Re: VSP-Error once more

From: silverdr_at_wfmh.org.pl
Date: Wed, 2 Jun 2021 19:28:43 +0200
Message-Id: <06395549-1DD9-4F86-933A-838653B36686_at_wfmh.org.pl>
> On 2021-06-02, at 17:33, Gerrit Heitsch <gerrit_at_laosinh.s.bawue.de> wrote:
> 
>>> Do we know / have any data about what board/VIV-II combination is the most
>>> prone to the said error? So far I was able to reliably reproduce it (with
>>> VSP-Lab) using 250425 board and 6567R56A VIC-II (which is in fact quite
>>> different from all other variants in respect to timings) but I need to test
>>> it on PAL too. While I remember finding hardware combination exposing the
>>> issue some good time ago but now I seem to be unable to find it again.
>> It depends very much on the individual components on the board (PLA. RAMs,
>> VICII, probably more).
> 
> The PLA should not be involved since /RAS doesn't go through the PLA.

Right! /CAS does. OTOH it can change between power cycles so who knows what factors affect it.

> From how I understand how it works (address lines changing too close to /RAS going LOW), it would be a combination of VIC and RAMs and maybe clock jitter. Plus, of course, temperature.

True. I am not looking for an exact recipe but am trying to increase my probability of finding a machine / combination of components which would expose the error at least somewhat repeatably.

-- 
SD! 
Received on 2021-06-02 20:00:49

Archive generated by hypermail 2.3.0.