The last days i’ve been selling a lot of my old computers. They have been in my collection for many years, but now its time to part. Time for others to enjoy them.
(Instead of posting which ones are being sold and which i’ve still got on this page i’ll make another post)
I started collecting when i studied computer sciences. It’s a wonder my parents attic wasn’t collapsing. (They let me store many computers on their attic, let me run a mainframe in the house (previous post) and let me have computer-parties (pre-lan) in their home. (They even left, and gave me the space) .. 15+ teens with computers … there was a pingpong table in the livingroom (besides the other tables in the house ) For all computers.
Then i’ve got even more, when living on my own. (At some point about 140. )
A few years later i got rid of uninteresting computers (to my taste at that time) and incomplete ones. Then i filtered-out the non working.
Still leaving with a lot of computers, i kept these for many years.
Now i only want the ones i’ve worked with, or are special to me.
My first computer was a Commodore Vic-20. Friends had the popular C64. So i kept 2 of both. In Junior Technical School i’ve used the BBC Acorn a lot (Funny story below) My then friend Richard had a Atari ST, loads of fun we had with that machine, so i’m keeping a Atari 1040STf. Another friend used a Apple SE, so that one i also keep for now. I’ve been programming a lot on 80×86, the first dos PC’s, i’m still looking for a old machine (Laser XT) which i used way back then. But for now i’ve got a Sinclair PC200. I’ll keep a old Commodore PET 2001, because its cute. Besides having a cute PET, i’ve got a Holborn System. Made in Holland (Enschede), there are only a few left according to some sites: only 200 made! (Holborn means Holland Born) One of the inventors was from Holten, my birthplace. (Polak)
Putting the system together in 2018
At school we kept a list of everyone’s collection.
Soo .. the story about the BBC Acorn.
When i was at school outside study hours, i went to the computer lab. This was one classroom with about 16 BBC Acorns and a master (teacher station). When they saw how enthusiastic i was, i got the key to the classroom. I even got access to the master system. And after a little hacking i’ve gained access to the teachers files. There was a simple network system, i think it was called Econet. The teachers system was the only one with a disk station.
I liked the ‘highres’ line graphics you could make on the machines. (640×256) So i’ve wrote a lot of programs using this mode. I even wrote a program which drew a 3D robotarm on screen using wireframe graphics. At that time my mathematics scores where .. bad. Wasn’t interested i think. But drawing 3D robotic arms are not possible using mathematics, like using sinus, triangulary etcetera. So when my mathematics teacher saw my program, he didn’t believe me. So .. fooling around in the computer lab, i missed start of classes. And later on .. worse .. i almost was not allowed to do my final exams. I was late several times (and one of the first to leave, …. straight from and to the computer lab. )
I’ve got some programs printed on paper, i will use my leftover BBC Acorn (or a emulator) to capture some screen examples.
Sold stuff
UPDATE : Selling a lot, but i’ve bought some others between 2020-2023
SDK-85
Laser Xt/3
80386 DX
Also a “new” 1084 monitor (CRT for a Commodore 64) Now i have to look for a VGA Crt to get old vga-register manipulation programs working.
Above is my design for a hex keyboard to enter opcodes in hex using a simple monitor program. i used a 74ls922 which can decode a 4×4 matrix. I’d rather had a 74ls723 which can encode 20 keys.
Nowhere to be found. So i have to think of a new plan.
Now it is configured as follows:
C
D
E
F
8
9
A
B
4
5
6
7
0
1
2
3
When pressing the alternate key
addr (to implement)
run (1/2 implemented)
reset (to implement)
step instruction (to implement)
memory next
memory previous
PCB design for matrix hexboard with place for notes
Meanwhile i’ve ordered new keys (the ones i’ve been using for my photomanager project and wnat to have a setup like this:
?
?
addr
run
reset
C
D
E
F
?
8
9
A
B
step
4
5
6
7
mem next
0
1
2
3
mem prev
When you want to show the status of busses and alike, you can’t use a led and restistor directly on the bus, it will require too much current. So i’ve been using below schematic which uses a darlington array.
Now i can display databus, address bus and i’ve been using this for address decoding logic and hex keyboard.
