February 26, 2022

Callan Multibus 68k CPU board - Memory

The CPU board itself has 256K of ram at addresses from 0 to 0x3ffff. No surprise there. The surprise is that it is not dual ported, i.e. controller boards on the multibus cannot see it. Access to it may be faster than access to multibus ram.

Multibus ram

I have two Micro Memory MM-8086D memory cards. One has 512K of ram, the other has 64K of ram. The "other" card could be upgraded to 512K by replacing the 4116 ram chips with 4164 chips and fiddling with some jumpers. I probably have such chips somewhere, but I am not sure I will ever bother.

Curiously, the 64K card is jumpered to start at address 0 and extend to 0x0fffff. The 512K card immediately follows it, i.e. it is jumpered to start at 0x10000 and to extend to 0x8ffff.

What I intend to do is to switch this around. In fact I intend to only use the 512K card and to change the switches so it starts at 0 and extends to 0x7ffff.

Note that although this card is based at zero on the multibus, the PM68K sees it at another location. Other multibus cards, like the hard drive controller will see it at base 0 when they do DMA. The PM68K sees it at 0x100000.

For some reason, when I install the card I see only 6*64 = 384K, i.e. addresses from 0x100000 to 0x15ffff respond (I expect it to respond up to 0x17ffff).

So, I switched the jumpers back how it has been all along. First however, I used a meter to ensure that the switches were not malfunctioning.

Amazingly enough, I now see 512K of memory! From 0x110000 to 0x18ffff. So, perhaps there is some fault in the card and it was just a workaround to ship it with the other silly card in place with 64K from 0x100000 to 0x10ffff.

Whatever the case, I don't care where the memory is, so long as I can find it and use all of it. I don't even think I will ever use the card that has only 64K.


Feedback? Questions? Drop me a line!

Tom's Computer Info / tom@mmto.org