Atom and Single Board retro-Computers

BigEd · Post by **BigEd** » Mon Feb 24, 2020 10:49 am

Maybe see this thread and the linked repo.
There are sources in an earlier revision, and notes about copyright, but that's gone - maybe moved. (Of course we all know that GPL or LGPL software should be shipped with source and the copyright notice should remain intact.)

hoglet · Post by **hoglet** » Mon Feb 24, 2020 12:01 pm

Thanks Ed,

Looks like it's based on a version of my (originally Alan D's) Atom core from ~2015 (before a big refactor to support multiple targets)

Dave

-B- · Post by **-B-** » Mon Feb 24, 2020 1:32 pm

hoglet wrote: ↑Mon Feb 24, 2020 9:55 am Is there any more information available on this core?

There is this link...
https://gitlab.com/victor.trucco/zx-spectrum-next-cores
It says that they are cores based on ports for ZX-Uno and have come from different sources, copyright has been kept on each file.

BigEd · Post by **BigEd** » Mon Feb 24, 2020 1:37 pm

Ah, indeed, I see now the sources are in a different place now:
https://gitlab.com/victor.trucco/zx-spe ... er/Sources

Post by **roland** » Mon Feb 24, 2020 8:51 pm

hoglet wrote: ↑Mon Feb 24, 2020 12:01 pm Looks like it's based on a version of my (originally Alan D's) Atom core from ~2015 (before a big refactor to support multiple targets)

Does that make you feel pissed or flattered?

Even my name is mentioned in the Godil part, so it's definitely yours:

Maybe we should point them to the newest releases with Real Time Clock, 64 colours, SAM and PAM etc...

hoglet · Post by **hoglet** » Mon Feb 24, 2020 9:17 pm

roland wrote: ↑Mon Feb 24, 2020 8:51 pm Does that make you feel pissed or flattered?

I'm very happy for people to re-use things I've worked on.

It would be nice to get a mention though!

Dave

gob33 · Post by **gob33** » Wed Jul 01, 2020 12:52 pm

Not an FPGA, but Geoff Graham and co had just released the new version 2 of Maximite Colour Computer.
The machine is always simple, very fast (270.000 lines of Basic interpreted in 1 sec). Kit or pre-assembled for 85 pounds.
There are some versions with direct integration of the ARM.
And Im wondering if an Atom (or BBC) port should be possible instead of MMBasic (which is very good) ?
There is something to do here, the machine is nice.

ARM Cortex-M7 32-bit RISC Processeur at 480 MHz / 2MB Flash /1MB RAM
VGA: max. 800X600 / 16-Bits (= 65536 couleurs)
5 MB Data RAM / 516 KB Basic programs
Build-in full screen code editor with colour coded text, unlimited line lengths and sophisticated search and replace.
Battery backed real time clock and calendar with software trimming.
USB Keyboard (wireless possible)
SD-card up to 128 GB (FAT16, FAT32 ou exFAT), build-in graphical file manager.
Video layers with selectable levels of transparency, multiple video pages with high speed copying, BLIT, SPRITE
Nunchuck Game Controller
Stereo Sound (WAV/FLAC/MP3/MOD)
28 I/O Pins (12 Analog IN)
Raspberry Pi HAT compatible.

Forum

Youtube demos

The Colour Maximite 2 by the 8-bit Guy

Debuglive's more in-depth video on the Color Maximite 2

About

The Colour Maximite 2 was conceived and developed by Peter Mather (matherp on the Back Shed Forum) who also led the development project.
It is based on the original Colour Maximite developed by Geoff Graham and uses the MMBasic interpreter written by Geoff Graham (http://geoffg.net).
A team of six people from around the world assisted with testing, advice and developing some initial games and programs. These are Phil Boyce, Jim Hiley, Graeme Rixon, Martin Round, Robert Severson and Mauro Xavier.

Support

Support questions should be placed on the Back Shed forum (http://www.thebackshed.com/forum/Microcontrollers) where there are many enthusiastic Maximite and Micromite users who would be only too happy to help. The developers of both the Colour Maximite 2 and MMBasic are also regulars on this forum..

