I spent more time on #NthPongWars today, looking into sound effects on the #NabuPC using the AY-3-8910 sound chip. I also converted some old assembler Pac-man sound code to C to get the hang of it, and found an abandoned sound experiment in the wacka-wacka code.
Longer story and technical details and code on my blog at https://web.ncf.ca/au829/WeekendReports/20240207/NthPongWarsBlog.html#2025.03.04
#NthPongWars continues on the #NabuPC. Today's head banging brought to you by #Z88DK / #SDCC compiler system for the Z80, where itoa() indirectly causes a crash or mysterious glitches as local variables suddenly change on their own.
Turns out the underlying utoa() assembler code uses more than the usual registers, including ix and the alternate set. And when compiling for speed, ix is the stack frame pointer! Hilarity ensues when you use a local variable pointer to write to memory :-)
More work on the #NABUPC Adaptor signal generation. Seems like I need to up my SPI game and learn how the DMA block works. The gaps between the bytes are almost as long as the bytes themselves.
I was obsessing over the inter-packet delays when the buffers were switched, but this is much more likely to be causing the issues I'm having with packets timing out and the NPC throwing "Adaptor failure" (which is a timeout or data corruption).
Good news! I've got the #NABUPC #NABU Adaptor's QPSK scrambler figured out and reimplemented in C. Next up, the packet format.
#retrocomputing
I've made some pretty good progress with the #NABUPC Adaptor reverse engineering, but I've got this unshakable feeling that I'll need to write an emulator and single-step the 6805 ROM code to see what the heck it's doing and figure out what the flags are...
edit: if you want to take a look at the work so far, it's on github: https://github.com/philpem/NABU-Adaptor/
