Beige G3 Musings

9 posts / 0 new
Last post
alk
alk's picture
Offline
Last seen: 2 years 10 months ago
Joined: Dec 20 2003 - 10:38
Posts: 369
Beige G3 Musings

Well, this is interesting. Has anyone ever taken a really close look at the beige G3 mobo?

I've got a spare I was using for my Wallstreet VRAM hack, and I noticed that there's a jumper near PCI slot C1 labeled "Mac Supply" and "PS/2 Supply". PS/2 Supply is set on both of my boards... Shades of CHRP!

Something else that's interesting is that the perch slot connector is the same connector used for the x100 PowerMac PDS. In fact, a 6100 right angle adapter can be used to tilt the personality card 90 degrees (so that the board "hangs" down toward the bottom of the case). Of course, the personality card is too large to actually fit in that configuration, but it might be good info for folks doing case hacks on with the beige G3 mobo.

Another thing is that there's aparently a space for a PC-style floppy drive connector (but not when the Mac style floppy connector is installed - the two overlap each other).

But what's MOST interesting is the blank cache DIMM slot! Just north of the CPU is an empty pad for a standard PCI Power Mac cache DIMM! I wonder... If I were to install such a connector and DIMM, would it appear to the Mac as L3 cache like on other PCI PowerMacs when you don't remove the on-board L2 cache before installing a G3 upgrade?

Hmmm.... I smell a hack coming on... Wink

Peace,
Drew

JetStar's picture
Offline
Last seen: 18 years 6 months ago
Joined: Dec 20 2003 - 10:38
Posts: 155
Hate to bust your bubble, but

Hate to bust your bubble, but all of those findings have been documented before. I'll give a quick run-down.

"These are the remains of the Gossamer project. Here is an exerpt from xlr8yourmac.com, since I'm too lasy and such a lousy writer/speller to type it out my self:
The motherboard of the beige G3 is known as "Gossamer". It was originally intended to be the basis of inexpensive white-box Macintosh clones. Based on the Motorola MPC106 SDRAM chipset (XPC/MPC-106 Memory/PCI controller-Mike), it would have used ZIF PowerPC 603e or G3 processors. However, as everyone knows, when Steve Jobs returned to Apple, his first victims were the Macintosh clones.

"At the time, the next generation Apple logic board was known as "PEX" or "Power Express" or the Power Macintosh 9700. (see http://www.applefritter.com/prototypes/pex/) As it would turn out, what was supposed to be a low end clone board, was, in fact, a great deal faster than PEX by the virtue of its SDRAM support and faster bus speeds. It was cancelled and Apple went ahead using the "Gossamer" logic board as the basis of the original G3 line."

The "ram" slot holes you see were for the 603e cashe slot. A g3 wouldn't need it, obviously. There are some reports of a pc floppy drive working, but you have to unsolder the mac floppy port in order to do so. The only thing I didn't get was the *100 slot thing. I'm guessing it's just coincidence.

alk
alk's picture
Offline
Last seen: 2 years 10 months ago
Joined: Dec 20 2003 - 10:38
Posts: 369
Don't worry, you aren't.

Hate to bust your bubble, but all of those findings have been documented before. I'll give a quick run-down.

Oh, you aren't bursting my bubble. "Has anyone ever taken a really close look at the beige G3 mobo?" was a rhetorical question. I didn't really think I was the first person on Earth to see these things. I was really wondering why no one had DONE anything with these bits.

I know the history of Gossamer quite well, in fact.


The "ram" slot holes you see were for the 603e cashe slot. A g3 wouldn't need it, obviously. There are some reports of a pc floppy drive working, but you have to unsolder the mac floppy port in order to do so. The only thing I didn't get was the *100 slot thing. I'm guessing it's just coincidence.

Well, I didn't call them "ram" slot holes. I clearly said that it was a blank spot for a PCI PowerMac cache DIMM - what you are calling 603e cashe. I also mentioned that the Mac floppy and PC floppy solder pads are too close and preclude the use of both floppy types simultaneously. And the x100 thing probably isn't coincidence - it's probably Apple saving money and reusing a connector design that had been in wide use by the company before (and available from 3rd parties - the connector is made by FoxConn). I was hinting at all those folks who want to use beige G3 mobos in hacks because they are so small. Usually they sacrifice the personality card to the hack because it sticks up a good 5" (and lose audio as the trade-off). Well, with a 6100 right angle PDS adapter, the perch cards still can be utilized in a low-slung case - it just has to be a couple inches longer.

But that cache DIMM slot. If one were added to the board and 1MB of cache installed, do you think the Mac would see it as 1MB of external L3 cache?

