What 2600 titles could/should have Trak-Ball support added?

[Lynxpro]

I know that the 5200 Trak-Ball emulates the analog joystick so it seems like every game should "support" the Trak-Ball.

It's really like playing with a self-centering analog joystick. Only in this case, you have to spin the ball to keep the "joystick" off center instead of just applying force to the stick.

 

If I recall correctly, Galaxian is coded to uniquely handle that analog input though. Instead of trying to match the position indicated by the joystick, it just moves in the direction the joystick is pushed as if you were playing with a digital joystick.

Returning to center results in no motion at all. If the player is sitting on the left side of the screen, it will sit there until you push the joystick to the right or spin the ball to the right.

 

I remember Galaxian being a special case when I built a couple of 5200 controllers from scratch. It actually has two speeds of motion. Pushing the joystick a little bit made the player move at one speed, pushing the stick farther made the player move faster in that direction. That was a test some people gave me to see if I was hitting full range with my analog joystick.

 

Disclaimer: It's been years since I had my 5200 fired up so my memory could be playing tricks on me.

 

In theory to the first part but 5200 games to have "native" Trak-Ball support have code in them for it.

 

As for paddles, yes, the Atari engineers built Paddle controllers out of their 5200 joysticks even before the system was released. There hasn't been a true 5200 Rotary Controller though which is why there isn't specific code in 5200 Tempest for it [unlike with the Jaguar's Tempest 2000].

[BigO]

 

In theory to the first part but 5200 games to have "native" Trak-Ball support have code in them for it.

 

As for paddles, yes, the Atari engineers built Paddle controllers out of their 5200 joysticks even before the system was released. There hasn't been a true 5200 Rotary Controller though which is why there isn't specific code in 5200 Tempest for it [unlike with the Jaguar's Tempest 2000].

I can see how the software would treat the input differently from a Trak-Ball to create a "real trackball" experience.

I vaguely recall a conversation a long time ago in which somebody said how 5200 hardware/software could detect the presence of a Trak-Ball. I'll have to see if I can find it.

Edited November 16, 2015 by BigO

[Andromeda Stardust]

So the 5200 trackball is rotary encoded or outputs analog like the joysticks?

[Andromeda Stardust]

wrong thread whoops!

[Lynxpro]

So the 5200 trackball is rotary encoded or outputs analog like the joysticks?

 

My understanding was that rotary controllers are digital like the 2600 Driving Controller. 5200 Trak-Ball is analog.

[BigO]

So the 5200 trackball is rotary encoded or outputs analog like the joysticks?

It uses the same encoding method to read the movement of the ball internally, but the output is effectively the same as the analog joysticks.

The actual quadrature encoded signals are not available to the console.

[Andromeda Stardust]

It uses the same encoding method to read the movement of the ball internally, but the output is effectively the same as the analog joysticks.

The actual quadrature encoded signals are not available to the console.

Interesting to note. Is there some calibration pots so that the pointer doesn't creep if you don't touch the trackball? You know since the joysticks aren't self centering and every game used a different method of detecting center?

[BigO]

Interesting to note. Is there some calibration pots so that the pointer doesn't creep if you don't touch the trackball? You know since the joysticks aren't self centering and every game used a different method of detecting center?

The trackball itself acts like a self-centering joystick. When you stop spinning the ball, the output drops back to a neutral position. I can see on an oscilloscope that the charge and discharge time on the console capacitor(s) increase when you spin the ball one way, decreases when you spin it the other way and returns to a baseline value when you stop spinning. It's a proportional output.

 

The standard calibration in the 5200 is all there is. I'm guessing that if you cranked that calibration pot too far, you'd see drift with the trackball.

I suspect that there's a significant deadband built into the software to help prevent that drift when the joystick isn't exactly at center.

Edited November 18, 2015 by BigO

[Andromeda Stardust]

 

The standard calibration in the 5200 is all there is. I'm guessing that if you cranked that calibration pot too far, you'd see drift with the trackball.

