Atari 7800 / 2600

[LamerDeluxe]

Paddles do not work via snac, they are analog.

Paddles do work with the Atari DB9 snac, but you have to connect a cable from its jack connector to that of an ADC board.

Not in the 2600/7800 core.

Would be great if it did support them.

[thorr]

Agreed, but I already beat that horse to death and managed to kill it again and unfortunately the answer is no. I keep a copy of the unofficial 2600 SNAC core around for paddle games if I want the authentic experience, or I do the rigamarole with my 2600daptor to get one paddle working with this core. I also built my own paddle for this core with a spinner and it is very very nice, but not authentic.

I haven't tried the driving controller lately with this core, but it would be great if the sensitivity with a third party spinner could be dialed back tremendously to match the low resolution sensitivity of the real driving controller. The driving controller has 16 transitions per revolution. I am not sure if this can be adjusted in the MiSTer.ini.

[LamerDeluxe]

Agreed, but I already beat that horse to death and managed to kill it again and unfortunately the answer is no. I keep a copy of the unofficial 2600 SNAC core around for paddle games if I want the authentic experience, or I do the rigamarole with my 2600daptor to get one paddle working with this core. I also built my own paddle for this core with a spinner and it is very very nice, but not authentic.

I haven't tried the driving controller lately with this core, but it would be great if the sensitivity with a third party spinner could be dialed back tremendously to match the low resolution sensitivity of the real driving controller. The driving controller has 16 transitions per revolution. I am not sure if this can be adjusted in the MiSTer.ini.

I find it odd that there seems to be so little paddle support in general, both on the hardware and software side. I often used paddle controllers on my 2600 and VIC-20.

I agree that a spinner sensitivity setting would be useful, if there isn't one yet.

[thorr]

I find it odd that there seems to be so little paddle support in general, both on the hardware and software side. I often used paddle controllers on my 2600 and VIC-20.

In my humble opinion, there are two main reasons there are issues with paddle support.

The first reason is because the core treats each individual paddle as a single player. IMO, this is incorrect. Each set of two paddles should be treated as a single player. There is an X and Y axis, and two buttons. The set of paddles plug into one port on the Atari. A second set of paddles can be plugged into the second port on the Atari, which would be player two. This one change would solve most of the issues, even if there continues to be no SNAC support.

The second reason is because the core is designed with a spinner in mind. Paddles are analog joysticks that have ranges, not spinners. Spinners are infinite. If the MiSTer detects an analog joystick with X and Y and two buttons, it should treat this accordingly as one pair of paddles plugged into the first port on the Atari. Natively, the Atari 2600 only uses a portion of the analog range in games, so writing MiSTer code to make use of the whole range is incorrect. When playing a paddle game, you would move the "joystick" only a part of the way back and forth to move the object in the game the full range, and that part of the range that gets used varies from game to game. Moving past the usable area is a deadzone and you would need to return back to the edge of the usable range to begin moving the object in the game again. If it was done this way, it would be correct and authentic.

I think that spinner support should remain in the core (the way it is now), and when spinners are used, then separate everything out so more spinners can be connected and control all the paddles. However, when an analog joystick is used, it should be treated as described in the paragraph above as two paddles.

In summary, just add support for a single analog joystick as a set of two paddles and all will be good. As a side note, I tried creating the MiSTer spinner paddles adapter, and it has the problem that there is some overlap where the left paddle can control the right paddle slightly and vice versa. In other words, it is not an acceptable alternative.

[HerrBerzerk]

That is right, the Indy500 controllers are basically spinners, but I did also not get them working with the 7800 (2600) core. Maybe the problem is, that the paddles are analog devices and not digital (though the schematics for the paddle is not very complicated)? Like we have problems with analog sound in many arcade cores? I heard it is hard to program analog circuits in FPGA?

But if I understand correctly then, at the moment we have no original Atari hardware support for the 2600 part of the 7800 core? Neither original joysticks nor paddles nor driving controllers.

