I’ve been busy programming Python and NodeRed for a client. But these are the things I’ve done in the last days.
C64 Assembly: Breaking borders, using sprites and multicolor font intro.
It does not look impressive, but I’ve learned a lot. Found a new way (for me) to open borders and change border colours on predefined raster lines. Sources will be posted.
KiCad tutorial, posted on YT also because I could not find many resources about the subject online. Maybe it’s helpful
Video editing using Kdenlive.
Edit: Even faster, use Netlabels, no need to join pins. Press L (uppercase) select pin 1, name 1. Press and hold insert until all pins named. Copy paste socket 5 times and goto your PCB tab.
This movie is about creating a backplane for a 6502 SBC I’m building. It is real-time and below 4 minutes.
Multi Keyboard
My small multitouch screen came in. This is for my previously mentioned multi-computer case.
It is going to show multiple keyboard layouts for different systems. (See previous posts about this)
Waveshare display, Raspberry Zero as HID device, using USB and pin emulated keyboards. (c64 matrix, AT (DIN) keyboard, ps2 keyboard)
Some example screens
Vic-20
Photo-realistic
Petscii C64
Another C64
I’m also going to make a layout like the keyboards on my 8085
Started working on my breadboard version of a 68k computer. When it’s working, I’ll make a PCB version. Using almost only parts I still have. (No 8mhz crystal)
The 68000 being 24 bit address and 16 bit data needs 2x 8-bit roms and 2x 8 bit ram, but i didn’t have the components yet in this picture.
While tinkering with above, my Fatter Agnus chip came in.
To make a 1mb chipmem version of your rev 5 amiga (PAL)
You need to have a newer version of the Agnus chip (I had 8371, and bought a 8372a) AND you need a 512kb trapdoor memory expansion.
An unmodded rev 5 will see 512kb Chip mem and 512 Fast mem.
Replacing the Agnus 8371 for 8372a: I lost my PLCC puller, so I modded a paperclip into a puller 🙂
MISSING IN ACTION
When placing the new chip, I had to tape pin 41 for PAL version. I used Polyimide Film tape.
Next I had to cut the jumper 2 connection and solder the other pads. (Bottom and middle disconnect and middle and top bridged)
Next was another cut on the PCB, this disables the trapdoor card detection.
Re-learning the little I knew (I never had a c64 as a kid). Back to basics, welll machine code I mean.
Programming a little demo using acme. Split screen bitmap and text mode plus sid music
Running a little demo in retrodebugger (missing the sid music in the recording)
Some useful commands
; Dump prg with offset 0x800 per byte and skip 00 00 lines xxd -o 0x800 -g1 icecrew.prg | uniq -f10
; Write symbol list acme -l icecrew.sym icecrew.asm
; png to kla (koala picture) retropixels icecrew.png -o icecrew.kla
; relocate a sid address sidreloc -r org.sid new.sid
Below code has some flaws:
Many empty gaps, creating a large file. Exomizer could fix this, but better memory management should be the better solution. The Koala file has many 0 bytes, the logo is small but the file is created for a full screen image.
Part of the program see $1000 of start of SID music
!cpu 6502
!to "icecrew1.prg",cbm
; Standard basic sys runner
basic_address = $0801
; sid addresses
; address moved using
; sidreloc -r Lameness_Since_1991.sid lame.sid
; addresses found using
;sidplay2 -v lame.sid
;+------------------------------------------------------+
;| SIDPLAY - Music Player and C64 SID Chip Emulator |
;| Sidplay V2.0.9, Libsidplay V2.1.1 |
;+------------------------------------------------------+
;| Title : Lameness Since 1991 |
;| Author : Peter Siekmann (Devilock) |
;| Released : 2017 Oxyron |
;+------------------------------------------------------+
;| File format : PlaySID one-file format (PSID) |
;| Filename(s) : lame.sid |
;| Condition : No errors |
;| Playlist : 1/1 (tune 1/1[1]) |
;| Song Speed : 50 Hz VBI (PAL) |
;| Song Length : UNKNOWN |
;+------------------------------------------------------+
;| Addresses : DRIVER = $1C00-$1CFF, INIT = $0FFF |
;| : LOAD = $0FFF-$1B25, PLAY = $1003 |
;| SID Details : Filter = Yes, Model = 8580 |
;| Environment : Real C64 |
;+------------------------------------------------------+
;
sid_address = $0fff
sid_play = $1003
sid_init = $0fff
; Character
char_address = $3800
screen_mem = $4400
; Koala address
bitmap_address = $6000
bitmap_data = $7f40
bitmap_color = $8328
bitmap_bgcolor = $8710
program_address = $c000
color_mem = $d800
reg_d011 = $D011
; VIC register
;Bit 7 (weight 128) is the most significant bit of the VIC's nine-bit raster register (see address 53266).