I suspect that there's a significant deadband built into the software to help prevent that drift when the joystick isn't exactly at center.

And that seriously defeats much of the precision afforded by true trackball control. IE barely nudge the trackball and your pixel would barely nudge onscreen.

 

I was exited about someday getting a 5200 console and trackball. Still am but I don't have room for this behemoth console atm. It's just with the trackball and a handmade digital joystick, most of the control issues would be eliminated. I guess it explains why nearly all 5200 games supported the trackball through some capacity.

[BigO]

And that seriously defeats much of the precision afforded by true trackball control. IE barely nudge the trackball and your pixel would barely nudge onscreen.

I don't disagree. But, at least you don't have to spin the daylights out of the thing to get it to move, at least as I remember it.

 

I'm with you. I do prefer the "directness" of a trackball.

 

It's been quite a while since I played any 5200 games with the Trak-Ball, so I don't really remember what the experience was like. I've watched waveforms on the o'scope screen a lot more than I have fought waves of aliens on the TV screen lately.

 

It would not be a significant technical challenge to make a 5200 read a "real" trackball. But, having to have a custom piece of hardware would severely limit the market for a homebrew game. Like, probably down to an audience of one: the guy who is able to build the hardware and write the game himself.

[Andromeda Stardust]

I don't disagree. But, at least you don't have to spin the daylights out of the thing to get it to move, at least as I remember it.

 

I'm with you. I do prefer the "directness" of a trackball.

 

It's been quite a while since I played any 5200 games with the Trak-Ball, so I don't really remember what the experience was like. I've watched waveforms on the o'scope screen a lot more than I have fought waves of aliens on the TV screen lately.

 

It would not be a significant technical challenge to make a 5200 read a "real" trackball. But, having to have a custom piece of hardware would severely limit the market for a homebrew game. Like, probably down to an audience of one: the guy who is able to build the hardware and write the game himself.

Well the analog control I imagine is still far better than using a 2600 trackball in JS mode. I imagine Tempest plays nice with it at any rate.

 

I found navigating the Harmony menu with my CX-80 trackball in JS mode extremely difficult, so much I had to hot swap my arcade joystick with it. Easy enough to do with a Genesis extension cable but still. The homebrew ROMs controlled like butter though.

 

I may see about getting a cheap USB encoder for the Trackball to use with my Raspberry Pi MAME barcade. I imagine I could simply put the trackball encoder into a project enclosure and trace the pins to a 9-pin male Dsub connector. It would probably make a neat PC mouse too except it's only got one button. Whatever. They sell a laser trackball USB mouse at Office Depot I could probably get one to use in a pinch.

[BigO]

I found navigating the Harmony menu with my CX-80 trackball in JS mode extremely difficult, so much I had to hot swap my arcade joystick with it. Easy enough to do with a Genesis extension cable but still. The homebrew ROMs controlled like butter though.

 

 

I don't know if it will work at all with yours. Is yours the one one that says CX-80 but has to use the CX-22 roms? If so, I doubt it would ever work properly.

 

 

It's pretty twitchy, but it does work with my trackball.

 

When first hacking my trackball, I only did the vertical axis as a test. Stupid menu didn't work at all. Flipped wires around...still nothing. Fiddle, fiddle, fiddle. Signals look good on the logic analyzer...

Did I break my console? Nope, driving controller still runs the menu just fine. Okay, what's going on here?

Eventually I looked at the driving controller wiring and realized that the driving controller hooks up to the console on what turns out to be the trackball's HORIZONTAL axis.

 

Move the ball horizontally to navigate up and down through the Harmony menu. Barely move it, almost imperceptibly, but when you do move it move it horizontally.

 

I still sit down to play and wonder why the stupid Harmony menu is just twitching randomly at me like that...oh, yeah, HORIZONTAL.

Edited November 19, 2015 by BigO

[Andromeda Stardust]

I don't know if it will work at all with yours. Is yours the one one that says CX-80 but has to use the CX-22 roms? If so, I doubt it would ever work properly.

 

 

It's pretty twitchy, but it does work with my trackball.

