News:

Printed Amstrad Addict magazine announced, check it out here!

Main Menu

Recent posts

#21
Mostly bad sectors came not from the magnetic surface.
Storing in a wet atmosphere is a problem (cellar f.e.) Then you have pimpels  out of water and dust on your disc...

xesrjb 
#22
Quote from: robcfg on Yesterday at 16:34Then you have the sprites which I don't remember if they use their own palette entries.

Yes, the sprites have a separate palette of 15 possible colours. 
#23
D
GFX + Tunes / Re: Amstrad music
Last post by dal - Yesterday at 18:03
#24
The CPC classic has a palette of 27 colours, of which it can show 16 on Mode 0, 4 on Mode 1, and 2 on Mode 2. Any of the 27 can be used as border colour.

Now, the CPC+ have a palette of 4096 colours while having the same limitations per Mode. Then you have the sprites which I don't remember if they use their own palette entries.

What is it exactly you are trying to do?
#25
It doesn't change the 27 colours in the "old" palette. If you set any colour using the "old" palette mechanism it re-writes that to the relevant fixed RGB value and stores it in the ASIC palette registers.

To select from any of the 4096 colours available you have to use the "new" method instead, which is basically:

Page in the ASIC registers at #4000
Write the relevant values to the appropriate ASIC register
(Optionally) Page out the ASIC registers again

How difficult that is to patch into any given game depends a lot on how it works. If the code is running (or has to access data) from #4000 to #7fff then you may have to relocate it or use a springboard function elsewhere so that the ASIC doesn't get in the way. In addition you're going to need at around twice as many bytes to store colours (unless they're fixed throughout).

For anything that changes the palette dynamically you also need to consider that changes to the Plus palette registers are (unfortunately) not atomic so you will potentially see a glitched colour if you change it at the point where that colour is on-screen.
#26
Hi,

In CPC 6128 we have "only" 27 colors
I ask about using more color in CPC+, does it just change a 27 colors palette ?
In other way can I theoricaly hack for sample super cauldron ?
~in this way I can hack only one CPC+ game, show me the best choice :)
#27
QuoteOr do you need the reeeaally old Pre-Autodesk Eagle 3.0 ?
Thanks for that. I'm not so sure now...

ChatGPT has been hallucinating bad and incorrectly assumed that LIB files were used by early versions of Eagle. I called on its BS after locating a Eagle 3.55 manual which discussed only the LBR (XML format)

So I went through the SCRipt file generated from the V and LIB files found the symbols it references

QuoteC050-025X075@rcl
E2,5-5@rcl
DO35-10@diode
DIL28@ic-package
2X04@pinhead
PAK100/2500-50@con-3m
PLCC-S44-SOCKET@ic-package
SD4-08@switch-dil
SIL9@resistor-sil

So now I reference these which I've added a 'USE' for in the SCR file:

Quote# Libraries
USE rcl ;
USE pinhead ;
USE switch-dil ;
USE resistor-sil ;
USE ic-package ;
USE diode ;

USE 50WEDGE ;
With this project the provided BRD files and 50WEDGE XML files were for Eagle 6.4. So I'm back to Eagle 6.6 now trying to run this SCP file.

I'm still getting 'Device not found' error and have now RTFM and found the scripts are only to build the BRD file, so cannot get a schematic from these. Kicad needs a BRD and SCH file to import, so now I'm trying nl2sch to convert netlist to schematic to get something loaded in
GitHub - tpecar/nl2sch: Netlist to KiCad schematic




#28
Hi,
     you can download Eagle 3.0 Trial version (Released 2023) here: https://www.filehorse.com/download-eagle/72076/
Or do you need the reeeaally old Pre-Autodesk Eagle 3.0 ?