Does anybody know if the paddles are working on the commodore 64 core? If I remember correctly, in the past, I was able to use the atari paddles on my c64...

Thank you.

[thorr]

But if I understand correctly then, at the moment we have no original Atari hardware support for the 2600 part of the 7800 core? Neither original joysticks nor paddles nor driving controllers.

There is support for original Atari hardware. From what i recall SNAC works with Atari 2600 joysticks, but not Atari 7800 joysticks (2 buttons), paddles, the driving controller or the keypad. If you get a 2600daptor D9, it works with Atari 2600 joysticks, Atari 7800 joysticks, and the keypad if you map the buttons to keys in the MiSTer core menu, and it works with one Atari Paddle, not two, but you have to go through a rigamarole to get the one working. I don't remember if I got the driving controller to work at all. Analog joysticks are supported in the MiSTer, but the Atari core doesn't map them to two paddles, and it uses the full range in the game, not a portion, so you have to turn the dial much further than you should normally have to. Right now the best way to play paddle games in this core IMO is on non-original spinner hardware and adjusting the spinner sensitivity to what feels correct. If you want to play on original paddles, use the Atari 2600 SNAC core linked earlier in this thread somewhere but you will also need an analog ADC audio cable and the Atari SNAC adapter.

[dmckean]

I'd like to see someone develop an open hardware super low latency usb potentiometer based paddle that anyone could build. Something that could work well with any MiSTer core that supports paddles.

It seems like currently that using original paddle controllers with adapters has a bunch of issues and buying two spinner based controls is prohibitively expensive and not authentic.

[HerrBerzerk]

There is support for original Atari hardware. From what i recall SNAC works with Atari 2600 joysticks, but not Atari 7800 joysticks (2 buttons), paddles, the driving controller or the keypad. If you get a 2600daptor D9, it works with Atari 2600 joysticks, Atari 7800 joysticks, and the keypad if you map the buttons to keys in the MiSTer core menu, and it works with one Atari Paddle, not two,

I have the analog board from Antonio Villena with the DB9 Adapter and I can only speak for this configuration. In my tests, the original Atari joystick is working fine in the 7800 core, but as soon as you load a 2600 game, the joystick stops working. I was not able to get paddles working at all, not one nor two, with e.g. Breakout for the 2600. Also not driving controllers. I did not test the keypad. Maybe that's a problem with the Villena config then.

Maybe the above linked older snac core is working better, I will try, thank you.

[thorr]

I'd like to see someone develop an open hardware super low latency usb potentiometer based paddle that anyone could build. Something that could work well with any MiSTer core that supports paddles.

It seems like currently that using original paddle controllers with adapters has a bunch of issues and buying two spinner based controls is prohibitively expensive and not authentic.

There is already MiSTer Spinner support for potentiometer paddles as you described, and it sets up each paddle as its own player, but there are two issues:
- The two paddles interact with each other, so you can steer the other player a little bit. This may be fixable in the Arduino code, but I haven't tried.
- The coding in this core uses the full range of the potentiometer so you have to turn it way farther than a real Atari paddle. This second reason is the reason I didn't bother fixing the first issue because I wouldn't want to play that way.

[thorr]

I have the analog board from Antonio Villena with the DB9 Adapter and I can only speak for this configuration. In my tests, the original Atari joystick is working fine in the 7800 core, but as soon as you load a 2600 game, the joystick stops working. I was not able to get paddles working at all, not one nor two, with e.g. Breakout for the 2600. Also not driving controllers. I did not test the keypad. Maybe that's a problem with the Villena config then.

Maybe the above linked older snac core is working better, I will try, thank you.

I am pretty sure you won't be able to use it with paddles because of the ADC cable requirement. If there is a USB-style SNAC port, you could get an Atari SNAC adapter with its own DB9 and audio jack and use that.

