Re: VIC-II DRAM refresh

From: Gerrit Heitsch <gerrit_at_laosinh.s.bawue.de>
Date: Fri, 14 Oct 2016 09:34:17 +0200
Message-ID: <5af5e232-3138-c352-d4c7-df962831bf8c@laosinh.s.bawue.de>
On 10/14/2016 09:11 AM, smf wrote:
>
> On 13/10/2016 20:07, Marko Mäkelä wrote:
>> Remember that the C64 KERNAL blanks the screen during tape access. I
>> think that it is possible that the memory accesses of the KERNAL tape
>> routines would not guarantee sufficient DRAM refresh.
>
> The blank screen not fetching character pointers might be the reason,
> although VIC2 still has it's bitmap fetches that it could use.

Yes, but disabling the display disables that too. And even if VIC was 
still doing bitmap fetches, the combined time for lower border, vertical 
retrace and upper border is longer than 2ms. That means you have to 
supply refresh cycles during that time. The 5 refresh cycles per 
scanline method is about the simplest way to solve all problems.


> I guess as their first dram system, they just picked what they thought
> was the easiest and safest method. It turns out that the VIC2 dram
> access was out of spec though, they sued micron for supplying faulty ram
> and it turned out the dram was faulty and the VIC2 was doing it wrong too.

Hm? They did use Micron DRAM, the dreaded MT4264-xx, in the C64 though. 
That must have been a very early revision of the VIC2.

I remember that differently, they tried to make their own RAM, bought a 
design from Micron and couldn't get it to work. Micron then admitted 
that they couldn't get it to work either and that's when Commodore sued 
them.

  Gerrit




       Message was sent through the cbm-hackers mailing list
Received on 2016-10-14 08:01:40

Archive generated by hypermail 2.2.0.