Custom 68000 esed on Pre-Production 128k logic boards?

22 posts / 0 new
Last post
JDW
JDW's picture
Offline
Last seen: 5 months 2 days ago
Joined: Mar 17 2006 - 19:08
Posts: 180
Custom 68000 esed on Pre-Production 128k logic boards?

I recently came across this EBAY Auction of a pre-production 128k logic board from 1983. In the photos you can see the CPU marked "SC87839L."

I wrote to Freescale about the chip and they replied as follows:

Unfortunately, the part number you have referenced was not a standard Freescale/Motorola Semiconductors device (see the "SC" prefix). It was a proprietary device manufactured by Freescale Semiconductors for a specific company to their own specification. The proprietary devices (also documentation) become confidential and can only be obtained through the company for whom we manufactured the products.

These devices cannot be crossed referenced.

If this device was found in an end product, please contact the manufacturer of the product or equipment for information. Due to nondisclosure agreements with this company, we are not authorized to release information. We can only release data on standard products. Please accept our apology for any inconvenience this may cause.

Have a nice day and thank you for contacting the Freescale Semiconductor
- Customer Care Center.

Of course, this reply simply means I would need to contact Apple about the chip. But certainly they would not be forthcoming with details. This is obviously a custom variant of the 68000 chip, but what kind of variant? Anyone know an Apple engineer who worked on the logic boards back in '83?

Jon
Jon's picture
Offline
Last seen: 12 years 10 months ago
Joined: Dec 20 2003 - 10:38
Posts: 2804
I don't know anybody of that

I don't know anybody of that sort, but I do know a guy who worked in fabs at that time. I only see him once a month, so I'd have to remember to ask in a couple weeks. He has old first-run 6502s and such. He's known guys like Chuck Peddle and Chuck Moore.

dankephoto's picture
Offline
Last seen: 9 months 1 week ago
Joined: Dec 20 2003 - 10:38
Posts: 1899
sure looks to me like a common 512K upgrade

Except for the different markings on the CPU (which is obviously a MC68000 otherwise the machine wouldn't even boot), it sure looks like a bog-standard 128K MLB that's been upgraded to 512K, with 'pirate' Plus ROMs as well I reckon. Pretty common setup in my experience, I've got a couple of very similar boards. In fact nowadays its tougher to find 'virgin' 128K boards that weren't upgraded to 512K.

dan k

coius's picture
Offline
Last seen: 10 years 2 weeks ago
Joined: Aug 25 2004 - 13:56
Posts: 1975
Re: sure looks to me like a common 512K upgrade

Except for the different markings on the CPU (which is obviously a MC68000 otherwise the machine wouldn't even boot), it sure looks like a bog-standard 128K MLB that's been upgraded to 512K, with 'pirate' Plus ROMs as well I reckon. Pretty common setup in my experience, I've got a couple of very similar boards. In fact nowadays its tougher to find 'virgin' 128K boards that weren't upgraded to 512K.

dan k

That could be true, but when were they first mass produced?
If they introduced them in 1984, but didn't start working on mass production till the middle of the year, then this very well could be a prototype. but if the machine started shipping the day it was introduced, then it would be likely that the boards were produced in 1983, and there was a build-up of them in 1983 just to release them at the beginning of 1984.

This guy could have just found a compatible CPU of it's type, slammed it on an early production board, then said it was a prototype. I didn't even think they would be able to achieve that much memory out of the box. I know apple would have shipped them with 256k, but it was too expensive to get the chips.
If they weren't shipped with 256, why would they mess with 512?

JDW
JDW's picture
Offline
Last seen: 5 months 2 days ago
Joined: Mar 17 2006 - 19:08
Posts: 180
Custom Chip

All I can say is that my Googling has found that the first two letters of this CPU's number indicates it is a "custom" chip from Motorola. It was either "custom" because Apple wanted the CPU before the 68000 was formally released or because they wanted a custom variant of the existing 68000. It would seem to be only one of those two possibilities.

I would never pay the outrageous price the EBAY seller is asking for the board, but his listing got me curious about the CPU itself.

Any further ideas?

Eudimorphodon's picture
Offline
Last seen: 3 months 3 weeks ago
Joined: Dec 21 2003 - 14:14
Posts: 1207
Re: Custom Chip

All I can say is that my Googling has found that the first two letters of this CPU's number indicates it is a "custom" chip from Motorola. It was either "custom" because Apple wanted the CPU before the 68000 was formally released or because they wanted a custom variant of the existing 68000. It would seem to be only one of those two possibilities.

