doc | ||
file | ||
rom | ||
src | ||
.gitignore | ||
README.md |
tlc-apple2
Some bits and bobs regarding the Tiger Learning Computer. The TLC was a "toy" computer that had an Apple IIe at its core. There was a UI that would come up by default that was an early, proprietary windowing system that was little more than a program launcher.
The TLC can be booted and you can get into Applesoft BASIC without any problem. The problem is that I/O is pretty difficult. No one with one of these machines has come forward with the ability to get anything into or out of the existing serial port.
The goal: get data of the TLC out of the machine. Hard to do without functional I/O.
Audio to the rescue
One thing the TLC does have is sound. And a headphone jack. So we have... output.
A long time ago...
In the dark ages, when dinosaurs roamed the Earth, there was cassette tape. As an analog-digital storage medium. It was slow, crappy, error-prone, and all that - but it at least worked. Early Apple II specimens came with this audio interface built-in - the original Apple ][, all the way through to the last Apple IIe. But the audio jacks and attendant ROM code was missing from the IIgs, the IIc, and the IIc+. And the TLC.
Steal all teh cassette ROM
Handily enough, the early ROM code is quite tidy and compact; this is before the days of the crazy gymnastics of the IIe and beyond that moved Heaven and Earth to keep the entry points constant (because programmers are bad, and used them literally). The angle of attack:
- Copy out just enough Apple II ROM code to do the moral equivalent of the monitor's save memory to tape command (i.e.
*300.400W
) - Change cassette output (
$C020
) to speaker output ($C030
) - Assemble and get a raw hex dump of that code
- Type that code into the TLC's RAM and run it, dumping all non-banked ROM (
$C100-$FFFF
) - Capture the audio on a modern computer (Windows hint: run
soundrecorder /file path\to\filename.wav
to get a.wav
, not a.wma
) - Run that audio through CiderPress, which knows how to take tape audio and reconstruct data from it
- ???
- Profit
The source code to do these activities is in the src
directory, and the resulting ROM from the Tiger Learning computer is in the rom
directory. A few pictures of the machine and its on-screen interface are in doc/images
.
ROM contents
Nonbanked ROM dump is available in binary and in monitor ROM disassembly form:
C100-FFFF.bin
- Raw binaryC100-FFFF.txt
- Monitor disassembly
Banked ROM exists in slots 2 and 3. The contents have been dumped as follows:
Slot 2:
S2-C800-CFFF.bin
- Raw binaryS2-C800-CFFF.txt
- Monitor disassembly
Slot 3:
S3-C800-CFFF.bin
- Raw binaryS3-C800-CFFF.txt
- Monitor disassembly
File contents
The TLC comes with a virtual disk available in slot 6, drive 1 as volume /LTWIN
. This is the volume that is booted when the machine turns on.
The contents of it are as follows:
]catalog
/LTWIN
NAME TYPE BLOCKS MODIFIED CREATED ENDFILE SUBTYPE
BASIC.SYSTEM SYS 21 6-DEC-91 16:48 6-DEC-91 16:48 10240
PRODOS SYS 35 6-MAY-93 17:10 2-NOV-92 21:09 17128
STARTUP BAS 1 <NO DATE> <NO DATE> 176
KERLT101 BIN 50 11-JAN-96 9:18 11-JAN-96 9:18 24764 A=$1400
BLOCKS FREE: 14 BLOCKS USED: 114 TOTAL BLOCKS: 128
STARTUP
Bootstrapping the TLC follows the normal Apple II boot sequence and loads up PRODOS
and then runs the STARTUP
Applesoft program from /LTWIN
.
STARTUP
looks like this:
]load startup
]list
10 PRINT CHR$ (4);"BLOAD KERLT101"
20 CALL 5120
30 A = PEEK (25)
40 IF A=1THEN GOTO 10
50 HOME
60 VTAB 10: HTAB 10: PRINT "To return to Desktop,"
70 VTAB 11: HTAB 11: PRINT "please type LTWIN"
80 POKE (49304),0
90 POKE (49326),1
100 END
The command LTWIN
is a command that Tiger inserted into the ProDOS external command list (clumsily)
that basically reboots the system so that you end up back at the graphical interface.
KERLT101
The contents of KERLT101
has been dumped and the beginnings of a 6502bench SourceGen project are here:
KERLT101.bin
- Raw binaryKERLT101.bin.dis65
- SourceGen disassembly projectKERLT101.bin_cc65.lst
- ca65 assembly output listing
Entry is at $1400 which immediately jumps over a data area and starts up for real at $171F. There are clearly a lot of softswitches in the $C0xx space that the TLC is using. Figuring out what they do should be an interesting challenge.