Open thepulserifle opened 3 years ago
@sorgelig This was identified in the old Arcade-SnapJack core but never reported, but we've now confirmed it happens in the new merged core as well so I asked @thepulserifle to report it here with evidence
This is something that rarely happens for me but I was able to find the issue.
Gating the Char & Color RAMs access here with clk_en_4mhz_i
causes much too short data output.
https://github.com/MiSTer-devel/Arcade-LadyBug_MiSTer/blob/962741e878b191f8e94436417940395b4e537a64/rtl/ladybug_char.vhd#L511
Please try the below core and see if it is fixed: Arcade-LadyBug_20240602.zip
I found that the Wait signal for the CPU is sometimes cleared too early.
The b1_ff
signal is supposed to delay the Wait clear when a new scroll value is loaded in at the falling edge of hx_ctrl_q
.
https://github.com/MiSTer-devel/Arcade-LadyBug_MiSTer/blob/962741e878b191f8e94436417940395b4e537a64/rtl/ladybug_char.vhd#L314
b1_ff
should go high at the falling edge of h_ctrl_s(3)
which is done here:
https://github.com/MiSTer-devel/Arcade-LadyBug_MiSTer/blob/962741e878b191f8e94436417940395b4e537a64/rtl/ladybug_char.vhd#L376-L379
The problem is that h_ctrl_n_rise_s
is wrong and b1_ff
goes high at the rising edge of h_ctrl_s(3)
.
https://github.com/MiSTer-devel/Arcade-LadyBug_MiSTer/blob/962741e878b191f8e94436417940395b4e537a64/rtl/ladybug_char.vhd#L265
It should be
h_ctrl_n_rise_s <= h_ctrl_s and not h_ctrl_d_out_s;
Here is a new core with the fix: Arcade-LadyBug_20240606.zip
https://www.youtube.com/watch?v=d12bKm9UEs4&t=25s
If the timecode doesn't work, you'll want to go to 25 seconds to see it.
Fairly consistently, during the "Power Dot Transition Phase" where the screen flashes, a purple tile artifact will appear onscreen (sometimes two, but rarely two). This issue has been around since at least October of 2020, but the recording supplied above is fresh from JimmyStones.