 

When first hacking my trackball, I only did the vertical axis as a test. Stupid menu didn't work at all. Flipped wires around...still nothing. Fiddle, fiddle, fiddle. Signals look good on the logic analyzer...

Did I break my console? Nope, driving controller still runs the menu just fine. Okay, what's going on here?

Eventually I looked at the driving controller wiring and realized that the driving controller hooks up to the console on what turns out to be the trackball's HORIZONTAL axis.

 

Move the ball horizontally to navigate up and down through the Harmony menu. Barely move it, almost imperceptibly, but when you do move it move it horizontally.

 

I still sit down to play and wonder why the stupid Harmony menu is just twitching randomly at me like that...oh, yeah, HORIZONTAL.

So you are saying that the JS mode is rotated sideways with horizontal and vertical axes switched? When I set the trackball to JS mode, and rotate the ball slowly up and down, it tends to oscillate between the top and second to the top menu selections. Maybe the JS/TB switch is busted?

[alex_79]

So you are saying that the JS mode is rotated sideways with horizontal and vertical axes switched? When I set the trackball to JS mode, and rotate the ball slowly up and down, it tends to oscillate between the top and second to the top menu selections. Maybe the JS/TB switch is busted?

A ST mouse (and the later production CX80 in TB mode) will work as BigO described, because they output the same sequence as the driving controller when moved horizontally.

Anyway in JS mode any trackball should, in theory, be detected as a joystick and work correctly without swapping directions.

 

Does your trackball work with original (joystick) games when it's set in Joystick mode? If not, then oxydation/bad contact in the mode selection switch might be the cause.

 

Otherwise it's the Harmony menu itself which causes this.

If the trackball in JS mode sends a series of pulses on the joystick direction pins with frequency proportional to the ball speed, instead of a continuos signal, and if there's code in the Harmony menu to ignore brief changes in joystick inputs (to "debounce" the joystick switches), then this could cause your problems. In that case also some original games using similar debouncing code could be affected.

 

But I'm just speculating here, because I don't know what is the output like on a real trackball in JS mode, nor how the Harmony menu handles the input...

Edited November 19, 2015 by alex_79

[Thomas Jentzsch]

But I'm just speculating here, because I don't know ... how the Harmony menu handles the input...

The Harmony menu debounces joystick input in two ways:

1. While the joystick is being pressed, it delays the repeated action by 1/4s. So when you keep the joystick moved e.g. down, the cursor moves down 4 times each second.
2. If you release the joystick in between, then you can repeat action immediately. So you can decide how fast you want to move the cursor.

The menu control problems in trackball mode are resulting from the way different sensitivity. A driving controller creates way less impulses per turn than a trackball does. And the code is designed to work with a DC, not TB. I suggest using the console's SELECT (works like joystick down) and RESET (works like the fire button) switches instead.

[Lynxpro]

And that seriously defeats much of the precision afforded by true trackball control. IE barely nudge the trackball and your pixel would barely nudge onscreen.

 

I was exited about someday getting a 5200 console and trackball. Still am but I don't have room for this behemoth console atm. It's just with the trackball and a handmade digital joystick, most of the control issues would be eliminated. I guess it explains why nearly all 5200 games supported the trackball through some capacity.

 

The CX-53 5200 Trak-Ball is precise. Dan Kramer prefers it over the CX-22 even though both are his babies. But for the record, he's tickled pink over all of these Trak-Ball hacks on the various platforms. He's promoted the hell out of Missile Command-TB in person ever since he got his copy from Albert for the 2014 Sunnyvale Atari Party.

 

The CX-80 was designed by the Atari Home Computer division based upon his CX-22. They wanted their own Trak-Ball to match the XL computer style. And Dan was not privy to GCC's hackery of the CX-22 to be deemed the "7800 Trak-Ball" which never made it out the door [yet its code is present in the commercial version of 7800 Centipede as Ken discovered].

