Potential new hardware

[foft]

@GlowingGhoul: Left channel audio is normal. The right channel is the 2nd pokey. I agree it would be convenient when only 1 pokey is used to send it to both channels.

 

@Vanfanel: Saving config changes will be done. However I want to say OSA is only required for about 0.1% of games. If you need it for more then you are either very unlucky or doing something wrong! Are you holding option on boot? e.g. hold f8 when pressing f10?

 

Anyway will get to these changes when I finish off a few 5200 bits...

 

There is also another exciting feature in the next build - I'll let Hias announce that though:-)

[+JAC!]

>However I want to say OSA is only required for about 0.1% of games.

Even those should be patched. There are separate threads here on AA for exactly that.

[foft]

New build up. The main feature is that it has the fantastic Turbo Freezer built in. Thanks Hias!

 

http://www.horus.com/~hias/freezer/turbo-freezer-2011/FreezerManualEn-3.10-130330.pdf

 

To enable it put freezer.rom with the other roms on the sd, then press scroll lock or delete.

 

(Don't use the cart emu option - this part is not included because we already have cartridge emulation via the menu.)

 

No keyboard support in firmware yet - it's not forgotten!

 

edit: Requires latest mist firmware

Edited October 30, 2014 by foft

[eeun]

Just tried the latest 800 core (VGA, non-cs) for MiST and it boots to a black screen. Pressing F12 displays a gray screen with some garbled color at the top.

I'm running the latest 141024 MiST firmware.

Previous 141006 core with same SD card, etc. works fine.

 

Can anyone else report success or issues with the latest 800 core?

[harbaum]

Just tried the latest 800 core (VGA, non-cs) for MiST and it boots to a black screen. Pressing F12 displays a gray screen with some garbled color at the top.

I'm running the latest 141024 MiST firmware.

Previous 141006 core with same SD card, etc. works fine.

 

Can anyone else report success or issues with the latest 800 core?

Garbled screen with f12 imho means the rom wasn't loaded. Are you sure the roms are all in place?

[eeun]

The roms are in the correct location. I've tried it with two SD cards now. On both, the 1006 core will work, but I just get black screen and the garbled F12 on the new core.

[greblus]

I can confirm this as well. Latest firmware and latest atari core starts to black screen and i see garbled screen when F12 is pressed.

It happens when I'm changeing the core from st core. When it's started from zx spectrum core once i got to atari basic but no files were available from the file selector. A few times it also boots directly to self-test. The freezer.rom i extracted from freezer2011.atr. Havent checked yet if it's not because of the freezer (maybe it's a wrong rom, but currently my tv-set is occupied by wife , will check it later)

[harbaum]

I can confirm this and i think i am to blame. The latest builds use some simulated sd card. This is due to the fact that the sd card is not directly connected to the FPGA in the MiST. In previous cores foft used some nasty (but obviously more reliable) hack to access the sd card. In the latest version he uses my fake sd card. This seems to be unstable from build to build.

 

The *_RGB cores e.g. do atually work. So does the 5200 VGA core.

 

I'll have a close look at this. Perhaps foft can trigger a rebuild for the two VGA cores until he gets working versions.

 

This may be a little difficult to understand. FPGA cores are not always built or "synthesized" in exact the same way. There's some randomness involved when the computer converts the human written FPGA description into the FPGA core you can finally download. Some logic may be wired in slightly different ways from build to build. This usually doesn't harm. But in case there's something implemented in a not so clever way this tiny differences can cause a core to work or fail.

 

Consider e.g. you generate two signals inside the FPGA in one place to change at exactly the same time and compare their state at another place inside the FPGA. The wiring inside the FPGA may be slightly different from build to build causing one of the two connections for the two signals just to be a tiny little bit shorter than the other one. If you now rely on one signal to change before the other one you have a 50% chance that this actually works since the signal you expect to arrive later actually was built using a longer wire inside the FPGA. I probably did something like that in my SD card implementation foft is using ... and now i need to do some bug hunting.

Edited October 30, 2014 by harbaum

[eeun]

That mostly makes sense And an interesting insight into FPGA workings.

 

Greblus, thanks for confirming I'm not crazy.

[greblus]

No worries, interesting people are always at least slightly crazy ;P

 

BTW I'm really looking forward to test turbo freezer. Foft, you wrote that there is no keyboard support yet, does it mean that saving to disk won't work?

[foft]

I meant navigating the 'f12' menu by keyboard.