Bryce.
#29
Quote from: tubs74 on 14:51, 20 August 25I think trying to get a strong magnetic wiper might restore them - no harm in trying.
A strong magnet did indeed help with many. In most cases it reduced the bad sectors, sometimes even made the full disk working again. But in some cases it didn't help at all. 
#30
I've been wanting to rebuild the @revaldinho CPC Sixrom card and are having issues. I'm wanting to add changes (JTAG connector and change Xilinx from 5V PLCC to  3.3V VQFP with regulator) and do this inside Kicad
https://github.com/revaldinho/cpc_ram_expansion/wiki/CPC-Sixrom-Card

Apparently I can convert from Eagle to Kicad if the files are saved in later Eagle versions (V9)


Whilst its great to have these on Github, the toolchain is so old, that sourcing and configuring it has taken quite a bit of time and I'm not even half way through yet.
 
Steps
1. There are BRD files inside the Eagle releases folder (noting here that Eagle 6.4 is required)
https://github.com/revaldinho/cpc_ram_expansion/tree/master/eagle_releases

For anyone trying to get Eagle downloads for older versions, it is very difficult to locate. What I did find was the Cadsoft FTP server (RIP) was archived and contains most versions (including Eagle 6.6 which can load these BRD files)
https://archive.org/download/falconk_archivebot_ftp_cadsoft_de_20170121

This uses Warcat to extract the files. I tried various PERL versions but those had issues. This one works:
https://warcat-rs.readthedocs.io/en/latest/downloads.html

*Edit* other versions (1.0 - 2.4)
https://alkesgroup.broadinstitute.org/Eagle/downloads/old/
Version 4.08
https://archive.org/details/spidla-elektro2

2. I would also like to get a SCH schematic for this and started with the 6 ROM source folder:
https://github.com/revaldinho/cpc_ram_expansion/tree/master/cpc_sixrom/pcb

What we see in here is a netlist file, footer file and a TCSH script which uses fiies in the common_pcb folders:
https://github.com/revaldinho/cpc_ram_expansion/tree/master/common_pcb
https://github.com/revaldinho/netlister

So these scripts convert the netlist to PCB files using some Jython Netlister program. Like most people I'm running Windows and have Cygwin and Msys/Mingw to do some Unix stuff.

This is looking a bit involved but we give it a shot:
https://www.jython.org/download
Install this and add it to the MSYS path
export PATH="/c/tools/jython/bin:$PATH"
Then to install TCSH to msys
pacman -S tcsh
The run the README TENEX CSH script. There are some errors like missing CLASSPATH variables
setenv CLASSPATH ""
Also incorrect hard coded folders inside the scripts (README and common_pcb/Makefile that need fixing) and problems with using MSYS drive mapping with /c/ translating to C: incorrectly. So those are fixed in both files

make cpc_sixrom.scr
Finally this works, and we have an Eagle SCR script file and the NETlist files to build the schematic.


3. Attempt to load the SCR file inside Eagle and get a schematic to use with our BRD files and update

We start getting errors. Unknown layer:
Quote# Netlist cpc_sixrom
Layer Top;
Grid Inch;
ChatGPT says this is a very old netlister (early 2000s)
QuoteThe .scr file is intended for Eagle only: you would normally open Eagle, run the .scr script, and Eagle builds a schematic and board file for you.
Quote.v → .net + .scr → run in Eagle 4–6 → builds .sch automatically.
After rolling back to Eagle 5.6 and then Eagle 4.16 and then 4.08 still no joy also with error 'Device not found' for
Quote"ADD E2,5"
So the devices are in the netlister repository. There is a newer Eagle LBR (50 pin connector)
https://github.com/revaldinho/netlister/tree/master/Eagle%20Lib
But then these older LIB files:
https://github.com/revaldinho/netlister/tree/master/lib

So apparently LIB files are in Eagle 3.x versions (there are apparently 3.0 and 3.5 versions made) that can be converted to the Eagle 4.x+ format (LBR combined library file)

Questions are:
1. Does anyone have an early version of Eagle (3.x) available to convert these LIB to LBR so then I can load/convert into later Eagle and then Kicad?

2. Is there another method to convert these across to maintain the schematic integrity before I remove/update the Xilinx 9536 to 9536XL with the TQFP footprint? Thanks!

Powered by SMFPacks Menu Editor Mod