The third possibility is that Apple just ordered bone-stock Motorola 68000s with custom labeling. It's not uncommon for computer and electronics manufacturers when ordering large quantitites of ICs from a supplier to ask for special labeling on them. Sometimes it's an attempt to obfuscate what's in the circuit, to make it more difficult for the hardware to be cloned by competitor, while in other cases it's just for inventory control.

Before getting too carried away it'd be worthwhile to look at a random sampling of 128K/512K Mac motherboards and see if their CPUs consistantly have generic Motorola markings on them or not.

--Peace

JDW
JDW's picture
Offline
Last seen: 5 months 2 days ago
Joined: Mar 17 2006 - 19:08
Posts: 180
Daniel Kottke

I have both a 128k and a 512k motherboard and the CPU on both is clearly marked 68000.

And while you could be right about Apple trying to hide the fact they were using a 68000, the obvious question is WHY? I would think they would want to tout that fact, seeing how "powerful" the 68000 chip was at the time. But perhaps they wanted to keep that knowledge from IBM.

Anyway, the EBAY seller of that board just wrote me an email saying he is in contact with Daniel Kottke about the matter. He didn't say how he knows Daniel Kottke. But it would be curious to hear if this yields any further details about the chip.

dankephoto's picture
Offline
Last seen: 9 months 1 week ago
Joined: Dec 20 2003 - 10:38
Posts: 1899
re: custom CPU chip

Now I could very well be waaay wrong, but IMNSHO Apple didn't put that CPU on that MLB at all. I reckon it was stuck there by some later owner of the board.

just my totally uninformed US$.02,

dan k

Jon
Jon's picture
Offline
Last seen: 12 years 10 months ago
Joined: Dec 20 2003 - 10:38
Posts: 2804
The board is clearly silk scr

The board is clearly silk screened with "68000" above the CPU socket. And the am27512dc are certainly UV EPROMs. 630-0101 shows that it is at least an early 128k board, but makes nothing clear about being such a rare prototype, worth $600+...

JDW
JDW's picture
Offline
Last seen: 5 months 2 days ago
Joined: Mar 17 2006 - 19:08
Posts: 180
Speculation

Gentlemen, I don't doubt your judgment about this board being overpriced. I am in no way interested in buying it at that price! But the photos in the auction got me interested in that chip to see if indeed somebody is playing EBAY games or if indeed Apple used a custom variant of the 68000.

Even though the board is silked with 68000, does that mean Apple could not have used a custom Motorola chip in the early boards? There is room for some level of speculation here. That's why I said that only an Apple engineer working on the Mac project back in '83 could say for sure.

Eudimorphodon's picture
Offline
Last seen: 3 months 3 weeks ago
Joined: Dec 21 2003 - 14:14
Posts: 1207
Re: Speculation

Gentlemen, I don't doubt your judgment about this board being overpriced. I am in no way interested in buying it at that price! But the photos in the auction got me interested in that chip to see if indeed somebody is playing EBAY games or if indeed Apple used a custom variant of the 68000.

Even though the board is silked with 68000, does that mean Apple could not have used a custom Motorola chip in the early boards? There is room for some level of speculation here. That's why I said that only an Apple engineer working on the Mac project back in '83 could say for sure.

Whatever the pedigree of that chip is, do note that you can buy one as electronic surplus:

http://www.lukechips.com/search/default.asp/FullPartNo/SC87839L/search/true

My personal pet theory is that whoever upgraded the memory on that motherboard snatched that chip out of a part bin at the company he worked for (which regularly ordered custom-model-number 68000s for some obscure piece of network or military hardware) to replace a bad or missing CPU. Perhaps he bought the motherboard sans-CPU out of an Apple reseller's upgrade junk pile in order to build one of these:

http://www.mandrake.demon.co.uk/Apple/ugly128.html

If you really, really want to speculate about custom CPUs being part of the Macintosh design cycle I'd suggest seeing if any of the folks on folklore.org:

http://www.folklore.org/index.py

Recall anything. My guess is "no". If Apple had gone down that path someone would of mentioned it.

--Peace

JDW
JDW's picture
Offline
Last seen: 5 months 2 days ago
Joined: Mar 17 2006 - 19:08
Posts: 180
Folklore

Actually, I thought about posting on Folklore. But no one there has even replied to my first post (see bottom 2 comments), made there a year ago! So I didn't think anyone at Folklore would care to comment this time either.

JDW
JDW's picture
Offline
Last seen: 5 months 2 days ago
Joined: Mar 17 2006 - 19:08
Posts: 180
SC87839L was definitely used in 1983 pre-production boards

It's funny how these things happen by accident...

