This is a bit early but I wanted to get a few things down now<\/p>\n
98 Program steps occupy 14 registers, 7 steps per register, in ram[32..45].
\nThe user program counter is in M2[2..0] as step, bank, register.
\nRegister starts as D (13) and works down to 0.
\nSteps are 6..0. Bank is always 2 (ram 32..47).
\nA value of 000 shows as step “00”.<\/p>\n
Memories are 0-9 in ram[0-9], .0-.5 (indirect 10-15) in ram[10-15], and indirect 16-29 in ram[16-29]. These are in ram banks 0 and 1.<\/p>\n
LastX is in ram[31]
\nX gets stored in lots of places (A\/B, C, M1, and ram[30] and ram[47])
\nThe magic number gets stored in ram[46]. If this is missing, the microcode reports an “Error” due to a loss of Continuous Memory.<\/p>\n
s3 is the PRGM-RUN switch flag. 1=RUN.
\ns5 is power ok. 0=low power.
\ns15 is keypressed<\/p>\n
The microcode includes 01120, 01220 and 01320 opcodes, which are unknown and presumed to be printer related.<\/p>\n
In the real world, only ram banks 0 and 2 are CMOS. In the emulator, all are.<\/p>\n
Whilst there is no way to share programs between a HP-29C and a HP-19C, they use the same hex codes for user program steps. This means the HP-29C could potentially attempt to execute “printer related” keys intended for a HP-19C. This lends credence to the unknown opcodes being “printer related”.<\/p>\n
User program step 1 is stored in the bottom 2 nibbles of ram[45]. The user program counter value for this is “62D”. “D” of bank 2 is ram[2*16+13]= ram[45] (ie ram[0x2D]).<\/p>\n