;Bit 6 controls extended color mode
;Bit 5 selects either the text screen ("0") or high resolution graphics ("1").
;Bit 4 controls whether the screen area is visible or not.
;Bit 3 selects 25 (when set to "1") or 24 (when set to "0") visible character lines on the text screen.
;Bit 0–2 is used for vertical pixel-by-pixel scrolling of the text or high resolution graphics.
; Rom routine to clear screen ( slow ! )
; Better to do this yourself
clear_screen = $e544
* = sid_address
!bin "lame.sid",,$7c+2
; standard charset
* = char_address
!bin "charset.chr"
; drawn with gimp converted using retropixel
; retropixels icecrew.png -o icecrew.kla
* = bitmap_address
!bin "icecrew.kla",,$02
; sys 49152
* = basic_address
!byte $0d,$08,$dc,$07,$9e,$20,$34,$39,$31,$35,$32,$00,$00,$00
* = program_address
sei
; init
lda #$00
tax
tay
jsr sid_init
jsr clear_screen
jsr load_bitmap
jsr init_text
ldy #$7f
sty $dc0d
sty $dd0d
lda $dc0d
lda $dd0d
lda #$01
sta $d01a
lda reg_d011
and #$7f
sta reg_d011
; move interrupt vector to bitmap
lda #<interruptbitmap
ldx #>interruptbitmap
sta $314 ; Low Address part IRQ vector
stx $315 ; High Address part IQR vector
ldy #$1b
sty reg_d011
lda #$7f
sta $dc0d
lda #$01
sta $d01a
; trigger interrupt at rasterline 0
lda #$00
sta $d012
cli
jmp *
interruptbitmap
inc $d019
; trigger interrupt at rasterline 128
lda #$80
sta $d012
lda #<interrupttxt
ldx #>interrupttxt
sta $314
stx $315
jsr bitmap_mode
jmp $ea81
interrupttxt
; ack IRQ
inc $d019
; IRQ at line 0
lda #$00
sta $d012
lda #<interruptbitmap
ldx #>interruptbitmap
sta $314
stx $315
jsr text_mode
jsr sid_play
jmp $ea81
bitmap_mode
; bitmap graphics multicolor
lda #$3b
sta reg_d011
lda #$18
sta $d016
; switch to video bank 2 ($4000-$7FFF)
lda $dd00
and #$fc
ora #$02
sta $dd00
lda #$18
sta $d018
rts
text_mode
; set text mode hires
lda #$1b
sta reg_d011
lda #$08
sta $d016
; switch to video bank 1 ($0000-$3FFF)
lda $dd00
and #$fc
ora #$03
sta $dd00
; set charset location
; 7 * 2048 = $3800, set in bits 1-3 of $d018
lda $d018
ora #$0e
sta $d018
rts
load_bitmap
lda bitmap_bgcolor
sta $d020
sta $d021
ldx #$00
copy_bmp
; screen memory
lda bitmap_data,x
sta screen_mem,x
lda bitmap_data+256,x
sta screen_mem+256,x
lda bitmap_data+512,x
sta screen_mem+512,x
lda bitmap_data+768,x
sta screen_mem+768,x
; color memory
lda bitmap_color,x
sta color_mem,x
lda bitmap_color+256,x
sta color_mem+256,x
lda bitmap_color+512,x
sta color_mem+512,x
lda bitmap_color+768,x
sta color_mem+768,x
inx
bne copy_bmp
rts
init_text
ldx #$00
copy_txt
lda text1,x
sta $0400+520,x
lda text2,x
sta $0400+640,x
lda text3,x
sta $0400+640+120,x
lda #$06
sta color_mem+520,x
lda #$0e
sta color_mem+640,x
lda #$0e
sta color_mem+640+120,x
inx
cpx #$28
bne copy_txt
rts
text1
!scr " back to oldskool demos in 2024 "
text2
!scr " greetings to bigred & tyrone & edk "
text3
!scr " a lot to relearn - keep coding! "
I wanted this to be a multipart loader, instead of a trackloader. A trackloader can load sector parts which I would like more. But the C64Pico can’t do disk images. (Mcume)
C64Pico based on MCUME see building of this in other posts.
2nd reason: While I’ve written a track loader for 8086, I never did it for C64. As a kid I didn’t have a C64, so all knowledge I have is from later years. I’ve written only a few C64 machinecode programs.