Copyright and Acknowledgments

The Maximite firmware and MMBasic is Copyright 2011-2020 by Geoff Graham and Peter Mather 2016-2020. 1-Wire Support is Copyright 1999-2006 Dallas Semiconductor Corporation and 2012 Gerard Sexton. FatFs (SD Card) driver is copyright 2014, ChaN.
MOD file support was written by Jean François DEL NERO (hxcmod.c) WAV, MP3, and FLAC file support are copyright 2019 David Reid PNG support is copyright 2005-2010 Lode Vandevenne and 2010 Sean Middleditch
The editor and file manager are based on code copyright 2016 Salvatore Sanfilippo and documentation from paileyq@gmail.com
The turtle graphics support including the polygon fill algorithm are copyright Mike Lam 2015
Marcel Rodrigues wrote the GIF decoder. Maury Quijada wrote the image resize and image rotate code.
The compiled object code (the .bin file) for the Colour Maximite 2 is free software: you can use or redistribute it as you please. The source code is available via subscription (free of charge) to individuals for personal use or under a negotiated license for commercial use. In both cases go to http://mmbasic.com for details.
This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

Volhout · Post by **Volhout** » Wed Jul 01, 2020 1:07 pm

It is not FPGA based, it has a STM32H743 480MHz ARM based MCU.

gob33 · Post by **gob33** » Wed Jul 01, 2020 1:09 pm

Yes but I didnt find a post relevant to old Maximite or STM32. So i continue on my post about retro-computers.

gob33 · Post by **gob33** » Thu Jul 30, 2020 11:55 pm

awilliams · Post by **awilliams** » Tue Aug 18, 2020 9:20 am

> I am asking this could run RISC/OS ??

I think it would be really hard pushed to do so.

When I saw that published in Silicon chips a couple of months back I wondered why they hadn't used one of the Raspberry Pi compute modules. If you really wanted to run RISC OS it would probably be quicker to build a CM3 adaptor than try porting the code. You would be missing the ADC converters though, and mapping the pi video to their VGA socket might need another chip & you might need to leave a lot of the video associated resistors off the main board.

I relatively quickly decided not to bother personally but I entertained the idea for a while.
Alan

gob33 · Post by **gob33** » Tue Aug 18, 2020 1:21 pm

Another option would be to port Richard T. Russel BBC Basic in replacement of the MMBasic (which is also powerfull).
STMicro has HAL libraries for developpement. I think there is even SDL somewhere in their graphical library called "TouchGFX".
The STM32H743IIT6 could be used up to 1024x768, it can interface a lot as in the OpenH743I-C dev board from Waveshare.
The advantage over a Raspberry Pi is all is integrated in the chip and you need few peripheral circuitry.

gob33 · Post by **gob33** » Sat Oct 10, 2020 11:27 pm

New polish version (Deluxe Edition) of the Color Maximite 2:

The Color Maximite 2 Deluxe Edition is an enhanced version of the Geoff design. Produced in Poland.
It is 100% compatible with it and has been extended with additional functions.

Main features:

built-in ARM Cortex-M7 32-bit RISC processor clocked at 400MHz (the processor is an integral part of the device, no additional development board with the processor is required)

3 sockets to connect 3 Wii nunchuk controllers (as opposed to 1 socket in the original design)

Socket for connecting a standard joystick, e.g. Atari (the prototype does not have such a socket)

Reset button (the prototype did not have it, which made programming difficult)

40-pin internal socket compatible with the Raspberry Pi extension which allows the module to be installed inside the housing (the original version did not have such an extension)

internal built-in prototype board (not included in the original version)

40 pin external socket

USB socket for keyboard connection

CPU power and programming socket

IR sensor

Temperature sensor

Audio jack

VGA socket

Switch

SD card slot

LEDs indicating power and activity

Built-in battery for clock settings

The kit includes:
Complete assembled and tested device with panels and casing
USB power cable

https://youtu.be/u1dHAOTn72A

davewoo999 · Post by **davewoo999** » Fri Dec 18, 2020 8:23 am

A while ago I ported the ICE40 atom fpga to the DE10-nano used in the MiSTer setup, It was a basic port and I thought I would update this, I wanted to use the mmc instead of the sddos but failed to make it work.
I then took the Atomfpga core and stripped it down to basics working with mmc and then added 32k ram, 7 additional roms with one slot ram for download, two character sets (lower case), tape in/out, sound - atom,sid,tape and off, atom or bbc basic, coulour palette and four keyboard selections, UK US original and game.
I thought I had everything working ok but on my last tests I found that Asteroids has a problem where the rocks get stuck in the top left corner, or they just move in diagonal lines across the screen.
I had noticed this before when running the 65c02 and 65816 cpu's and thought this was because of the cpu differences but perhaps it was highlighting a problem with my code.
I presume it is a timing issue and before I go back to the drawing board - Has anyone come across a similar problem?
Thanks Dave

