• 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.

630 solder pads

bbraun

Well-known member
Both LEM (http://lowendmac.com/quadra/lc-630-ram-expansion.html) and AppleFritter (http://www.applefritter.com/node/24415) refer to the solder pads on some versions of the Quadra/Performa/LC 630 logic board as taking another 72pin SIMM slot.

On my board, I see the solder pads, but there are only 64 pads, and in a much shorter configuration than the neighboring 72 pin slot. They don't seem to be setup to accept a 72 pin SIMM socket as these resources suggest.

So, does anyone here know what they are for? Am I misinterpreting these posts or the board?

Thanks

 

Gil

Well-known member
ROM SIMM, most likely. I have had several 5400 board with pads like that, and the diagram shows that they are for a ROM DIMM - except they ended up soldering the ROM chips directly to the board, when they went into production. Boards that actually have the socket are probably either early production units, or development units.

 

bbraun

Well-known member
Thanks, that makes sense. I guess the LEM article must be either incorrect or referring to a different revision of the board.

 

Trash80toHP_Mini

NIGHT STALKER
Nope, they weren't just for development systems!

REALLY Ancient Apple History:

After the "Dirty ROM" debacle, Apple started shipping all (?) systems with soldered ROMs AND a ROM SIMM slot, enabling them to "clean up any further ROM boo-boos," by shipping out new ROM SIMMs. A far less costly proposition than paying anyone an "undisclosed amount" for a software fix.

Apple made the folks at Connectix quite wealthy in order to provide MODE-32 (a patented software fix) to its customers at no cost, thereby avoiding a MASSIVE class action suit by the angry multitude of customers stuck with 24 bit limited systems that had been advertised as, and bought with the expectation of, being 32 bit capable systems.

After a while, Apple stopped populating the SIMM socket pads, which remained a standard design feature for quite some time.

IIRC, of course.

old f**t mode>

ISTR, that among other MLB revisions, DOS Compatible Systems may have had extra pads added to those already present on the MLB for the ROM SIMM. This change from a ROM to a RAM SIMM Slot would have doubled the memory limit for the DOS Compatibles, an intelligent move. However, Apple most likely chose not to implement this change due to increased DRAM densities, which allowed larger RAM SIMMs to become available.

guess mode>

This, of course, was an intentional move by Apple, to cripple mid-range systems to "protect the high end!"
vent.gif


old f**t mode>

[:eek:)] ]'>

 

LCGuy

LC Doctor/Hot Rodder
jt: LC630 DOS boards have solder pads for a ROM SIMM slot? I've never noticed them on mine - I'll have to pull apart the 630 tomorrow and have a look (and even post a picture, I can do these things now with my awesome new high speed internet connection!)

 

Trash80toHP_Mini

NIGHT STALKER
I Seem To Remember, but I could be wrong, therefore I state that I'm only making an educated guess!

BTW, isn't there a scan of the 630 DOS Compatible MoBo over on 'fritter?

NOPE! xx(

 

johnklos

Well-known member
One of the links refers to a post I made. I was wrong - the solder pads were for a 64 pin ROM SIMM, not a 72 pin SIMM. I didn't really look at it much before reading that some Quadra 630-type motherboards have two SIMM sockets, but realized when I pulled out the motherboard to add a SIMM socket.

Needless to say, 128 megs (or 132 including the built-in memory) isn't bad for a Quadra 630-type machine.

 

LCGuy

LC Doctor/Hot Rodder
And now, for some pics!

Here is my LC630 DOS board stripped of all its extra goodies, basically a standard two-slot 630/580 series board, which has neither a ROM SIMM slot, nor solder pads for one:

1000755.jpg

And here's my LC630's original board, which is the earlier revision board and does have the solder pads for the ROM SIMM slot:

1000757.jpg

I'd say that they put the ROM SIMM slot in the first revision board, so that it was there in the event it was needed, but found it unnecessary in the next revision. These are the only two revisions of the 630/580 series board that I'm aware of - if yours is different, please, post pics!

 

Trash80toHP_Mini

NIGHT STALKER
Actually, the pads are still there for the ROM SIMM, they just added the additional pads/lines to turn it into your second RAM SIMM slot! It's all the same bus, just different "Chip/SIMM select" lines, the ROM version was likely only 24 bit, maybe even 16 bit. With just a few rework wires, you could very likely hack your second RAM SIMM Slot . . .

. . . into a completely useless ROM SIMM Slot. :lol:

8-o . . . just thought of something! I'll bet you could do the "same" thing to the ROM SIMM slot! just add the jumper lines for the additional bits from the RAM SIMM! You'd need to use a wire wrap prototyping slot and clip common the legs to a workable length and the additional RAM signals/control lines down to nubbins for the jumper wires (shielded of course!) [;)] ]'>

You could even fab an adapter PCB, assuming you can hijack third/fourth SIMM select lines, to make three/four slots out of . . .

:O }:) :rambo: :eek:)

 

LCGuy

LC Doctor/Hot Rodder
So they used the exact same solder pads, but turned it into a RAM SIMM slot instead? Hmm, interesting. They do some strange things that Apple, much like my MacBook which won't downclock DDR2-800 RAM to DDR2-667 (even though just about any PC will downdlock faster RAM)...all I can say is that I'm not surprised.

