Commander X16

could the extra ram chips be moved to an expansion board?

On 7/30/2022 at 5:39 AM, Fabio said:

could the extra ram chips be moved to an expansion board?

In theory, but it would require changes to the expansion slots, and it would make the extra RAM more expensive, rather than less. 

At this point, if they want people to actually be able to use 2MB of RAM, the cheapest alternative is to leave the sockets in place on the board, but not populate them. 

 

Since I don't consider the Panasonic MSX Turbo-R Machines to be 8-bit computers (the ASCII R800 featured 16-bit integer instruction arguments, registers, and at least one 16-bit data bus), the 8-bit computers with the most standard RAM would be the final production year Tandy CoCo 3 and Fujitsu FM-77, with 512-640K system RAM each (plus 128-384K separate VideoRAM in the case of the FM-77 AV40 Level 3).  That said, the Apple IIc+ and IIe Anniversary Edition, Sharp X-1 Turbo Z and X, and NEC PC8816/18 have O/S memory maps capable of addressing up to 2MB, and Atari 8-bit computers equipped with the FREDDIE MMU/DMA can address up to 4MB.

Back in the day, two full megabytes of High RAM+124K Video RAM, 4K PCM Buffer, and 64K Low RAM would together be seen as overkill until Ca. 1990-91, with only the Apple IIGS and Amiga 750 and then 1200 possessing similar memory amounts among contemporary non-PC Clone competition.*

 

*I do not consider the Macintosh III and Quadra, Sharp X68000, High End Amigas, or Acorn Archimedes A4000/A6000 series to be anything close to hypothetical market segment competitors.

If it counts, the later Apple IIgs models came with 1MB installed... actually it was 1MB + 128K.

True, however, the 65816 features 16-bit integer instruction arguments and register sizes.

Make a DDR2 etc DRAM interface using an FPGA which have a DIL socket under it. Behaves like a RAM chip but way more availability.

The important thing is to make sure all relevant (address) signals are available in the socket. One option is to place another smaller DIL socket inside the larger than such that ordinary RAM chips use the standard pins but a special board may enable even more (D)RAM by using extra high address signals from a center connector.

 

On 10/19/2022 at 3:43 PM, neutrino said:

Make a DDR2 etc DRAM interface using an FPGA which have a DIL socket under it. Behaves like a RAM chip but way more availability.



The important thing is to make sure all relevant (address) signals are available in the socket. One option is to place another smaller DIL socket inside the larger than such that ordinary RAM chips use the standard pins but a special board may enable even more (D)RAM by using extra high address signals from a center connector.



 

Once again, this isn't going to happen. There are no more planned changes to the hardware, and if you have an idea for a hardware product, you should be prepared to design and build it yourself. 

An REU-like memory expander is possible, but there is no way to fully bank out memory on the mainboard. That design feature was intentional, as David wanted the memory banking to be as simple as possible. So if you have an idea for your own memory expansion module, you'll want to figure out how to build it within the constraints of the expansion ports. As in:

1. 
   Each port gets 32 bytes of I/O space. If I recall, there are actually give 32-byte blocks, so your card will need to be able to select from one of those blocks.
   


2. 
   Your card can read from and write to main memory, but it must follow the DMA process outlined by @Lorin Millsap here: 
   


3. 
   There are four expansion slots, so your card must get along with cards plugged into the other slots. IE: it needs to stick to its 32 bytes of address space and not try to tie up all the I/O space for its own use. If you want to store stuff in expansion RAM, it needs to go through the 32 bytes in that card's slot. How you do that is up to you, but I suggest you read up on how the Commodore 1750 REU works as an example.