IIfx ROM SIMM in a IIsi . . .

[dougg3]

Whew! Thanks for restoring my sanity! Glad you got it working.

I don't know the IIsi RAM chip pinout, but there is probably a single line you can disconnect from the chip and then tie that pin high at the chip to disable it completely (like the RAS pin on a DRAM chip's data sheet I'm looking at). What is the IIsi RAM chip part number?

Will the SIMM RAM take the soldered in RAM's place, or is it set up to always appear above the first 1 MB of RAM?

[Trash80toHP_Mini]

NoPro! It's not like this kinda crap hasn't happened before . . . even Macs got gremlins!

I've got to tear the rig down to get to the part numbers and I'm too tired to do it tonight!

Will the SIMM RAM take the soldered in RAM's place, or is it set up to always appear above the first 1 MB of RAM?

Could you re-phrase that please? That does not compute . . . zzzzzzzzzzzzzzzzzzzzzzzzzz

I'm hoping to do a complete replacement of both Banks of RAM with 72 pin SIMMs that are FAR more readily available to max the sucker out.
Picture a reverse SIMM Saver kinda deal.

[dougg3]

Sorry, let me try that again

What I mean is, even if the soldered RAM is not responding, will the IIsi try to assume it's still there? Like for instance are the SIMM sockets hardwired to only map from addresses 1 MB and higher? Or does the MMU transparently handle all of that? (I honestly don't know anything about how the RAM works in these computers...this question is mainly for my curiosity)

[Trash80toHP_Mini]

From the DevNote, it sounds as if the IIsi polls the sense lines for monitor type, and likely the PRAM for B/W or grayscale/color bit depth, in determining how much of Bank A needs to be buffered for Video Memory.

Assumptions/Ramifications

If - Apple went to the trouble of tweaking the IIci ROM to limit Bank A to 1MB in the IIsi's Memory Mapping
Then - that limitation should disappear upon substitution of a ROM with a less limited memory mapping scheme.

If - hacking the IIsi's soldered in memory allotment in Bank A entails giving up the use of the Mobo's Vampire Video
Given - the IIsi's I/O expansion limitations imposed by the single PDS slot/PDS Adapter Or single slot NuBus Adapter
Then - Apple likely assumed that no IIsi user would bother to hack the IIsi's soldered in memory allotment Bank A

If - hacking the IIsi's soldered in memory allotment in Bank A does not affect the buffering of the first MB of a hacked Bank A
Given - the IIsi's Memory Map Limitation of 1MB for for Bank A
Then- that limitation should disappear upon substitution of a ROM with a less limited memory mapping scheme.

Assuming - a Hackintosh was a practical option at that point, the IIsi would be immediately dismissed as a viable candidate

WAG - No PROBLEM!!!!!!!!!!!!

[dougg3]

Thanks for the info! I guess we'll just have to see what happens!

You guys are gonna love this one:

I just burned a Quadra 650 ROM dump onto my SIMM for fun. It boots the IIci! WITH the IIci's startup chime!

About this Macintosh still identifies it as a IIci. The only thing that's weird is I think it resets the SCSI bus or something, because my hard drive makes its initialization ticks twice (normally it does a few ticks while it's warming up). The blinking question mark comes up before the ticking is finished, but then a couple seconds later the hard drive finishes initializing, the happy mac appears, and the IIci boots right up.

Amazing, isn't it? I mean, the Quadra is a 68040 Mac. I assumed they would have just thrown out all the old Mac startup stuff, but it still works with the IIci. Does anyone else smell the need for us to create a Macintosh ROM compatibility matrix project?

[Trash80toHP_Mini]

[Trash80toHP_Mini]

IIRC, some of the extra ROM storage includes what we've assumed to be repeated copies of the same ROM.
If so, could some be copies of earlier ROMs with the proper startup sound for, in this case, the IIci?

Is there anything in the Inside Macintosh startup procedure documentation that indicates that a CPU might look for a machine ID indication (bus multiplier type resistor array?) or poll the system for Mac specific ASICs before banging on one particular ROM image's door?
Dunno, just spitballing!

[dougg3]

Well, the repeated copies of the same ROM don't repeat inside the size of the ROM chips -- they are repetitions of the entire ROM contents. They do that because some of the address pins are left unconnected, depending on how big the installed ROM chips are.

In this case, it means that they definitely left earlier IIci startup code somewhere in the ROM. In fact, they went so far as to leave in the old *synthesized* startup sound code for II series Macs (along with the sampled startup chime that the 650 uses).

http://mess.redump.net/mess:driver_info ... ical_notes

The above list seems to show on some II series Macs, a few of the VIA data register bits will identify the model.

[Trash80toHP_Mini]

Interesting question:

IF . . . the IIfx requires SIMMs of 1MB minimum capacity . . .
IF . . . the IIfx RAM banks need to be of equal size . . .
Then . a IIfx needs a minimum of four 1MB Simms installed in both banks to boot . . .

Might a IIsi with 1MB soldered into Bank A, actually boot with 4 256Kb "Keychain" SIMMS in Bank B with a IIfx SIMM installed?

Maybe it's time to rummage around for some Keychain fobs!

[Trash80toHP_Mini]

The saga continues . . . quotes pulled from the Trading Post:

I can't believe that I can't locate any of these pestiferous POS SIMMs in my MacPartsHorde . . . but I can't!
I'd appreciate it it someone would sell me a set of these Paleolithic keychain fobs for a hack project test.

TIA,
jt

I have a matched set of 4 if you still need them. Pulled from a Mac Plus, just a few weeks ago, so they definitely work.

What are you building that needs 256k SIMMs?

AichEss has already made the generous offer of mailing a set out for me tomorrow, but thanks!

I'm re-testing the IIfx ROM SIMM in a IIsi . . . Hypothesis, in hopes that the minimum requirement of 1MB SIMMS to work in the IIfx, would be that 1MB was the smallest custom SIMM Apple ever made available for that $10,000 beast.

Even in 1990, a 4MB SIMM as the lower limit probably ought to have been the choice of any other Unix capable Machine Mfr for anything they had the stones to declare as "wicked fast!"

The IIsi has 1MB soldered in as Bank A, so before hacking 4MB or more into Bank A . . .
. . . I might as well see if installing 1MB in Bank B will do the trick for equalizing the memory banks!

If the IIsi boots from the IIfx ROM SIMM in a 2MB config, it'll save me a lot of bother . . .
. . . my original, bought new, Rocket 33 that's slated to go in there already has 32MB on board right now . . .
. . . and it'll only be using the IIsi to copy its ROM into RAM, as an I/O processor, Video Card Cage & networking port.

Forgot to mention, with any luck, as an ATX PSU upgrade enclosure for itself . . .
. . . and two or three PDS Cards slung under the MoBo!

To finish up the "thread" on a Good, Bad & Ugly note:

The good: thanks ever so much for the 256k SIMMs, AichEss. Safely arrived & tested.
The Bad: I still get the 5(?) note death chime with my stock IIfx ROM and the Hacked ROM SIMM in the IIsi.
The Ugly: gotta start looking into maxing out Banks A & B with a 72 pin SIMM Hack again . . .

@*&/*0\&^*&%&^$%!!!!!!!!!!!!!!!!

Thanks again, aichess! Time to lock that thread . . .

[dougg3]

Hmmm....this is gonna be tough to figure out. The death chime doesn't necessarily indicate a RAM problem, so something totally different could be causing it too.