I’ve implemented a second VIA chip, and ordered components to amplify the SID sound part
Flashing ROMs .. (eeproms). It used to be a pain in the *$$. Burning took a looong time. But clearing one with UV took .. 20 minutes or so. Using one of these:
Altered clock module
Changed button press
Dipswitches for more speed control (red .. upper left)
Changed Rom/Ram
Changed addressing
Added RAM
ZIF Socket for ROM
VIC 6522
Fixed clock
Added buttons for interrupt
Display
Display works now
To test: Create Address logic to access display without VIA Can work, but not at high speed clock. Stays behind VIA
To buy: st7920 lcd 128×64
Generic improvements
Rewired most parts, using color codes (Blue data, Yellow Address and so on)
Added leds on data and address bus using ULN2803 darlington arrays
100nF Decoupling capacitors on the power rails
To do’s or ‘have to look into’s’
For sound i planned to use a General Instrument AY-3-8910, it is somewhere in my Lab, i know it is. I saved this chip and a SID for my Amiga addon soundcard. Where are my plans for the simple v1 setup? (FOUND IT)
I have to start writing rom functions for display usage. Like JSR $ff00 – Clear screen subroutine .. etc
I’m scraping information from websites, to get started on my clock controller. ATmega328 with ssd1306 display and rotary encoder/dip switches
Notes about the movie: Left side is Arduino IDE monitor reading Addressbus and Databus. (I’m going to try to rewrite this to realtime disassemble) Resetting system. Stepping CPU with manual clock pulses. Start vector being read at $FFFC/$FFFD. Program being run from $8000. Set clock on automatic ( ~ about 150 Hz ) Last opcodes you see a JMP loop 4C 2F 80, that is JMP $802F Display enlarged on video, was not visible on movie i took on mobile. (Wrong angle?)
Breadboard overview
Clock module
Reset module + Crystal
CPU + nmi/int buttons
RAM and ROM
Address decode + Bus divide
Addres/Data bus leds
6522 VIA + Display
2nd via + Buttons
?
(sound board)
TIL: 6502 can run without ram only rom,expect when using JSR … which uses a program stack in RAM
Cartridge printEeproms 8k and 32k (also for 6502 project)Eeprom programmer
I’ve got the tools and Bigred made me enthusiastic again. My goal is to make a C64 Cartridge from a PRG. And Not any program, it is the 8085 Emulator from Sepp.
Serveral problems i have to ‘fix’
The program is 17K, Cartridges can only be 16K. So i have to use 2x 8K and compress the data. This means it have to be uncompressed at start time. ( I was thinking of using exomiser for this )
Program starts normally at $0820 and probably is not optimised to run anywhere else. So a starting routine has to copy the program from cartridge memory to the correct location
Luckily i have the source! How cool is that
For version 4.73 it states : Starting at $0820 .. but my hexdump is off by one??!?
00000020 00 20 ec starts with 00 at $0020 .. and not 20 ?!?!
Tools used until now:
Vice – C64 Emulator x64 -cartcrt 8085.crt
c1541 – Linux disk tool for C64 images. Used this to extract the 8085emulator PRG
prg2crt.py – a convertor from PRG to a cartrid file which can be used by Vice python2 prg2crt.py 8085.prg 8085.crt
minipro – eeprom programming tool for Linux minipro -p AT28C64 -w /tmp/test.bin
cartconv (tool from vice to convert crt <-> bin) cartconv -t normal -i test.bin -n ‘my cart’ -o test.crt
xa – Cross assembler 65xx/R65C02/65816
ACME – the ACME Crossassembler for Multiple Environments
Memory Map C64 – source c64-wiki.com
Card Low starts at $8000, so that’s the place where those roms are going to be. To place on this address:
Copy routine : from ($8000 + this copy routine) to $0820 When to decompress?? jmp routine to $0820
A cartridge file >16K and with his emulation headers seems to work??!
