On 04/08/2013 11:12 PM, email@example.com wrote: > > On 2013-04-08, at 22:53, Gerrit Heitsch wrote: > >>>>>>> If you tristate the cpu, you still have half the cycle left for the >>>>>>> write. Not much more difficult. >>>>>> >>>>>> You also need to monitor the BA (RDY) line, otherwise you'll run into trouble when the VIC does a badline and uses the complete cycle. >>>>> >>>>> But - generally - if this is done on power-up, then it could possibly be done before VIC gets initialised (I assume - maybe wrong now - that it powers up with "screen disabled" state)? >>>> >>>> It will still do its own read cycles even if they're dummy cycles, including the refresh cycles. You might be able to get rid of the badlines though. >>> >>> That's what I meant. Clearing the "screen disable" bit (bit 4 at SCROLY register) is the soft way to get rid of bad lines, used in many timing critical operations. If this bit is (as I expect) cleared on power-up then at least the bad lines are not interfering and there would be no need to monitor the extra line. >> >> Problem is, only a power down will get VIC into that state. Just a normal RESET will not. > > That's exactly the case here. We (out of list) are discussing the options how to make my long-wished ultimate ROM replacement, which is to be triggered only on power-up. > >> So how do you make sure that the extra circuit only accesses memory after a power cycle (that is long enough to really reset VIC) but not after a user pushes the RESET button? > > This is not in the use cases ;-) Then you need a little extra circuit that makes it impossible for it to happen, otherwise it will. You know, Murphy. Should be easy enough... When RESET is let go and the system starts up, a flipflop flips and prevents access until it's reset by a power cycle. As for finding out what VIC does at power up... Keep the system in RESET at power on and take a look at the BA line. If the display is indeed disabled, nothing should happen on it. Gerrit Message was sent through the cbm-hackers mailing listReceived on 2013-04-08 22:01:00
Archive generated by hypermail 2.2.0.