As an aside, I also wonder if the board can take a 603e ZIF from one of the Umax clones. And, by corrollary, can a Umax clone take a G3 ZIF? Those are questions I cannot investigate as my only clone is a PowerTower Pro.

Peace,
Drew

Mackie's picture
Offline
Last seen: 17 years 9 months ago
Joined: Dec 19 2003 - 17:34
Posts: 92
Wallstreet VRAM hack??? More

Wallstreet VRAM hack??? More info plz, this could get interesting ::)

JetStar's picture
Offline
Last seen: 18 years 6 months ago
Joined: Dec 20 2003 - 10:38
Posts: 155
Heh. Sorry if I didn't read y

Heh. Sorry if I didn't read your post correctly last time. I sit corrected. Smile But that cache DIMM slot. If one were added to the board and 1MB of cache installed, do you think the Mac would see it as 1MB of external L3 cache? I have no idea if it would actualy *work* since it *was* desined for a 603e, but I'm guessing that it woudln't show up as a L3 cashe, as a best case senario, it would add to the l2 cashe that is on the die. The worst probably would be that it simply wouldn't work, since nobody knows if the support chips and such were left on or not incuded in the final production board. The ROM would have to work with it too, not to mention the 603. At best, it's probably just a mixed bag; if you have a spair board, the time to solder on a connecter, and the 603 from your clone, plug away.

alk
alk's picture
Offline
Last seen: 2 years 10 months ago
Joined: Dec 20 2003 - 10:38
Posts: 369
Re: Wallstreet VRAM hack??? More

Wallstreet VRAM hack??? More info plz, this could get interesting ::)

Hmm. Please don't hijack this thread for the Wallstreet VRAM topic.

I originally posted this hack to AF 1.5 back in JANUARY. There was zero interest (most people are still skittish about hacking G3 PowerBooks given their relative expense).

References:
http://www.applefritter.com/cgi-bin/YaBB/YaBB.pl?board=hacks;action=display;num=107389006
http://www.alksoft.com/projects/wallstreet_vram.html

Dan K and I went in on an order to Digi-Key (Ok, Dan made the order and asked if I wanted anything). I'm getting a pair of SGRAMM SO-DIMM slots, but don't hold your breath waiting for me to complete the hack.

Peace,
Drew

Offline
Last seen: 8 years 7 months ago
Joined: Dec 20 2003 - 10:38
Posts: 354
I had also noticed the 6100 P

I had also noticed the 6100 PDS/Perch slot similarity. I didn't pursue it because of the direction of the adapter; it would make the overall length longer (out over the PCI slots). My other concern would be to find out if the adapter has electronics or missing connections due to the different purpose. Worth finding out though, because that stupid $%()&* Perch card p$*(%&sses me off. That thing sticking up so high ruins a lot of nice hack ideas.

alk
alk's picture
Offline
Last seen: 2 years 10 months ago
Joined: Dec 20 2003 - 10:38
Posts: 369
Yeah. I'd give it a try, but

Yeah. I'd give it a try, but I've only got one power supply and two boards. The working board is hosting my AppleShare server full-time.

But from inspection, it looks like most, but not all the connections are made for the PDS adapter. ... Apple would have no way of knowing what cards third parties were making for the 6100 and which pins were being used, so they would have had to connect every pin on the right angle adapter to a pin in the 6100 PDS. To do otherwise would have been inviting incompatibilities. Well, they'd have to connect all the active signals in the 6100 PDS, anyway. Who knows - the perch slot may use a different physical pinout that wasn't supported on the 6100 PDS... So I guess I'd have to take a look at the edge of my Wings card to figure out if the same pins are/aren't being used. Oh well, that was worth a short thought.
Sad

For the intrepid hacker, though, it should be pretty easy to connect up the missing traces... The board is completely passive (no logic on it at all) and has great big through-hole solder points for the angled PDS. It should be easy to wire these directly to the board-edge connector...

Peace,
Drew

jt
Offline
Last seen: 17 years 1 month ago
Joined: Dec 20 2003 - 10:38
Posts: 447
Buz the connections . . .

. . . on the MLB AND the 6100 adapter FIRST! I'd be a lot more worried about alternating ground lines and voltage pins being tied together in mismatches than missing signals. Some of those "missing" connection might be made in a ground or voltage plane. IIRC the DevNotes have pinouts for the 6100's PDS, but not for the identical 7100 and 8100's?

* or was that the other way around? *

That'd shortcut your testing requirements quite a bit!

jt Wink

Log in or register to post comments