As far as the joystick stops working issue, there is probably a setting in the core menu that is not right. Try playing with the various settings in the core menu and it might work for you. I don't have it in front of me right now, so I can't tell you what settings, but it probably needs to be SNAC and maybe other things.

[LamerDeluxe]

I find it odd that there seems to be so little paddle support in general, both on the hardware and software side. I often used paddle controllers on my 2600 and VIC-20.

In my humble opinion, there are two main reasons there are issues with paddle support.

The first reason is because the core treats each individual paddle as a single player. IMO, this is incorrect. Each set of two paddles should be treated as a single player. There is an X and Y axis, and two buttons. The set of paddles plug into one port on the Atari. A second set of paddles can be plugged into the second port on the Atari, which would be player two. This one change would solve most of the issues, even if there continues to be no SNAC support.

The second reason is because the core is designed with a spinner in mind. Paddles are analog joysticks that have ranges, not spinners. Spinners are infinite. If the MiSTer detects an analog joystick with X and Y and two buttons, it should treat this accordingly as one pair of paddles plugged into the first port on the Atari. Natively, the Atari 2600 only uses a portion of the analog range in games, so writing MiSTer code to make use of the whole range is incorrect. When playing a paddle game, you would move the "joystick" only a part of the way back and forth to move the object in the game the full range, and that part of the range that gets used varies from game to game. Moving past the usable area is a deadzone and you would need to return back to the edge of the usable range to begin moving the object in the game again. If it was done this way, it would be correct and authentic.

I agree with both points. The MiSTer system should present original paddle controllers data in such a way that a core can implement them exactly like it worked on the original hardware. On the Atari 2600 the first set of paddles was used for player 1 and 2 and the second set player 3 and 4. IMO preserving a system means preserving the ways they could be controlled as well.

I think that spinner support should remain in the core (the way it is now), and when spinners are used, then separate everything out so more spinners can be connected and control all the paddles. However, when an analog joystick is used, it should be treated as described in the paragraph above as two paddles.

I think this is the whole reason the paddles were split into two players by the system, to allow for two regular analog controllers to control player one and two. Analog stick controllers, dual shock / Xbox style, work really badly as paddle controllers though, IMO, because of the centering spring and much lower precision.

In summary, just add support for a single analog joystick as a set of two paddles and all will be good. As a side note, I tried creating the MiSTer spinner paddles adapter, and it has the problem that there is some overlap where the left paddle can control the right paddle slightly and vice versa. In other words, it is not an acceptable alternative.

It could be added as an ini file option. I looked at the MiSTer spinner paddles adapter, but don't want to ruin a set of paddles for it. Disappointing to hear that it doesn't really work well.

I have been planning to create the JogCon one, as it allows for spinner and paddle mode, but I would prefer it as a separate adapter, instead of modifying the JogCon itself. It turns out it is a bit difficult to find a good housing for it. And then still, the JogCon is not very comfortable to operate.

[thorr]

Here is a link to the post with a picture of the paddle I made. It works great with this core, but it is only one paddle and not authentic. It's better than the original by far, but I prefer the original for that original feel. viewtopic.php?p=39990#p39990 I just ordered the Tron knob for it so it will look better: https://www.amazon.com/ThunderStick-Tro ... B097RBJN8J

[Chilli_Vibes]

Just a thought but, wouldn't it be a better option to split the core into two cores, 2600, and 7800 ?
Or, back-port any progress in this core, to the older core with working SNAC + Paddles ?
That way, all future fixes, and options can be focused in the 2600 core, without affecting the 7800 part.
At the moment, this core doesn't offer me anything over a real 2600, except for convenience. Playing games on a real 2600 + harmony cart is still far the better option than the frustrations of the 2600 part of this core.