Also nice: Magic Desk Cartridge Generator V3.0
UPDATE: 20220811
exomizer sfx 0x0820 8085.prg -o data.exo # Compress and start at 0x0820
xa frame.asm -o frame.bin # Add code and write binary
x64 --cart16 frame.bin # Test cartridge with Vice
frame.asm
;----------------------------------------------------------
; example usage
; xa frame.asm -o frame.bin
; cartconv -t normal -i frame.bin -n 'my cart' -o frame.crt
; x64 -cartcrt frame.crt
;----------------------------------------------------------
;no load-adress for bin-file, so no header here
*=$8000
.word launcher ;cold start
.word launcher ;warm start
.byte $c3 ;c
.byte $c2 ;b
.byte $cd ;m
.byte $38 ;8
.byte $30 ;0
launcher
stx $d016
jsr $fda3 ;prepare irq
jsr $fd50 ;init memory
jsr $fd15 ;init i/o
jsr $ff5b ;init video
;make sure this sets up everything you need,
;the calls above are probably sufficient
ldx #$fb
txs
;set up starting code outside of cartridge-area
move_starter
ldx #(starter_end-starter_start)
loop1
lda starter_start,x
sta $100,x
dex
bpl loop1
jmp $100
;---------------------------------
starter_start
ldx #$40 ;64 pages = 256 * 64 = 16384 Bytes
ldy #0
loop
src
lda exomized_data,y
dst
sta $801,y
iny
bne loop
inc src+2-starter_start+$100
inc dst+2-starter_start+$100
dex
bpl loop
;make sure settings for $01 and IRQ etc are correct for your code
;remember THIS table from AAY64:
; Bit+-------------+-----------+------------+
; 210| $8000-$BFFF |$D000-$DFFF|$E000-$FFFF |
; +---+---+-------------+-----------+------------+
; | 7 |111| Cart.+Basic | I/O | Kernal ROM |
; +---+---+-------------+-----------+------------+
; | 6 |110| RAM | I/O | Kernal ROM |
; +---+---+-------------+-----------+------------+
; | 5 |101| RAM | I/O | RAM |
; +---+---+-------------+-----------+------------+
; | 4 |100| RAM | RAM | RAM |
; +---+---+-------------+-----------+------------+
; | 3 |011| Cart.+Basic | Char. ROM | Kernal ROM |
; +---+---+-------------+-----------+------------+
; | 2 |010| RAM | Char. ROM | Kernal ROM |
; +---+---+-------------+-----------+------------+
; | 1 |001| RAM | Char. ROM | RAM |
; +---+---+-------------+-----------+------------+
; | 0 |000| RAM | RAM | RAM |
; +---+---+-------------+-----------+------------+
lda #$35 ;cart is always on instead of BASIC unless it can be switched off via software
sta $01
jmp $80d ;for exomizer, i.e.
starter_end
;----------------------------------
exomized_data
.bin 2,0,"data.exo"
;syntax for exomizer 2.0.1:
;exomizer sfx sys game.prg -o data.exo
main_file_end
;fill up full $4000 bytes for bin file ($c000-$8000=$4000)
.dsb ($c000-main_file_end),0
Exomiser info
Reading "8085.prg", loading from $0801 to $4CE9.
Crunching from $0801 to $4CE9.
Phase 1: Instrumenting file
-----------------------------
Length of indata: 17640 bytes.
[building.directed.acyclic.graph.building.directed.acyclic.graph.]
Instrumenting file, done.
Phase 2: Calculating encoding
-----------------------------
pass 1: optimizing ..
[finding.shortest.path.finding.shortest.path.finding.shortest.pat]
size 80273.0 bits ~10035 bytes
pass 2: optimizing ..
[finding.shortest.path.finding.shortest.path.finding.shortest.pat]
size 80039.0 bits ~10005 bytes
pass 3: optimizing ..
Calculating encoding, done.
Phase 3: Generating output file
------------------------------
Encoding: 1101112133423160,1122,2010223445667788,032144406789BBCD
Length of crunched data: 10034 bytes.
Crunched data reduced 7606 bytes (43.12%)
Target is self-decrunching C64 executable,
jmp address $0820.
Writing "data.exo" as prg, saving from $0801 to $304C.
Memory layout: |Start |End |
Crunched data | $07E7| $2F18|
Decrunched data | $0801| $4CE9|
Decrunch table | $0334| $03D0|
Decruncher | $00FD| $01C0| and $9F,$A7,$AE,$AF
Decrunch effect writes to $DBE7.
Decruncher: |Enter |During|Exit |
RAM config | $37| $37| $37|
IRQ enabled | 1| 1| 1|
UPDATE:20230126
; CODE COPY FROM http://www.lemon64.com/forum/viewtopic.php?t=60786&sid=2559442c8b963d7aac27cb13b493f372
; Thanks for posting: Richard of TND
; this is for a 16KB cart, using ACME!!
