
used to create "state machines" (they use feedback loops, i.e. they 
have some memory).  If you have a PAL or GAL which was programmed as 
a State Machine it can take significant brain sweat and time to 
figure out how it's programmed.

This task, if you must do it, is by far the most complex part of the 
whole operation.

3)  If any of the parts are non-volatile storage, you'll need to 
remove them from the board and read them out.   Non-volatile storage 
would be a ROM, EPROM, EEPROM or Flash chip.  Given the era and the 
probably low volume of the poduct, I'd guess EPROM in this case. 
Reading the contents is very straight-forward.  The difficult part is 
removing the chip from the board but if you took up that project you 
could probably find one or more skilled solderers willing to help 
with that part.

Once you have the chip off the board, you simply need access to an 
EEPROM programming machine.  There is a vast variety of these 
machines, but basically they have a socket on top (usually an 
interchangeable socket for different packages) and they can program 
EPROMs, EEPROMs, Flash chip and often PLDs as well.

Programmers cost from a couple hundred on up to whatever you want to 
spend.  Reaslistically a pretty nice one seems to cost about $400 - 
$600 but definitions of "nice" may vary.   However, there is probably 
a local PC shop near you where you can get teh chip read and copied. 
PC shops have programmers to reprogram BIOS chips in PCs.   You may 
need to supply blanks of the chips (for copies) and an adapter 
socket, if, e.g., the PC shop only programs DIP packages and your 
example chip is in a PLCC package.

4)   If you overcome 2 and 3, then it's worth tracing the circuit 
board of the device to develop a schematic.   The board may have more 
than two layers so you really need to check connections with a 
continuity meter and it helps to have some idea of how the device is 
built so that you can make guesses about what should connect to what. 
This part is not difficult but can be very tedious.

5)   Once you've got the parts are the programming together you'll 
want to test it.   You can do a wire wrap kind of thing on perforated 
board or use a proto-board but neither of these is all that great if 
any of your chips are not DIP (straight pins sticking out). 
Another option is to get a couple of proto-type circuit boards built. 
This will likely cost you in the neighborhood of $150.

6)  You can use Osmond PCB to lay out the board. 
http://www.swcp.com/~jchavez/index.html  It's pretty nice software 
for that kind of thing, though the auto-routing function isn't too 
useful on multi-layer boards.  Still, it works on the Mac, and there 
is a 68K as well as a PPC and OSX version available.

7)  Sierra Proto Express seems to have the best deal on proto-type 
circuit boards.  They'll build you a two layer board for $31 each 
(minimum two) or a four layer board for $51 each (minimum two). 
There's a service charge of $20 and shipping so the total is about 
$35 more than the board cost.   www.4pcb.com has a similar special on 
two layer boards.

Conclusion:    This is probably too much project as a first project 
if you don't have any experience with hardware hacking.   There are 
several tasks, none of which (except the PLD analysis) are all that 
difficult.  But if you're starting from a base of not knowing how to 
do any of them, you'd have a lot learning curve to overcome.  It's 
better to learn each task as part of some other project and then put 
them together into a big project after you have more experience.

Heh, heh.  Someone could buy the one on Ebay, open it up, do a high 
resolution scan of the thing, and we can make reverse engineering it 
a Compact Macs List project.  :-)

If there aren't any PLDs in the thing, I'd say that it would be 
pretty straight forward, with the most difficult part tracking down 
any rare or obsolete parts.

I'm trying to clone the external floppy mechanism for the Outbound 
Laptop Model 125 at the moment.   It used a "standard" PC notebook 
floppy mechanism with a circuit card controller stuck on the 
interface.   This is interesting, because I've heard it said (well, 
read it written) that it isn't possible to get a PC floppy drive to 
do the 800K floppy dance that Mac floppies do, yet this one does.

The bad news is that there is a GAL 16V8 in the thing which is 
programmed in register mode (is a state machine, uses feedback).   So 
before I go any further I'll need to figure out how the thing is 
programmed.  On the bright side, the 16V8 is about the smallest PLD 
you're likely to run into.   There are only eight inputs and eight 
outputs.  The thing could be much larger.

The other pain is that the thing uses a 37C65 floppy controller chip, 
92C32 floppy controller (somethign to do with clocks), and a Xicor 
9103 digital potentiometer, amongst others.   These are obsolete 
parts and I'm getting quotes of $6 - $11 each for the various chips. 
You'd expect to pay $2 - $3 for chips of their complexity 
ordinarily--at least I would.  Maybe I'm overly optimistic and cheap. 
There is also a handful of regular logic chips (74 series) and an 
85C30.   Anyway, suddenly this project would cost more than $20 each 
just in parts before even considering the circuit board and such. 
Sigh.

Maybe I should take Phillip's advice and rather than reverse engineer 
it, design it anew with more modern parts.  That might get the 
component count down.  The problem is that I don't really know the 
specs to which it was designed.   The interface back to the Outbound 
laptop was some kind of proprietary interface.   It must be some kind 
of interface that the 85C30 understands because that's the first 
thing (logically speaking) on the interface card.  And the SCSI 
adapter, which connects to the same interface uses a 53C80 SCSI chip, 
so the interface must speak something a 53C80 understands as well.

Jeff Walther

-- 
Compact Macs is sponsored by <http://lowendmac.com/>.

      Support Low End Mac <http://lowendmac.com/lists/support.html>

Compact Macs list info: <http://lowendmac.com/lists/compact.shtml>
The FAQ:                <http://macfaq.org/>
  --> AOL users, remove "mailto:";
Send list messages to:  <mailto:[EMAIL PROTECTED]>
To unsubscribe, email:  <mailto:[EMAIL PROTECTED]>
For digest mode, email: <mailto:[EMAIL PROTECTED]>
Subscription questions: <mailto:[EMAIL PROTECTED]>
Archive:<http://www.mail-archive.com/compact.macs%40mail.maclaunch.com/>

Using a Mac? Free email & more at Applelinks! http://www.applelinks.com

Reply via email to