As for hacking in further RAM SIMM slots...hmm, that could be a possibility...but for now I'm quite happy leaving the LC630 DOS board in there with its two slots, both filled with 32MB SIMMs, giving me a grand total of 68MB :O

 

johnklos

Well-known member
So they used the exact same solder pads, but turned it into a RAM SIMM slot instead?
The pinouts are very likely not the same, but they're also very likely not tremendously dissimilar. They're definitely not 16 bit - not sure what you were referring to, Trash80toHP_Mini, but if you're talking about address lines, the ROM slot would at very minimum support up to 4 megs, which would require 20 address pins (the ROMs are 32 bits wide, so 2^20 would give 1 million 32 bit words, or 4 megs).

Hmm, interesting. They do some strange things that Apple, much like my MacBook which won't downclock DDR2-800 RAM to DDR2-667 (even though just about any PC will downdlock faster RAM)...all I can say is that I'm not surprised.
That actually is an issue with the memory. I sell Macs as an Apple value added reseller, and I know that most 800 MHz DDR2 memory will run just fine in MacBooks, MacBook Pros, iMacs, and Mac minis, but CERTAIN 800 MHz won't, and it has nothing to do with what kind of Mac it is. I have two 2 gig DIMMs which are waiting for a machine because they (the memory) are 800 MHz only.

As for hacking in further RAM SIMM slots...hmm, that could be a possibility...but for now I'm quite happy leaving the LC630 DOS board in there with its two slots, both filled with 32MB SIMMs, giving me a grand total of 68MB :O
The problem with adding more slots is that you need to generate chip select lines based on the address space being accessed. It's possible that the massively integrated chips on the Quadra 630 type machines could support more than 256 megs, but if nobody has documentation about them and their pinout, it might be difficult or impossible to ever do this.

Just FYI, you have a two SIMM slot motherboard, so you can have two 128 meg SIMMs plus the four on the motherboard, or 260 megs.

 

Trash80toHP_Mini

NIGHT STALKER
So they used the exact same solder pads, but turned it into a RAM SIMM slot instead?
The pinouts are very likely not the same, but they're also very likely not tremendously dissimilar. They're definitely not 16 bit - not sure what you were referring to, Trash80toHP_Mini, but if you're talking about address lines, the ROM slot would at very minimum support up to 4 megs, which would require 20 address pins (the ROMs are 32 bits wide, so 2^20 would give 1 million 32 bit words, or 4 megs).
I was going from my eratically behaving memory. I guess I assumed that the RAM SIMM was 32 bits wide, not 72, and the next steps down would then have been 24 or 16 bits, as I stated . . . incorrectly.

The re-arrangement of the dissimilar pinouts was the reason for choosing a wire wrap socket for the hack. You'd cut to nubbins anything that wasn't identical and solder rework wiring to nubbins for the dissimilar/non-existent lines.

The problem with adding more slots is that you need to generate chip select lines based on the address space being accessed. It's possible that the massively integrated chips on the Quadra 630 type machines could support more than 256 megs, but if nobody has documentation about them and their pinout, it might be difficult or impossible to ever do this.
Exactly, but the 68040 should have plenty of memory select lines available and it's very well documented. The Pro Level Quadras have at least four RAM slots, do they not? It wouldn't be as easy as hacking a IIfx by hijacking the existing parity lines with a few cuts and very short rework wires, but it might be possible. You'd need to be careful about keeping all the lines at similar lengths to avoid timing issues . . .

. . . but it might be a lot of fun to look into! }:)

My IIfx memory hack notions take precedence though. :rambo:

 

Trash80toHP_Mini

NIGHT STALKER
So they used the exact same solder pads, but turned it into a RAM SIMM slot instead?
What he said! [;)] ]'>

They kept the pads in the same place, more or less, that's an assumption based on the similarity of the slot placement, and added the rest of the pads for the SIMM slot, but they'd have altered traces on the MLB to correct for whatever mismatched lines were involved in the upgrade.

 

johnklos

Well-known member
Exactly, but the 68040 should have plenty of memory select lines available and it's very well documented. The Pro Level Quadras have at least four RAM slots, do they not? It wouldn't be as easy as hacking a IIfx by hijacking the existing parity lines with a few cuts and very short rework wires, but it might be possible. You'd need to be careful about keeping all the lines at similar lengths to avoid timing issues . . .. . . but it might be a lot of fun to look into! }:)
The issue is that if the address decoding is complete, which one would guess it would be on a last-generation Quadra, then the RAM cycle generation wouldn't happen when addresses are accessed outside of the normal RAM area. So while one could simply put a 74HTC138 onto A31, A30, and A29 and assume that everything else mapped in the Quadra is in the 0-512 meg range and activate a bank of memory in the 512-1024 meg range, you'd also have to have the circuitry to strobe in the full address.

Now if someone were to make some huge straight memory chips (meaning all of the address lines are run out to pins and the data lines are separate), this would be much more doable. Wiring memory to, say, an 8 bit processor was something a 12 year old could do (I did!)

 
Top