[thorr]

I respectfully disagree. This core is absolutely stunning, except for original paddle support. I also have a real Atari 2600 + harmony cart + Svideo mod, but it does not look nearly as nice as this core does using component. It's the same as pretty much any other console. It just looks better/cleaner on the MiSTer. Despite all of my vocal ideas about paddle issues, I absolutely love this core. As far as separating the 2600 part, there really is no point since the 7800 is basically just additional functionality tacked onto the 2600 part.

[Chilli_Vibes]

Looking cleaner doesn't bother me, I grew up with RF, Composite, and SCART. I even had a black and white CRT TV 12" for my ZX81, and that looked crap.
What bothers me, are games that were meant for a specific controller just suck playing on anything else, especially something modern. This core is good, but it's more good for 7800, a lot of 2600 stuff still isn't in the core that really should be there.
Accepting anything less than what real hardware was capable of, just goes against the whole preservation concept, and for now I am sticking with the real thing.

[thorr]

I hear you and agree, but there is nothing more that can be done about it unless the developer changes her mind and gives us proper paddle support. For non-paddle games though, it rocks! Paddle games with a nice spinner also rock, but are a different experience that is nice to have only sometimes when in the mood for the upgraded, non-authentic experience.

[Stinky]

You could always use the old 2600 core for paddles. It wasn't perfect but it was OK.

[HerrBerzerk]

For non-paddle games though, it rocks! Paddle games with a nice spinner also rock.

Can you or someone please explain or make a video playing a 2600 game in the 7800 core with the original Atari Joystick over SNAC? I tried again and I don't get it working. In contrary, the Joystick is working in the menu and in the 7800 part but stops working in the 2600 part.

Thank you.

[thorr]

I am not sure about it working in the menu. With SNAC, it can't usually control the menu. Make sure it is in SNAC mode and not USB emulation or something. SNAC working in the menu may be a special feature of the Antonio Villena version. When my menu is open, it moves the Atari player when moving the stick.

The settings that work for me are:
In Peripherals...
Swap Joysticks No. (Try switching this one if it doesn't work)
Port1 Input SNAC
Port2 Input Auto
SNAC Analog No
Sega No
Swap Paddle No

[Lightwave]

I hear you and agree, but there is nothing more that can be done about it unless the developer changes her mind and gives us proper paddle support. For non-paddle games though, it rocks! Paddle games with a nice spinner also rock, but are a different experience that is nice to have only sometimes when in the mood for the upgraded, non-authentic experience.

If the creator of the 2600daptor d9 can't provide a firmware that splits the 2 paddles, then perhaps another alternative would be someone creating an arduino adapter similar to the d9 that presents the 2 paddles as separate devices. That way no destruction of donor controllers is required, and we could have something close to the original experience.

[skooter]

He can. He just doesn't want to.

[thorr]

I wouldn't say that is true at all. I emailed with him months ago and he said he was planning to get a MiSTer and see what he could do. I don't know if he ever got one. I am still waiting for mine from December from Mouser.

[skooter]

I also emailed him. He said all the switches combinations are already used and another alternative firmware would be another thing for him to maintain.

He wouldn't actually need a MiSTer to implement the changes. It is just a matter of presenting the two paddles as two different controllers to the system instead of a single one.

I think this is the right way to do it. They are two controllers intended for different players. The choice of Atari to make them attached and connected to a single port doesn't change that.

[HerrBerzerk]

I am not sure about it working in the menu. With SNAC, it can't usually control the menu. Make sure it is in SNAC mode and not USB emulation or something. SNAC working in the menu may be a special feature of the Antonio Villena version. When my menu is open, it moves the Atari player when moving the stick.

Yes, I got it working now, no idea what went wrong before, thank you.

My results are:

7800/2600 Core: One Joystick, One Spinner Controller/Indy500 Controller, no paddles.
old 2600 Core: Twp Joysticks, Two Spinner/Indy500 Controllers, no paddles.

I have one technical question. Maybe someone can answer. I have no knowledge about FPGA, I would just like to understand, out of pure interest, what the challenge is, to implement the Paddle controllers? When I look at the schematics, that is just a Potentiometer and a Capacitor scanned by the TIA. So is the TIA not yet fully build in FPGA or is it so hard to build analog circuits?