[BigO]

So you are saying that the JS mode is rotated sideways with horizontal and vertical axes switched? When I set the trackball to JS mode, and rotate the ball slowly up and down, it tends to oscillate between the top and second to the top menu selections. Maybe the JS/TB switch is busted?

I don't have JS mode, but I doubt that would be the case since that would be different from how it works with a real joystick. I'm saying that on my trackball that works with the CX-80 ROMs, menu navigation is through horizontal trackball movement as opposed to vertical. I expected vertical. But, I eventually realized that there was no reason that I should have expected vertical.

[BigO]

 

The CX-53 5200 Trak-Ball is precise. Dan Kramer prefers it over the CX-22 even though both are his babies. But for the record, he's tickled pink over all of these Trak-Ball hacks on the various platforms. He's promoted the hell out of Missile Command-TB in person ever since he got his copy from Albert for the 2014 Sunnyvale Atari Party.

 

The CX-80 was designed by the Atari Home Computer division based upon his CX-22. They wanted their own Trak-Ball to match the XL computer style. And Dan was not privy to GCC's hackery of the CX-22 to be deemed the "7800 Trak-Ball" which never made it out the door [yet its code is present in the commercial version of 7800 Centipede as Ken discovered].

Interesting history. I wonder if the CX-80 labeled TB floating around here that acts like a CX-22 is a result of a factory effort or of a hobbyist somewhere along the line patching something together with the parts they had available.

 

For the record, I'm not in any way putting down the CX-53. I think it's cool and I intend to study it a bit more to understand the one last piece of it that I don't fully grasp yet. I think I may want to leverage the methodology for another project.

 

I could be wrong about that proportional response as it's difficult to measure and I only observed it in passing.

 

I need to hook up the CX-53 and get some more recent experience with actual games. I haven't played much in several years when I was building controllers. My question is not one of precision, but of sensitivity. With a less processed trackball signal, I can make one little twitch and the game can respond instantly. Does it take a more sustained effort to make the CX-53 provide meaningful activity on the screen? If so, how much of that is the fault of the software?

 

One thing is keeping me from looking at some meaningful technical details. I can't get my multi-cart to respond correctly to the input from the CX-53 so I can't get to the diagnostic cart selection. Probably need to tweak the pot in the 5200. I'll probably go there after I finish another project.

Edited November 19, 2015 by BigO

[Andromeda Stardust]

Interesting history. I wonder if the CX-80 labeled TB floating around here that acts like a CX-22 is a result of a factory effort or of a hobbyist somewhere along the line patching something together with the parts they had available.

I somehow doubt it. If someone had modded the CX-80 trackballs to function as CX-22, there would likely be only one documented case of a CX-80 acting like a CX-22. Additionally when opened, there would be evidence of modding/tampering. I can confirm mine is 100% stock. I popped the rubber feet off to clean it and it left double sided tape which I had to puncture with a screwdriver. There was still enough adhesive left for me to reapply the feet after I sealed it back up.

 

Albert and others have documented that some CX-80s behave as CX-22s while others behave as ST mice. As for it not behaving well in JS mode, I'm going to investigate the trackball using the testcart ROM on my Harmony. It measures joystick, paddles, and keypad inputs. I should be able to diagnose the behavior of the joystick inputs with the CX-80 trackball in both JS and TB mode.

 

For the record, I'm not in any way putting down the CX-53. I think it's cool and I intend to study it a bit more to understand the one last piece of it that I don't fully grasp yet. I think I may want to leverage the methodology for another project.

 

I could be wrong about that proportional response as it's difficult to measure and I only observed it in passing.

 

I need to hook up the CX-53 and get some more recent experience with actual games. I haven't played much in several years when I was building controllers. My question is not one of precision, but of sensitivity. With a less processed trackball signal, I can make one little twitch and the game can respond instantly. Does it take a more sustained effort to make the CX-53 provide meaningful activity on the screen? If so, how much of that is the fault of the software?

 

One thing is keeping me from looking at some meaningful technical details. I can't get my multi-cart to respond correctly to the input from the CX-53 so I can't get to the diagnostic cart selection. Probably need to tweak the pot in the 5200. I'll probably go there after I finish another project.

