MiSTer FPGA Forum

Wanted to start a thread for Atari’s System 1 & 2 boards, since they are some of my favorites.

The games using the boards are:

System 1: (Hardware overview)

• Indiana Jones And The Temple Of Doom
• Marble Madness
• Peter Pack Rat
• Road Blasters
• Road Runner

System 2: (Hardware overview)

Similar to System 1, but with a medium-resolution CRT (512×384) — this is where most of my favorites are.

• 720°
• Accelerator
• APB
• Championship Sprint
• Paperboy
• Super Sprint

There is progress on the Atari Gauntlet board, see this thread — and the community has banded together to get its developer a DE-10 nano.

Are the System 1 & 2 boards similar to the Gauntlet boards? Would completion of this core help in any way towards these two? Or are they completely different architectures?

(Update: The author did say in the thread I linked that he’s started System 1 — GitHub)

The same author as the Gauntlet core has also started System 1:
https://github.com/d18c7db/atari_system1_fpga

Based on http://system16.com/hardware.php?id=766, it does seem like Gauntlet is very similar hardware. The only major difference I can see is the main CPU is running twice as fast as System 1 hardware.

Looks like Gauntlet is running on MiSTer now:

https://twitter.com/misteraddons/status ... 3688196096

http://mk64.com/ron/gauntlet.html

Friend has this mod done to his Gauntlet 2

Here's a playble .rbf of Indiana Jones (work in progress, sound has some issues), you need both atarisy1.zip and indytemp.zip ROMs available.

I just gave it a quick try and it works really well! I'm really looking foward to Roadblasters and Marble Madness.

Was a big fan of Road Runner!

Has there been any new progress on this core recently? I would love to play Marble Madness on my Mister

Marble Madness was iconic.

Looks like more games are working now — from the GitHub page:

Supported game cartridges

Marble Madness (1984)
Peter Pack Rat (1985)
Road Runner (1985)
Indiana Jones and the Temple of Doom (1985)
Relief Pitcher (1986) (unreleased prototype)
RoadBlasters (1987)

Core is obviously still a work in progress.

I've successfully compiled the rbf. It involved finding, converting and copying the correct ROM files, with some hand-editing of a few of the converted files. Also there was a line missing in the conversion script.

Road Blasters is not included in the conversion script, maybe its ROM files are not needed for the build.

The GitHub repository only has mra files for Indiana Jones and Peter Pack Rat. Peter Pack Rat works, but has grey bands above and below the sprites, looks like that might be a DMA problem.

The mra files contain parts that describe the board components. I have no idea how to create mra files for the other games.

I had a quick look at both existing MRAs which are quite similar (same board). It looks like mapping ROM to banks and so on. And when also taking a look at MAME's source code for atarisy1, one can find all mentioned ROMs again. Seems that creating an appropriate MRA is "only" like a logical sliding puzzle...

RealLarry wrote: ↑Tue Oct 18, 2022 11:36 am I had a quick look at both existing MRAs which are quite similar (same board). It looks like mapping ROM to banks and so on. And when also taking a look at MAME's source code for atarisy1, one can find all mentioned ROMs again. Seems that creating an appropriate MRA is "only" like a logical sliding puzzle...

It is a bit of a puzzle alright. I think I should be getting close with creating the mra file for Road Runner, but right now it only gives me a black screen.

LamerDeluxe wrote: ↑Tue Oct 18, 2022 8:46 pm It is a bit of a puzzle alright. I think I should be getting close with creating the mra file for Road Runner, but right now it only gives me a black screen.

Don't forget to set the correct slapstic (protection) for road runner, should be 6C

It looks as though the movable objects (sprites in atari speak) are quite complex, proms set the number of GFX banks, number of bits, colour masking etc. and I am not sure if any of that is currently implemented other than what is needed for Indy - trying to understand mames video code for starters!

macro wrote: ↑Tue Oct 18, 2022 9:37 pm

LamerDeluxe wrote: ↑Tue Oct 18, 2022 8:46 pm It is a bit of a puzzle alright. I think I should be getting close with creating the mra file for Road Runner, but right now it only gives me a black screen.

Don't forget to set the correct slapstic (protection) for road runner, should be 6C

Thanks for the reminder! I had those written down correctly and still forgot to set the right value in this mra file. Still a black screen unfortunately. I think I know what I've done wrong, but it is already getting late here.

It looks as though the movable objects (sprites in atari speak) are quite complex, proms set the number of GFX banks, number of bits, colour masking etc. and I am not sure if any of that is currently implemented other than what is needed for Indy - trying to understand mames video code for starters!

I'm also not sure the mentioned games are actually working in practice. Starting Peter Pack Rat on the previous core also gives a black screen.

ok, peter packrat odd graphics can be overcome with a change to the MRA file. It's still not 100% but looks much better!

it seems the graphics data is inverted on it's way through the system, so fills need to be 0xFF for unused roms.

LamerDeluxe wrote: ↑Tue Oct 18, 2022 9:57 pm I'm also not sure the mentioned games are actually working in practice. Starting Peter Pack Rat on the previous core also gives a black screen.

