[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [N8VEM-S100:1814] S-100 68K CPU board V3
I guess the real question to me is what are we doing? Are we building / designing hardware because we can or is there a higher purpose. I am basically a software guy that likes to dabble in hardware and I don't see a lot of software being created for all these boards (maybe the software folks are being really quiet). Like I tell my hardware compatriots at work - you are just building an expensive pile of sand (silicon) without the software to be able to use it.
I think SBC is probably the better way to go as the S100 bus is becoming a more expensive I/O bus for these more powerful processors and in the end limits their speed and adds more complexity. Maybe we should explore a better bus structure for these more powerful processors?
Just my thoughts - would be nice to hear what others think. No use building a jet engine to place in a Model T Ford ?
On Wednesday, August 14, 2013 10:27:01 AM UTC-5, lynchaj wrote:
On the S-100 80386 CPU board there is a special connector “over the top” for memory boards with full 32 bit addressing and 32 bit data width. The S-100 bus serves for accessing base memory and IO mostly since it is limited to 24 bit addressing and 16 bit data width. The S-100 80386 CPU board can work with just bus memory but is limited to 16MB of 16 bit wide memory and is going to be slower than with the special memory board.
Assuming we move forward with a more advanced 68K series processor (68020, 68030, 68040, etc) on the S-100 bus I recommend a similar approach if not the same “over the top” memory boards to make them more economical and provide a useful amount of RAM to more powerful CPUs. I was considering using the PAK68 approach for a 68020 or 68030 CPU socket mezzanine board for the S-100 68K board but in retrospect it would probably be better to design a new board instead.
Thanks and have a nice day!
Not sure it makes sense on an S-100 board as the address space is limited to 16 MB. It will probably be an SBC with a couple hundred MB of memory - when we start getting in the GB range it is not quite as practical and the old chips are limited to 4GB anyway without a lot of additional memory decode tricks. I can probably get old version of NetBSD flavor running on even a 68K with some tricks - at least I will be able to gauge the complexity.