Re: Different ROM 3 ROM chips ?????

[spector@vax2.concordia.ca (Mitchell Spector)]

In article yeo@wattle.qut.edu.au (Willie Yeo) writes...
>
>	The other day, fixing my friend's ROM 3 motherboard with wierd
>problems (certain programs worked on one but not the other ROM 3 machine),
>I discovered that both of his ROM 3 machines have different ROMs !
> 
>	Looking at a closer look, and comparing with my ROM 3 machine
>back home, I found out that the ROM 3 machine in question, had a 32 pin
>ROM FF-FE (located on UJ9, next to the 65816), part 341-0749 and sitted
>next to it, is also a different ROM FC-FD, a 28 pin ROM, part 341-0728.

	These specific ROM part numbers were never publically released.
341-0749 and 341-0728 had some slight bugs that were caught just before
(or days after) the ROM 3 went into production, yet again, _none_ should
have made it into the comsumer market. What your friend has is either a 
ROM 3 prototype that slipped out of Apple, or a dealer-demo that had the
prototype ROMs installed. Where did your friend purchase this ROM 3?

>	Both my friend's 'working ROM 3' and my machine had 341-0748
>for ROM FF-FE and 341-0737 for ROM FC-FD, both 28 pin ROMs.

	341-0737 and 341-0748 are the standard (newer) ROM part numbers in 
the ROM 3. The only difference between the two ROM 3's is in the 256K of
firmware, but no one (expect in _extreme_ rare circumstances, such as your
friend's ROM 3) has the prototype ROMs.

>	Anyone have a different part number ?

	Nope, I've only heard of one other case where someone had the 
prototype ROMs. David Empson has disassembled these unreleased ROMs and
listed the differences, I'm sure he'll follow up shortly with a review
of what they are. I recall the most serious bug was in QuickDraw II or
the print manager.

	FYI there is a third type of ROM 3, though it has nothing to do with
the firmware. It has to do with the vaule found at E1:017D:xx in memory at
cold-boot. The vaule _should_ be '00' but on my ROM 3 (and a hundred or so
others) the vaule is 'FF'. This bug causes the programmer CDAs be invoked 
whenever the machine is powered on, even the bRAM settings don't help here.
No one has quite figured out the source of the problem, but a guess is the
type of DRAMs used in the 'Standard RAM' (128K) bank, just next to DOC RAM.

Solutions? If you have the prototype ROMs (rare) swap them with a copy
of the current ROMs. Alltech might have the parts or perhaps you can burn
a copy if none are available. It's important that it be done...

CDA-bug: Install a small INIT in your System folder that temporarily purges
them from memory. At the same time it changes the address at E1 to '017D:00' 
so the first warm-boot keeps them from returning (until you power down the GS).
The extra CDAs are really just a minor annoyance and eats only a couple of 
bytes. This, like the other ROM 3, is rare. I found a batch of them at 
Apple Headquarters in Montreal though, possible they only made it to this
part of Canada. You could do a motherboard swap too, if your luckly enough
to find one (the Apple dealer out here tossed a dozen or so in the dumster
a few weeks ago, wish I had been there. :-(

>Willie Yeo (Kingfisher)    | Apple IIGS ROM 3, 8Mb OctoRAM, RAMFast D SCSI,
>Brisbane Apple Users Group | ********** InterNet : yeo@qut.edu.au **********

Mitchell Spector
m_spect@vega.concordia.ca / spector@vax2.concordia.ca