Am 31. Januar 2019 09:26:57 schrieb vossi <vossi@ceffy.de>: > Hi, > at testing my new CBM2-v9958 graphics card I had a problem to write to the > VDP-VRAM. > After some non-successful testing with longer delays between reg and vram > access, I started my LA and found an interesting fact: > The 6509 performs a short read access before it writes with STA(ZP),Y from > bank1 to bank15 (screenshot)! What do you mean by "short" read? Without Phi2 in this picture it is difficult to judge. And yes, the 6502/6509 have bogus reads while doing address calculations, most of the time when crossing page boundaries during indexing. This is known behaviour. André > I tried again with code in bank15 $0400 using STA absolute,X and the 6509 > only writes correctly. > I also measured a read from bank1 to 15 and thats only a plain read. > > > My 6509R4 and the 6509R7 have both this "feature" ;) > Is this a know bug of the 6509? > > > PS: After setting up the v9958 VRAM-address I write to the VRAM via port #0, > but the VDP saw the read access and increases the address counter so that > the write-byte land one byte later in the VRAM and I had a displacement by > one char on the screen. > > > BTW: The v9958 has 128k dedicated VRAM and does 256x212 with 256 colors, > 512x212 with 16 colors, 512x424 interlaced and 256x212 with 19k colors. It > also supports 32 sprites, sprites collision detection, hardware commands > like higspeed-move, line, pixel, search and H/V-softs-crolling. It has > direct RGB output with PAL or NTSC software selectable ans screen adjusta. > > > <http://cbm-hackers.2304266.n4.nabble.com/file/t374762/6509indirect.jpg> > > > > > > > ----- > greets > Christian > -- > Sent from: http://cbm-hackers.2304266.n4.nabble.com/Received on 2019-02-01 09:00:03
Archive generated by hypermail 2.2.0.