hoglet · Post by **hoglet** » Fri Dec 18, 2020 9:56 am

Hi Dave,

davewoo999 wrote: ↑Fri Dec 18, 2020 8:23 am I presume it is a timing issue and before I go back to the drawing board - Has anyone come across a similar problem?

I just tried Asteroids on Atom 2K18 (which is uses the latest AtomFpga with the 65C02 core written by Alan D). I also tested it on a real Atom with a 65C02 fitted. It seems to work fine on both.

Some questions:
- Is the problem with Acornsoft Asteroids? (there are other similar games)
- Are you using the version from the Atom Software Archive?
- Does it matter whether you choose the joystick version or not?
- Is the problem immediately obvious as soon as you start playing?
- Which 65C02 core are you using?
- Which 65C816 cord are you using?

In my experience, the main source of timing issues is an incorrectly timed write enable signal to the external SRAM.

Is your latest project up on github, or available elsewhere for me to take a quick look at?

I did find this, but it's over a year old.

Dave

davewoo999 · Post by **davewoo999** » Fri Dec 18, 2020 10:56 am

Thank you for your quick reply.
65c02 from AlanD and the 65c816 from the super nintendo port on MiSTer.
I just tried the JS Asteroids and it has the same problem.
I have re-written the logic for RAM/ROM access to be closer to your original and am now recompiling.
I have not placed it on Github yet and was hoping to put it on the Main MiSTer site if that is ok with you.
Can I also point them to the Software Archive?

Thanks Dave

hoglet · Post by **hoglet** » Fri Dec 18, 2020 10:58 am

davewoo999 wrote: ↑Fri Dec 18, 2020 10:56 am I have not placed it on Github yet and was hoping to put it on the Main MiSTer site if that is ok with you.
Can I also point them to the Software Archive?

That's all fine with me.

gob33 · Post by **gob33** » Sun Dec 20, 2020 1:10 am

3D printed modulable case for Mister FPGA or RPi 3 by Lee "Legion" Smith (nottingham).
The most approaching case which ressembles an Atom and that can be build on a small personnal 220mm wide 3D printer.
(but with the new 45° Crealty 3D printer, lengths can be very long)

Your custom colored USB TKL keyboard goes directly blocked inside the case (size max width 360mm and depth 135mm
screw size m3 self tappers 20mm and 8mm).
You can also use a FULL SIZE keyboard by mounting inserts in the middle and approach an Archimedes / Pi 4B / Risc OS.

Youtube / Channel
Twitter
Files for 3D print

For best results, read: Teaching Tech 3D Printer Calibration

NOTE 1: It is more economical and less complicated to put the Mister FPGA into one of its cases (personalized with stickers or not) and to add a custom separated (WASD) keyboard labeled with Acorn logo.

NOTE 2: See also similar AjSTer: a Mister case for A-JAZZ AK33 keyboard

davewoo999 · Post by **davewoo999** » Mon Dec 21, 2020 11:19 am

HELP
I have tried several variants of my code and need to go back to basics.
It is unusual as my error does not affect any other programs apart from Asteroids.
As it takes around 20 minutes to recompile each change so I want top strip down the code to bare minimum but still need the Asteroids program for testing.
Am I correct in thinking that I can take the 3k Asteroids program and place it in memory and Link to it.
I believe that Invaders and Galaxians was previously done this way.
What address do I need to load it to and what is the link address, I have #2800 in my mind but not sure why.
My last code changes made all the boulders disappear until you fired into the corner and they appeared but either moved horizontal or vertically.

Thanks Dave

hoglet · Post by **hoglet** » Mon Dec 21, 2020 11:25 am

davewoo999 wrote: ↑Mon Dec 21, 2020 11:19 am What address do I need to load it to and what is the link address, I have #2800 in my mind but not sure why.

