Re: IIgs problem...

[Rubywand <rubywand@swbell.net>]

Labelas Enoreth writes ...
> 
> More on the IIgs memory problem and error $0682...
> 
> Moved the chips around again, now the card is laid out so that the top row
> is 16 150ns chips (8 soldered on) and the bottom is 16 120ns chips.
> I just don't have another 16 150ns chips to put there.
> Anyway, here's the thing. Startup without card in...computer works fine.
> Startup with card with no jumpers (only first 8 chips in upper bank used,
> right?)...works fine.
> Startup with upper jumper on (upper bank used, right?)...works fine
> Startup with bottom jumper on (bottom bank used, right?)...works fine
> Put both jumpers on, and wham. There's the errors and crashing.
> So, I'm guessing that it's not the chips themselves, but an interaction
> between 150ns and 120ns RAM? Or am I missing something?
> 
> Here's the card, sort of...in case anyone wants a bad yet gratuitous ASCII
> art...
> 
> 15-15-15-15-15-15-15-15 15-15-15-15-15-15-15-15        ..
> 12-12-12-12-12-12-12-12-12-12-12-12-12-12-12-12        ..


     Part of the confusion in messing with the 1MB card is that you have four
possible jummper settings but only three legal memory configurations. According to
Apple's manual for the board ...


no Jumper= 256k (upper left bank of chips)

Jumper on bottom= 512k (upper left and right banks) adds upper right bank

both Jumpers= 1MB (all banks) top Jumper adds lower left and right bank


     As you can see, adding the top jumper activates the lower two banks. If
_just_ the top jumper is added, you are enabling access to the lower two banks,
which, altogether, gives you three 256k banks the GS can access.

     The problem with this setup is that you are leaving a gap in memory address
space because the upper right bank is not enabled. So, if this setup works, my
guess is that the GS detects no memory where the second bank should be and stops
looking-- i.e. you are really still testing only the first 256k bank. Part of the
memory in the lower two banks may be bad but it would not matter.

    _If_ the above guess is correct, you would not get to test the lower two banks
via the method you are using unless both jumpers are in place.

     So, it seems pretty clear that you have a bad RAM IC in the lower bank-- or,
possibly, one of the interface IC's is screwing up.



Rubywand