Alternate install for both the video and BIOS mod, working perfectly. Notice the chip on the far right? It’s the 74LS04, touching the board with only 5V and GND. 🙂
Also, one of the signals going to the 174 has been intercepted near the potentiometers in the middle of the board.
… I can now intercept and execute a real breakpoint, on demand by button, or a software breakpoint via the BRK opcode. The 7800 will send back it’s state and the address it broke on. Super cool, here is an example of the 2 flavors of breakpoints…
Now the trick is to request the breakpoint memory region through the I/O board and display a disassembly of it. 🙂 If that works fast enough, the gates are open, due to the virus like modification method I use to get debug functions working.
RevEng and me have been pondering on different things in regards to possibly adding a button to the cart, that allows to dump state/RAM contents back to the command line. I’ve setup a test rig for doing exactly this, I’ll be using the always overlooked cart interrupt for realizing this.
And it’s fast. Still a WIP, currently only works in Windows command line mode. 🙂
Time to test it. 🙂
PCB porn…
I basically told my program that there was a ROM already loaded (but it’s all $FF) and ran the diff (cool way to see big blocks of $FF data in ROM’s).
The output is formatted as follows: BANK, BLOCK_NUMBER, ADDRESS, NUMBLOCKS, BANKSWITCH_FLAG (SWITCH means do the bank switch).
Test output from the control software on OSX…
Port opened.
Loaded /Users/CPUWIZ/Desktop/3khotspotroms/jinks-3k.bin (131072 bytes).
Loaded /Users/CPUWIZ/Desktop/3khotspotroms/jinks-3k.bin (131072 bytes).
[BANK 0] Block: 0 @ $8000 (57) SWITCH
[BANK 1] Block: 64 @ $8000 (64) SWITCH
[BANK 2] Block: 128 @ $8000 (54) SWITCH
[BANK 2] Block: 191 @ $BF00 (1)
[BANK 3] Block: 192 @ $8000 (7) SWITCH
[BANK 3] Block: 208 @ $9000 (8)
[BANK 3] Block: 224 @ $A000 (8)
[BANK 3] Block: 240 @ $B000 (8)
[BANK 3] Block: 255 @ $BF00 (1)
[BANK 4] Block: 256 @ $8000 (7) SWITCH
[BANK 4] Block: 272 @ $9000 (8)
[BANK 4] Block: 288 @ $A000 (8)
[BANK 4] Block: 304 @ $B000 (8)
[BANK 4] Block: 319 @ $BF00 (1)
[BANK 5] Block: 320 @ $8000 (28) SWITCH
[BANK 5] Block: 352 @ $A000 (25)
[BANK 5] Block: 383 @ $BF00 (1)
[BANK 6] Block: 384 @ $8000 (28) SWITCH
[BANK 6] Block: 416 @ $A000 (25)
[BANK 6] Block: 447 @ $BF00 (1)
[BANK 7] Block: 448 @ $C000 (63) SWITCH
🙂
I managed to power the MCP IO kit from the 7800 and replaced the connection with a BT connection. It works, but even at 115200, it is super slow. I am not sure if this is a buffering issue or what, but instead of 8 seconds for a 32K file, it takes 2 minutes. However, it does work!
The SuperCart Rev.B board (16K to 256K with SARA support), has been finalized and submitted to the PCB manufacturer (you can thank ‘theloon’ from AtariAge for that). Super-banking (0880) is confirmed working without SARA, but PLD code for traditional F8/F6/etc. access has also been done and extended to 256K. Here is a quick PCB capture of the design that is now being built (black PCB’s, vs. Rev.A purple). 🙂