Showing first part assembly (without text Hello 2nd part)
Showing second part (no sysheader) needs to be loaded at $2000
Compile using Acme
make disk image
and run using autostart x64 (Vice emulator)
You see the first text from the 1st assemby code, then it will load the second at $2000 and does a jmp to this address. Second text will but displayed.
While i’ve been using KickAss in the past and some other 6502 compilers, I manly use acme.
Makefile I created to compile, create a C64 diskimage and run the program is as below. (No exomizer tools in this Makefile)
Today we worked on this project again. (Bigred and me)
There were some problems we needed to fix since last time:
It was quite hard to get the correct parts. Our display connector was only fitted with connection pins on the wrong side of the connector. (up/down) So I bought a connector with both positions populated. So we had to replace this hard to solder (40 pin) connector.
It was not clear what the orientation should be of the atmega328pb. We looked at the pinout, and followed the VCC/GND. But these are also available of the opposite side of the chip. (We missed that) Later, we saw a tiny line on the PCB, which showed the pin 1 placement. So we had to remove and replace the chip. When turning on the power, (with incorrect placement) probably fried R5 (10k resistor), on both our boards. Had to replace those also.
Programming the atmega328pb was not easy, see below fixes.
Compiling the pico firmware resulted in a black screen. Below the fixes I had to make to get the screen working.
Other things still to fix.
Bigreds screen.
atmega328p didn’t work for Bigred, so probably needs to replace with the pb version.
My battery controller is not charging. See bottom of page
Some of my buttons are working. The pewpew and some of the cursor keys (not as I expect, there are some up/down issues) And none of the other keys are working.
Some other things we noticed.
sdcard: remove partitions, format using mkfs.exfat Create a c64 directory on this filesystem where you can put the d64 files!
0402 SMD is far too small for me. There is enough room on the board to use 0805 for example. Even THT is possible, there are only a few components.
Some components are TOO close together, removing a component resulted in other small parts disconnecting also.
My friend Bigred said: If I can see it, I can solder it. But it is not easy. This probably keeps a lot of people from building it!
I modded it and was planning to rebuild this using pluggable eurocard-prints. Then it got lost, somewhere.
Today I went to Almelo with Tyrone. In the morning reverse engineering a lift controller print, and afternoon going to a guy selling a lot of Retro stuff.
And there it was, an Action Replay II for Amiga just catching dust.
I had to buy it, and got a sh*tload of 27256 Eproms for free!
With my own copy of the manualRipping graphics and music
Action Replay Mk I
This version is compatible with the A500/A1000 version only. It also plugs into the side expansion port. It introduces the following features:
1.0 Version
Shows and modifies registers (even read-only ones) and memory contents.
Trainer maker.
M68000 assembler / disassembler.
Copper assembler / disassembler.
Sprite editor.
Virus detector.
Picture / music (tracker format) / sample ripper.
Save computer memory (freezed programs) to disk.
Shows computer status (disk parameters, ChipRAM, FastRAM…).
Features added to the 1.5 Version
Mempeeker.
Ability to save freezed programs to RAM.
RAM testing.
Illegal opcode – jumps to freezer mode.
Action Replay Mk II
A special A2000 version is available for this particular revision. Instead of plugging into the side expansion port it plugs into the 86 pin CPU slot.
Features added since MK 1.5 version
Boot selector.
Picture editor.
Sound tracker.
Turbo fire manager (separately for both joysticks).
Disk encoder.
Start menu.
Disk monitor.
Integrated DOS commands (Dir, Format,…).
Diskcopy.
80 characters display with two-way scrolling.
Calculator.
Notepad.
Memory and drive switch (enabling / disabling).
Music ripper now finds all tracker formats (SoundTracker, NoiseTracker, other formats with 32 samples).
I was afraid to start this myself, SMD is on another level for me. But my good friend Marco said … No problem!
So I ordered components online, which was not easy. Selecting the correct parts, sizes and options.
Finding orientations of the componentsThe master at work, he has always been our soldering master (see GPC)Using a microscopeFluxWickI have to do one myself
These things are really really small
1.6mm x 0.8mm40 connections / 20mm !
Using tweezers to place the components was even difficult. The slippery tiny bastard got catapulted everywhere. (Or got stuck on fingers, soldering iron and alike) Many small components got lost into the 7th dimension. Never to be found again.
Awesome to work on this together, but Marco said that I have to try it myself. Welllll, I got 3/4 of the ATmega328PB-A perfectly soldered, then I notished that it was crooked. Desoldering was a mess, and I heated the PCB TOO much with the heatgun.
My messed-up PCB, and f*cked-up IC. Leave it to the professionals.
Next step for me is soldering the 75 mini buttons!
Got a Trinitron display from him, I was looking for this for a long time.
"If something is worth doing, it's worth overdoing."
