[Coco] [Color Computer] Re: POKE 65495,0

Mike Pepe lamune at doki-doki.net
Sun Feb 4 19:07:42 EST 2007


jdaggett at gate.net wrote:
> On 3 Feb 2007 at 22:37, James Diffendaffer wrote:
> 
> You had good dram then. Because your dram didn't loose data does not preclude 
> that all dram will work after 4.5  seconds of no refresh . 
> 
> james
> 

The thought occurs that this question: "Does the SAM really not refresh 
DRAM in 'Fast' mode?" is actually pretty easy to test without resorting 
to logic analyzers and such.

In a machine language program:

Disable interrupts
Put the machine into all RAM mode.
Move your code somewhere, say $0000.
Write a defined pattern into the rest of the memory map.
Run a tight loop for some time.
Switch back to Slow so you can see a display :)
(Assuming you survive) checksum the RAM and see if any of it changed.
You can then repeat and increase the delay until something gives.

It would sure be interesting to run this program among a wide gamut of 
CoCo 1 & 2 machines to see what happens.

The loop would conceivably not be "big" enough to refresh all 256 rows 
by the nature of it simply running.

-Mike



More information about the Coco mailing list