Thank you very much,

Frank

[thorr]

There is nothing particularly hard about it. We apparently just have a case of two immovable forces with both sides not budging to make this work. I have done everything I can to make it work short of learning to code this stuff myself. There are currently only two solutions with this core. Get multiple 2600daptor D9's and extra paddles and only use one paddle per 2600daptor and ignore the second one in each pair (or just use one paddle), or, build multiple spinner controllers and enjoy them and ignore the fact that they are not original hardware.

I suppose I could develop my own Arduino solution, and it probably would not be that hard. I would code it so that it would only use a portion of the paddles so it feels right. Maybe I will just do this and solve this for everyone (assuming they have the skills to solder and work with Arduinos), time permitting... If I get real ambitious, maybe I will make it with two DB9's and have it detect and work with other Atari controllers besides paddles. Basically just make my own dual-port equivalent to the 2600daptor D9 that is actually compatible with the MiSTer 7800 core (and probably not much else).

[remowilliams]

Therein lies the problem approaching it from that end. A bespoke D9 firmware that a) doesn't work like real paddles do and b) isn't compatible with anything else because of a) isn't appealing.

[thorr]

At this point, it is either that or live with the current situation, unless further development occurs to properly solve the problem. Someday I will need to learn how to update FPGA code for the MiSTer and then I can fix the stuff that bothers me even if it has to be with unofficial cores that break the rules. I probably could if I dedicated the time to it. I have written code for and programmed Altera FPGA's in the past. I just have too many other hobbies and change interests frequently.

If I had to pick one simple change that doesn't break the rules with this core, it would be to eliminate the auto-detection of the paddle edges for each game when an analog joystick (potentiometer) is used. Leave the auto-detection there for spinners, but take it off for analog joysticks. That way it would be simple to make a proper paddle controller that would replicate a real 2600 using only a portion of the potentiometer range. If I could pick a single more advanced change, it would be to add the option in the core menu to change the ADC function to SNAC Paddle so it works with the Atari SNAC adapter with Paddles. That would be the cat's meow.

P.S. Rest in peace Fred Ward aka Remo Williams.

[skooter]

Therein lies the problem approaching it from that end. A bespoke D9 firmware that a) doesn't work like real paddles do and b) isn't compatible with anything else because of a) isn't appealing.

The current way D9 works is not like real paddles either. It detects if a joystick is connected and map its joystick directions to analog X and Y axis. The real thing is digital... The button is mapped to button 1.
When a pair of paddles is connected, it maps their wheels to the same analog X and Y axis and the buttons to button 1 and 2. The real thing would map the buttons to the left and right digital directions.

[dmckean]

If I had to pick one simple change that doesn't break the rules with this core, it would be to eliminate the auto-detection of the paddle edges for each game when an analog joystick (potentiometer) is used. Leave the auto-detection there for spinners, but take it off for analog joysticks. That way it would be simple to make a proper paddle controller that would replicate a real 2600 using only a portion of the potentiometer range. If I could pick a single more advanced change, it would be to add the option in the core menu to change the ADC function to SNAC Paddle so it works with the Atari SNAC adapter with Paddles. That would be the cat's meow.

I'm not sure if that setting belongs in the core or with the potentiometer paddles themselves. But I agree either way, there should be a way to correctly use such paddles with the core.

[thorr]

I'm not sure if that setting belongs in the core or with the potentiometer paddles themselves. But I agree either way, there should be a way to correctly use such paddles with the core.

Definitely in the core, if it is to be done correctly. On a real Atari 2600, each game has a different spot where the paddles work. If it is not done in the core, then every game would have the same single spot which would be incorrect. My Arduino proposal above is to do it the wrong way with a single spot, which would work and be a good enough compromise if we are stuck with the current state of the paddle support forever.