!to "mycart.crt",cart16crt
scr = $0400
DecrunchADDR = 2061 ;SYS 2061 (HEX $080D)
*=$8000
!word launcher
!word launcher
!byte $c3,$c2,$cd,$38,$30 ;CBM 80
launcher
sei
stx $d016
jsr $fda3 ;prepare irq
jsr $fd50 ;input memory
jsr $fd15 ;initialise i/o
jsr $ff5b ;initialise video memory
;For a more professional boot up. Make
;the border and screen black. AFTER
;the video memory, etc has finished.
lda #$00
sta $d020
sta $d021
cli
;Switch off the screen.
lda $d011
and #%11101111
sta $d011
;Move transfer code over to the screen
;memory.
ldx #$00
tloop lda transfer,x
sta scr,x
inx
bne tloop
jmp scr
transfer
ldx #$00
tr1 lda linkedgame,x ;Move from linked address
sta $0801,x ;Direct to BASIC start address
inx
bne tr1
inc scr+4
inc scr+7
lda scr+4
bne transfer
jsr $e453 ;load basic vectors
jsr $e3bf ;init basic ram
ldx #$fb
txs
;Execute the game, by jumping to the
;de-cruncher's start address.
;jmp $0820
jmp DecrunchADDR
;Link crunched game as a PRG file to memory after
;the cartridge build code.
linkedgame
!bin "8085sys.prg",,2
FileSize = *
!if FileSize >$c000 {
!error "FILE SIZE IS TOO BIG TO FIT 16KB CARTRIDGE"
} else {
*=$c000
}
Exomizer:
exomizer sfx sys 8085.prg -o 8085sys.prg
Reading "8085.prg", loading from $0801 to $4CE9.
Crunching from $0801 to $4CE9.
Phase 1: Instrumenting file
-----------------------------
Length of indata: 17640 bytes.
[building.directed.acyclic.graph.building.directed.acyclic.graph.]
Instrumenting file, done.
Phase 2: Calculating encoding
-----------------------------
pass 1: optimizing ..
[finding.shortest.path.finding.shortest.path.finding.shortest.pat]
size 80273.0 bits ~10035 bytes
pass 2: optimizing ..
[finding.shortest.path.finding.shortest.path.finding.shortest.pat]
size 80039.0 bits ~10005 bytes
pass 3: optimizing ..
Calculating encoding, done.
Phase 3: Generating output file
------------------------------
Encoding: 1101112133423160,1122,2010223445667788,032144406789BBCD
Length of crunched data: 10034 bytes.
Crunched data reduced 7606 bytes (43.12%)
Target is self-decrunching C64 executable,
jmp address $0820.
Writing "8085sys.prg" as prg, saving from $0801 to $304C.
Memory layout: |Start |End |
Crunched data | $07E7| $2F18|
Decrunched data | $0801| $4CE9|
Decrunch table | $0334| $03D0|
Decruncher | $00FD| $01C0| and $9F,$A7,$AE,$AF
Decrunch effect writes to $DBE7.
Decruncher: |Enter |During|Exit |
RAM config | $37| $37| $37|
IRQ enabled | 1| 1| 1|
exomizer sfx $\0801 8085.prg -o 8085out.prg
Reading "8085.prg", loading from $0801 to $4CE9.
Crunching from $0801 to $4CE9.
Phase 1: Instrumenting file
-----------------------------
Length of indata: 17640 bytes.
[building.directed.acyclic.graph.building.directed.acyclic.graph.]
Instrumenting file, done.
Phase 2: Calculating encoding
-----------------------------
pass 1: optimizing ..
[finding.shortest.path.finding.shortest.path.finding.shortest.pat]
size 80273.0 bits ~10035 bytes
pass 2: optimizing ..
[finding.shortest.path.finding.shortest.path.finding.shortest.pat]
size 80039.0 bits ~10005 bytes
pass 3: optimizing ..
Calculating encoding, done.
Phase 3: Generating output file
------------------------------
Encoding: 1101112133423160,1122,2010223445667788,032144406789BBCD
Length of crunched data: 10034 bytes.
Crunched data reduced 7606 bytes (43.12%)
Target is self-decrunching C64 executable,
jmp address $0801.