A couple months ago I found an old used copy of Doug Clapp's "Macintosh Complete," published in 1984, on Amazon for $2.50. I had been reading through the book little by little over the past few weeks. While flipping through the book today, I happenstanced on a photo (page 204) of a pre-production Mac128 motherboard. What do you know! The SC87839L is indeed soldered into that board!

There is no visible 512k RAM hack as is the case on this EBAY auction, and the ROMs are not erasable windowed EPROMs (although they are socketed), but the SC87839L is there. This proves the board in the EBAY auction was not "hacked" by someone simpy grabbing a CPU off the shelf.

The only difference with the chip in the Clapp book photo and the EBAY auction photo is the third line of text on the chip. The 3rd line on the chip in the photo in the Clapp book reads: GN78249 -- which I assume is a manufacturing lot number or date stamp of some sort. Middle text is the same. ROMs in the Clapp book photo read as follows:

ROM HI
23258-1008
341-0220

ROM LO
23258-1007
341-0221

Interestingly, Clapp makes no mention of a custom CPU chip in his book. Despite the photo he used, he only refers to it as an MC6000 or simply "68000." But perhaps that's due to his not eyeing the photo he used? Who knows. But another interesting point is the photo used on page 215 of his book, which shows a true 68000 chip with the following markings:

MC68000L8
GN78301

Well, if any of you have contact with Doug Clapp, maybe you can get a clarification from him.

Thoughts?

dankephoto's picture
Offline
Last seen: 9 months 1 week ago
Joined: Dec 20 2003 - 10:38
Posts: 1899
more 128K logic board infos

See new thread here, didn't want to clutter up this one with unnecessary details.

the date code works thusly - eg: 8301 is the 1st week of 1983. eBay MLB has a cpu date of 8304.

The part numbers printed on the ROMs are for an original 128K Macintosh, but it is odd they should be EEPROMs. Hmm, I wonder if the inked '433U' indicates the production date? For comparison, the three 128K boards I just looked at have 054J, 054U and 304J inked on them.

Still, everything I can see about the actual item for sale is that it's a production 128K board.

*edit
Heh, I just noticed it's missing its paper serial number label.
*/edit

dan k

JDW
JDW's picture
Offline
Last seen: 5 months 2 days ago
Joined: Mar 17 2006 - 19:08
Posts: 180
Another photo from this site

Here's another photo found on this very site of an SC87839L chip used on an older Mac128 motherboard.

Eudimorphodon's picture
Offline
Last seen: 3 months 3 weeks ago
Joined: Dec 21 2003 - 14:14
Posts: 1207
Occam's Razor

To quote the Wikipedia Definition:

Occam's razor states that the explanation of any phenomenon should make as few assumptions as possible, eliminating, or "shaving off," those that make no difference in the observable predictions of the explanatory hypothesis or theory. In short, when given two equally valid explanations for a phenomenon, one should embrace the less complicated formulation. The principle is often expressed in Latin as the lex parsimoniae (law of succinctness):

entia non sunt multiplicanda praeter necessitatem,

which translates to:

entities should not be multiplied beyond necessity.

This is often paraphrased as "All things being equal, the simplest solution tends to be the best one." In other words, when multiple competing theories are equal in other respects, the principle recommends selecting the theory that introduces the fewest assumptions and postulates the fewest hypothetical entities. It is in this sense that Occam's razor is usually understood.

The *simplest* explanation for these oddly-labeled CPUs being ocassionally found in early Macintosh 128 motherboards is that Apple ordered a batch of MC68000s from Motorola under a vendor-specific part number, *possibly* with a mind towards obfuscating the Macintosh's innards. (The "SC87839L" chips available as surplus are listed with a 1983 manufacture date, which would jive with Apple perhaps wanting to be coy about telling the world it was about to go into large scale production of a "mass market" 68000-based computer. Sure, they were already ordering 68000s for the Lisa, but that was technically another division developing that, and companies can get weird and stupid when trying to be clever.) Once the Mac was in production and the 68000 became a major selling point of the machine (and developers started programming for it) there was no point in ordering any more obfuscated CPUs, but a few of them were left in the parts bin and were used in production machines interchangably with more-conventionally labeled units.

That's *much* more likely then any theory about "customized" CPUs. It's worth noting from some of the other motherboard pictures posted that sometimes the 68000 CPU in these machines has "Motorola" markings on it, while in other instances it's stamped "APPLE". The "APPLE" units clearly indicate that Apple *did* have a relationship with Motorola which extended to ordering chips with their name scribed on the top, so it seems to me perfectly reasonable that they might at some point have ordered chips with a "private" vendor part number on them as well.