I built my own, has a couple of changes - see if that is any better!

gfx likely to be incorrect for some games as it seems to have gfx roms for the lower two S.L.A.G chips set to ignore the images, so any used by those are not going to work.

Does someone have the missing MRA files?

rcade wrote: ↑Tue Oct 18, 2022 10:27 pm Does someone have the missing MRA files?

I'm working on creating an mra file for Road Runner at the moment.

Looking at the cartridges there needs to be some more connections made for the S.L.A.G.s as Indy and Peter Packrat are the only games that use the top two chips. (code currently has lower chips connected to 0xFF for all addresses)

All of the other (released) games use at least 3.

Road Runner and Marble Madness use 0,1 and 2 (with 0 being the one nearest the cartridge edge connectors)

Road Blasters uses 0,2 and 3

Additionally, 32 bits of data are passed into the cartridge module, this is mapped to the top two chips as well.

I guess we need to find out the maximum size for graphics roms loaded for each of the 4 chips and set the addresses accordingly (and use fills on the other MRAs so that all are loaded for all of the games) plus find out how the video data is fetched into the other chips so it can be catered for somehow. (and hope we don't hit bandwidth problems as graphics data is currently stored in SDRAM)

macro wrote: ↑Wed Oct 19, 2022 4:44 pm Looking at the cartridges there needs to be some more connections made for the S.L.A.G.s as Indy and Peter Packrat are the only games that use the top two chips. (code currently has lower chips connected to 0xFF for all addresses)

All of the other (released) games use at least 3.

Road Runner and Marble Madness use 0,1 and 2 (with 0 being the one nearest the cartridge edge connectors)

Road Blasters uses 0,2 and 3

Additionally, 32 bits of data are passed into the cartridge module, this is mapped to the top two chips as well.

I guess we need to find out the maximum size for graphics roms loaded for each of the 4 chips and set the addresses accordingly (and use fills on the other MRAs so that all are loaded for all of the games) plus find out how the video data is fetched into the other chips so it can be catered for somehow. (and hope we don't hit bandwidth problems as graphics data is currently stored in SDRAM)

Ah, so they can't work at the moment. I didn't notice a difference with your build, what did you change?

Documentation on creating mra files is pretty basic. I've been trying to figure out the exact logic.
With interleaving, four banks of chips are interleaved into 32-bits. But Road Runner has six banks of chips (2 x 3, like you mention) I have now interleaved those six banks at 32 bits, because if I would expand the map bits with two extra ones, it would be 48 bits (which is probably not even possible).

For the correct order I'm looking at the order of the chips on the board and then using a scrambled order similar to that used in the other mra files.

With the Peter Pack Rat mra, one thing that is unclear to me is how the two <part repeat="32768">00</part> blocks relate to the 3 x 2 empty sockets at the top-right of the board. Each of the chips used there is 16KB. The address gap in the MAME source is 0x60000, or 0x10000 per empty socket (32KB), which makes more sense than the odd value used in the mra file.

SYS1 core is in the works...

macro wrote: ↑Tue Oct 18, 2022 10:19 pm ok, peter packrat odd graphics can be overcome with a change to the MRA file. It's still not 100% but looks much better!

it seems the graphics data is inverted on it's way through the system, so fills need to be 0xFF for unused roms.

Hello.
I have a Black Screen ,“ Peter pak rat„.Can you help me?

Hi all, sorry I've been on a bit of a break from MiSTer so System-1 hasn't seen much progress but I've now spent a couple of days and fixed the sprite issue with Peter Pack Rat. A new rbf is published and you will definitely need the updated mra files for Indy and Peter to go with it.

For Marble Madness to work I'd need some suggestions as to how the player inputs should be mapped (to what) as the original game uses trackballs so it's not controllable through joysticks unless some fakery can be implemented so a joystick emulates a trackball. Hmm more work required....

The other two games Road Runner and Road Blasters use significantly more ROMs, eg 68K ROM areas 0x50000-0x80000 which are currently not mapped and cannot be loaded from a rbf, as well as more TILE ROMs. Some substantial jiggery-pokery is required to get these done so don't waste your time trying to create a mra file thinking that's all that's needed.

Hi Alex. You could ask jimmystones how he handled track ball control in Missile Command.

Oh thanks for the pointer, I just quickly browsed the source for Missile Command, I'm gonna steal trackball.v from it.

Hi just tested the latest rbf and mra's Indy still has the bug where the whip input freezes the game if held down.
This has been the case for me in all but one of the mra-core combinations in the past.
p.s I'm using Mister on a LG C9 Freesync enabled, 1440p, integer scaling, lowest lag option selected.

Thanks.

Which mra didn't have this problem, was it one from my repository or was it provided/compiled by someone else?

Sorry I'll have to guess from memory, I do remember it was a month of may build, it obviously had worse audio reproduction though.

The new core doesn't crash the game, it just pauses until the button is let go, which means any quick whip attacks cause super fast pauses which cause deaths when attacking enemies..

I'm using a 8bit do 2.4ghz pad, fast USB polling enabled in mister, and same pauses happens on CRT or OLED.