Writing "8085out.prg" as prg, saving from $0801 to $304C.
Memory layout: |Start |End |
Crunched data | $07E7| $2F18|
Decrunched data | $0801| $4CE9|
Decrunch table | $0334| $03D0|
Decruncher | $00FD| $01C0| and $9F,$A7,$AE,$AF
Decrunch effect writes to $DBE7.
Decruncher: |Enter |During|Exit |
RAM config | $37| $37| $37|
IRQ enabled | 1| 1| 1|
This looks okay: (monitor in vice)
Attaching crt in vice
Maybe one of these problems:
1) you CAN NOT use BASIC routines when a cart is inserted (without weird tricks, i.e.
storing BASIC routines on cart etc)
2) you need to be careful about $01 as you may map in ROM at $8000 without expecting it.
Please refer to this if in doubt:
http://unusedino.de/ec64/technical/aay/c64/memcfg.html
[3] You should also be careful about the usage of KERNAL routines as some of them
sweep across BASIC-code as well!
After a whole day soldering yesterday, ending up with a wire mess. Which didn’t work at the end…
Starting measuring some things, and create some test sketches (led blinky tests) I found out that the main problem was not having the red switches connected to GND. Blue switches where upside down, this was a easy fix. Because these are ON-ON switches, and where already connected to a common line. Then a mixup between D0 and D6 (wires crossed) And it is working! Made some lines and lettering on the frontplate after some playing around.
65c22 connected, new data, and address-bus ribboncables!
First led on Register B blinking!
Notes: Temporary display wil be 2×16 Chars. Ram in place, but not connected (is emulated by the Arduino Mega at the moment) Rom is somewhere halfway the atlantic ocean .. still waiting on that one. Ben Eatons clock module is disconnected, i’m using the Arduino as programmable clock right now. (There wil be a little display and a rotary encoder to set clock speed.)
lda #$ff ; all bits
sta $6002 ; set direction (out) for B register
lda #$80 ; set 1 bit
sta $6000 ; set register B
lda #$00 ; reset bit
sta $6000 ; set register B
jmp $8005 ; jmp to bit set part
Due to eeproms being scarce, i’m going to use a arduino as Rom emulator. Below is a test setup i’m going to build.
Made the drawing in Kicad.
KiCad is a free software suite for electronic design automation. It facilitates the design and simulation of electronic hardware. It features an integrated environment for schematic capture, PCB layout, manufacturing file viewing, SPICE simulation, and engineering calculation.
Memory assignment:
$8000-FFFF - Rom
$4000-7FFF - Ram ?
$2000-3FFF - Multiple times the 6522 *
$0000-???? - Ram probably
* This is due to the fact i am only using Address lines: 0,1,2,3,13,14,15
As posted before, i found a cartridge in my collection a while ago. I don’t know where i’ve got this one from.
While searching on the internet for more information, i really couldn’t find anything about it. Not even on collectors sites.
Where did it come from, what does it do?
It says: Data Manager 64 (1984) , Softworld and Handic.
I can find many cartridges by Handic. I’ve checked all database/filers i could find. Non looked like this.
Some screenshots
Next to try .. Dumping the cartridge to file. Probably by changing the way the C64 starts a autostart cartridge. So i have to modify the cartridge port, and put a cross connection on the lines EXROM and GAME, this will change the memory addresses ($8000/$A000) where a cartridge is placed. Then the commodore will start normally, and i can dump the cartrige memory locations to a file.
So pins 8 and 9 have to be switched around. The C64 autostarts a cartridge when it find certain data on $8000
Meanwhile i try to contact some collectors of cartridges.
Next week i’ll be going to May Contain Hackers 2022, what to bring? My old friend Bigred will be there, many others couldn’t make it …
What to bring and do:
Laptops
Arduino touch project?
My new 6502 breadboard computer?
The DVB-T / DAB / FM stick
Besides the emulators on my laptop, maybe i’ll bring this little thing (Or a real C64?)
Booting in 4 seconds! Running Vice in 50 or 60 Hz Low latency video! Can emulate cartrides, floppy’s When you connect joysticks or a real C64 keyboard to the GPIO pins it will use that. SID sound using ReSid CRT emulation (look for BMC64 or combian)
"If something is worth doing, it's worth overdoing."