You need to load it to #2900. Then to start it you would enter:

Code: Select all

OLD
RUN

davewoo999 · Post by **davewoo999** » Mon Dec 21, 2020 11:46 am

You need to load it to #2900. Then to start it you would enter:
Code: Select all
OLD
RUN

Many thanks

hoglet · Post by **hoglet** » Mon Dec 21, 2020 11:53 am

Dave,

I wonder if your issue is down to RAM initialization?

Asteroids uses the Atom BASIC random number generator, whose seed is in locations 08-0C. Atom BASIC relies on this getting a random value on power up, which happens if this is external RAM.

I've just done a couple of tests in Atomulator.
- If 08-0C are initialized to 0x00 then you end up with no Asteroids at all.
- If 08-0C are initialized to 0xFF then you with an Asteroid stuck in the top left corner.

Is it possible you are hitting this issue?

Try doing this before running Asteroids:

Code: Select all

!8=#12345678
?12=#90

Dave

davewoo999 · Post by **davewoo999** » Mon Dec 21, 2020 1:18 pm

I wonder if your issue is down to RAM initialization?

Even though RAM initialization has been discussed on this forum before I failed to take that into account.
Its exactly the problem I have had.

Many many thanks.

Now just a sticky shift key to fix in my keyboard code, which is the reason I started this as I could never remember to press shift8 for the open brackets etc - Its so much easier for an old brain like mine to press the key with character you can see.

Dave

hoglet · Post by **hoglet** » Mon Dec 21, 2020 1:38 pm

davewoo999 wrote: ↑Mon Dec 21, 2020 1:18 pm Even though RAM initialization has been discussed on this forum before I failed to take that into account.
Its exactly the problem I have had.

Many many thanks.

Excellent stuff.

It took a while for me to think of this as well!

Dave

TheCorfiot · Post by **TheCorfiot** » Thu Dec 24, 2020 8:45 pm

I have the latest core downloaded to my MiSTer FPGA system, created the AcornAtom folder in the Games directory and proceeded to create a boot.vhd file containing the archive but for the life of me cant get it to run. keep getting the Not Ready error..

Does anyone have a compiled vhd ready or can help me create a suitable one please

davewoo999 · Post by **davewoo999** » Fri Dec 25, 2020 8:29 am

Does anyone have a compiled vhd ready or can help me create a suitable one please

There is now a blank.zip file in the releases folder.
Unzip - rename to boot.vhd and try that one.

Dave

TheCorfiot · Post by **TheCorfiot** » Fri Dec 25, 2020 9:05 am

Thanks Dave

I uploaded the file renamed as boot.vhd firstly as is with no files added.
Still getting the Not ready error when trying to *CAT it

Used WinImage to inject the Atom Software archive but still getting Not ready error

I'm assuming with the new release that the VHD file goes into fat/games/Acornatom on the main SD

confused here lol

hoglet · Post by **hoglet** » Fri Dec 25, 2020 11:49 am

The Not Ready error from AtoMMC occurs when the AtoMMC Firmware (running on the embedded AVR/PIC microcontroller) has been unable to initialize the SD Card. I'm not sure how this works in Mister, but in a real AtoMMC this usually occurs if there is a fault in the SD Card interface, or the SD Card is missing completely.

This likely won't resolve the Not Ready error, but here's a link to a boot.vhd file that contain V11.01 of the archive:
https://www.dropbox.com/s/4y3eailpagftq ... d.zip?dl=0

Dave

TheCorfiot · Post by **TheCorfiot** » Fri Dec 25, 2020 1:45 pm

Thanks Dave I will give it a go

Merry Christmas

TheCorfiot · Post by **TheCorfiot** » Fri Dec 25, 2020 5:37 pm

hoglet wrote: ↑Fri Dec 25, 2020 11:49 am The Not Ready error from AtoMMC occurs when the AtoMMC Firmware (running on the embedded AVR/PIC microcontroller) has been unable to initialize the SD Card. I'm not sure how this works in Mister, but in a real AtoMMC this usually occurs if there is a fault in the SD Card interface, or the SD Card is missing completely.

This likely won't resolve the Not Ready error, but here's a link to a boot.vhd file that contain V11.01 of the archive:
https://www.dropbox.com/s/4y3eailpagftq ... d.zip?dl=0

Dave

You Guessed it, "Not Ready" lol

stardot.org.uk

Atom and Single Board retro-Computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers

Re: Atom and FPGA based retro-computers