If even *one* of these chips slipped into a production machine then it blows the "custom 68000" theory totally out of the water. It also just seems unrealistic to me that Apple would be investigating custom versions of the 68000 so late in the machine's design cycle. The Macintosh ROMs were largly composed of hand-tuned assembly language code, so the *last* thing you'd want to do in 1983/84 is introduce possible CPU hardware bugs. (It'd also be too late to really take advantage of any instruction set changes.)

Anyway. "Custom 68000" introduces (many) more "hypothetical entities" then "Custom Label", so it's the less satisfactory explanation to this puzzle. The fact that no one on Folklore.org recalls "the big day when the custom chips showed up" is a pretty strong argument against it as well, since the site goes into pretty gruesome detail about other Macintosh hardware design dead-ends, like the "IBM" chip.

Feel free to speculate, but the only way to *really* answer the question at this point would probably be to obtain one of these "SC87839L" CPUs and run an assembly language regression test against it (probably with it inserted into a CPU analyzer) and see if any instructions return different results then a Motorola-labeled 68000 would. (Or if any invalid opcodes do anything interesting, suggesting an instruction set enhancement, although on old cpus like that the results returned by invalid instructions can often vary between mask revisions, so you *could* get misleading results.)

--Peace

Offline
Last seen: 2 years 3 months ago
Joined: Dec 20 2003 - 10:38
Posts: 234
I recall reading that Steve J

I recall reading that Steve Job negotiated a crazy low price for 68000 CPUs.

So it could have just as easily been someone at Motorola wanting to had the fact that it was just a plain ol' 68000. I would imagine it would be easier to slip a low price on a "custom" part than a standard part past management.

I agree that no matter the motivation that someone wanted to had the fact that it was a 68000 is best conclusion.

JDW
JDW's picture
Offline
Last seen: 5 months 2 days ago
Joined: Mar 17 2006 - 19:08
Posts: 180
Re: I recall reading that Steve J

...someone wanted to hide the fact that it was a 68000 is best conclusion.

I'm not so sure. Here's why... The board we are talking about is an early Macintosh 128k logic board with a late 1983 build date. I've seen Lisa motherboards with a 1982 date on them that have MC68000 chips in them. Have a look at this Ebay Auction and check out the photos toward the bottom. Flipping the CPU 180 degrees and blowing it up in Photoshop reveals it to be an MC68000L8, GN78303 chip.

So if the theory about Apple wanting to "hide" the chip is true, then why didn't they hide it in the Lisa?!

UPDATE: Here's another photo showing a 1982 manufacturing date for a Lisa that clearly shows the CPU as a 68000, with no attempt by Motorola to obfuscate the CPU silk data.

Offline
Last seen: 3 years 7 months ago
Joined: Mar 9 2005 - 03:00
Posts: 79
Re: Another photo from this site

Here's another photo found on this very site of an SC87839L chip used on an older Mac128 motherboard.

That is actually a 512k mac. I guess Apple held on to those SC chips for a while...

Grant

Offline
Last seen: 3 years 7 months ago
Joined: Mar 9 2005 - 03:00
Posts: 79
Re: Another photo from this site

Here's another photo found on this very site of an SC87839L chip used on an older Mac128 motherboard.

That is actually a 512k mac. I guess Apple held on to those SC chips for a while...

Grant

Opps, its actually a 128k board! I never noticed that! : ) My hacked motherboard is much better than this fake prototype. My 128k mac has SCSI and 4MB ram. ; ) (the motherboard was in a 512k case when I bought it in 1997)

I actually have two 128k motherboards with that processor on it. Henry Courbis had one that he gave me, but he desoldered it first because it looked "cool". Wink

I seriously doubt its a prototype at all. I would say it isn't.

JDW
JDW's picture
Offline
Last seen: 5 months 2 days ago
Joined: Mar 17 2006 - 19:08
Posts: 180
Mac board photo

Whether a 512k board or no, the 128k boards are almost identical. The point here is that it is not clear whether a custom variant of the 68000 existed, more less whether it was used in early Mac boards. The evidence I see is non-conclusive.

Offline
Last seen: 16 years 1 week ago
Joined: Mar 29 2006 - 10:12
Posts: 5
Bit9 Card!

I think this is a 128K board updgraded to 512K. It was common to socket upgrade RAM chips. I know as I cut the 128K ram chips off my 128K mother board and installed RAM sockets myself.

Next to the CPU is a little daughter card. Looks like a Bit9 card that was a common upgrade to decode the extra RAM. As to the CPU, lots of upgrades beyond 512 involved removing the CPU and installing a socket. The daughter boards would plug into the socket and the CPU would be placed in the daughter board. Later on the Killy clip was invented which was a lot easier to use.

IMHO this is no "pre-production" 128K motherboard.

Tony

Log in or register to post comments