Yes, I do plan on getting a 5200 with trackball controller someday...

 

Having slept on it since last night, I would imagine that the CX53 probably uses a timer delay as digital logic output to the analog paddle lines, then increases or decreases the delay proportional to the tick rate of the trackball optical encoder. Shorter clicks = greater positive or negative delta to the timer delay. I imagine any "dead zones" would be applied by the game logic itself since the 5200 analog joysticks are fairly sloppy, lacking a self centering mechanism.

[+Mitch]

Interesting history. I wonder if the CX-80 labeled TB floating around here that acts like a CX-22 is a result of a factory effort or of a hobbyist somewhere along the line patching something together with the parts they had available.

 

Actually the opposite, the CX-80s natively mimic the CX-22. The CX-80s the mimic the ST mouse are the ones that are factory or hobbyist modded and are rarer than the unmodded version.

 

Mitch

[BigO]

 

Actually the opposite, the CX-80s natively mimic the CX-22. The CX-80s the mimic the ST mouse are the ones that are factory or hobbyist modded and are rarer than the unmodded version.

 

Mitch

There never was a production Atari CX-anything Trak-Ball that output what we've taken to calling "CX-80" straight Gray code?

 

So...does that mean I'm virtually the only one using the CX-80 ROMs with my hacked up hardware?

I'm glad the guys made those versions or I wouldn't get to play with them at all. Now I feel kind of exclusive in an inclusive kinda way.

Edited November 20, 2015 by BigO

[alex_79]

There never was a production Atari CX-anything Trak-Ball that output what we've taken to calling "CX-80" straight Gray code?

Some of the latest units were patched directly at the factory and they shipped already modded.

Here is an old discussion about that on the stella list:

http://www.biglist.com/lists/stella/archives/199902/msg00012.html

 

 

So...does that mean I'm virtually the only one using the CX-80 ROMs with my hacked up hardware?

Not for long. I just won a MSX trackball on ebay that will soon go under surgery...

Edited November 20, 2015 by alex_79

[Lynxpro]

Since this is a suggestion thread, I'm going to throw it out there… 2600 Gyruss. Sure, it would probably play best with the Driving Controller but having played 5200 Tempest with the Trak-Ball made me believe Gyruss could also benefit from the Trak-Ball experience. I guess the same could go for that prototype of 2600 Tempest.

 

2600 Pole Position would also be interesting since the 5200 version does support the Trak-Ball but I admit a Driving Controller hack would play better. But hey, this is about expanding the Trak-Ball experience!

Edited November 22, 2015 by Lynxpro

[Andromeda Stardust]

Since this is a suggestion thread, I'm going to throw it out there… 2600 Gyruss. Sure, it would probably play best with the Driving Controller but having played 5200 Tempest with the Trak-Ball made me believe Gyruss could also benefit from the Trak-Ball experience. I guess the same could go for that prototype of 2600 Tempest.

 

2600 Pole Position would also be interesting since the 5200 version does support the Trak-Ball but I admit a Driving Controller hack would play better. But hey, this is about expanding the Trak-Ball experience!

Wouldn't the driving controller be better for Tempest/Gyruss style play?

[Lynxpro]

Wouldn't the driving controller be better for Tempest/Gyruss style play?

 

I said that in my original comment. However, based upon the playability of 5200 Tempest with the 5200 Trak-Ball, it might be cool to also play 2600 Gyruss with a Trak-Ball as well. Of course, a Driving Controller version would also be [extremely] cool.

 

I mean, the same goes for Pole Position. 5200 Pole Position can be played with its Trak-Ball. Of course, Dan Kramer's goal was to have the code added into the game so that he could take advantage of it as well with an actual 5200 Driving Controller - he was working on adapting the arcade Atari Pole Position 2 steering wheel and shifter into a product for the 5200 - but alas, the industry crash happened and then management didn't support him continuing with that project and as well with a 5200 Yoke controller. So the code is there but is only taken advantage of by the Trak-Ball.