[greblus]

Hi foft. There is some progress, the new core is booting now to Atari basic, but file selector under F11 is not showing any files and after pressing del or scroll lock nothing is showing up. And booting takes 4-5 seconds longer.

[foft]

New build up, should all be working now.

 

I'm not sure where to find freezer.rom (Hias?), for now I put it on my server under autobuild.

Edited October 31, 2014 by foft

[greblus]

Great! (I've freezer rom from your ftp and also the one extracted from freezer atr).

 

Thank you again for your hard work! And big thanks also to Hias for porting freezer to fpga.

 

Update: it works! And boot time returned to normal. Freezer also works (well, sort of). I can activate freezer menu, but could not find a way to save to disk (mounted an empty image as D1, but freezer can't save due to i/o error). Another strange behavior: after turning off the freezer's menu F11-F12 no longer work.

Edited October 31, 2014 by greblus

[NML32]

Update: it works! And boot time returned to normal. Freezer also works (well, sort of). I can activate freezer menu, but could not find a way to save to disk (mounted an empty image as D1, but freezer can't save due to i/o error). Another strange behavior: after turning off the freezer's menu F11-F12 no longer work.

 

I too get an error when trying to save to disk. All the other save options work but are lost after shutting down the FPGA. I also tried to save to a blank disk mounted to D2 and also tried saving to the games's atr file.

I didn't have any issues with F11 or F12 after leaving the Freezer menu. I tried this on my MIST and MCC-216.

[harbaum]

Ah yes, SD card writing ... i think i also need to look at that ...

[Farb]

foft, the new NTSC color palette looks great!

 

BTW, in the release notes it says:

 

MIST:2600daptor or analog joystick paddle support

 

Is the 2600daptor required to use paddles? Sorry, the "or" in that statement is confusing me

[foft]

The paddles will not work on the built-in db9's. So to play paddle games an analog input device over USB is needed. Such as paddles on the 2600daptor. An analog joystick should also work.

[vanfanel]

@foft: the compatibility problems I was having where entirely my fault. Pressing F8 when pressing F10 to reboot did the trick as you said!

[Farb]

The paddles will not work on the built-in db9's. So to play paddle games an analog input device over USB is needed. Such as paddles on the 2600daptor. An analog joystick should also work.

 

I got my 200daptor but it doesn't seem to work correctly on my Mist. The left/right is reversed and the button doesn't work. I tested the adapter and paddles using Stella on my Mac and it worked fine. This is using the 20141104 build.

 

Is anyone else seeing this problem?

Edited November 7, 2014 by Farb

[foft]

Hah! Yes you're right, will get this sorted out...

 

I'll change the direction in the next core build. It will need a new firmware to map the buttons. My core receives them as fire buttons while they should be on the joystick directions.

Edited November 7, 2014 by foft

[Farb]

Hah! Yes you're right, will get this sorted out...

 

I'll change the direction in the next core build. It will need a new firmware to map the buttons. My core receives them as fire buttons while they should be on the joystick directions.

 

Thanks for verifying, foft! I'm glad it wasn't just me. Looking forward to getting some Kaboom in with real paddles

[foft]

I've got this fixed. The cores will be online in a few hours.

 

To switch to paddle mode press 'left windows key'. The reason this is needed is that the 2600daptor maps the paddle fire buttons to fire (on the 9-pin connector the fire buttons are mapped to the joystick directions).

 

Another couple of new features on the mist:

i) Left alt key to toggle fake 'scan lines'.

ii) Arrow keys, escape and enter can be used to navigate the menu.

 

Does anyone want to volunteer to write/maintain a manual for these cores? :-)

Edited November 8, 2014 by foft

[Farb]

I've got this fixed. The cores will be online in a few hours.

 

Thank you, foft! I can confirm that the paddle button now works (when the left windows key is pressed) and the direction is correct.

 

However, I am seeing some "jerkiness" with the paddles -- the buckets in Kaboom don't smoothly follow paddle movement. They even wiggle sometimes when the paddles are idle. I didn't see this behavior in Stella.

 

It also appears that keyboard navigation doesn't work when I have both a USB keyboard and the USB 2600daptor plugged in simultaneously. Is that expected?

 

On another note, I might be interested in writing a manual for the cores. Shoot me a PM so we can discuss.

[foft]

Re menu: In 'normal mode' the mist mirrors the analog stick to the digital stick. So the paddle position can be used to control the menu (painful)! The keys work if you eg stop holding up/down/right/left. Press windows key to toggle ''paddle mode' then the menu will work.