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

[Andromeda Stardust]

Regarding Tempest/Gyruss style control, I'm aware the driving controller has much lower resolution, but it's a lot more common than either type of Atari trackball. In fact a Driving controller would probably work in a CX-80 mode ROM since it uses gray code on the Y joystick axis, but movement would be really sluggish.

[Lynxpro]

Regarding Tempest/Gyruss style control, I'm aware the driving controller has much lower resolution, but it's a lot more common than either type of Atari trackball. In fact a Driving controller would probably work in a CX-80 mode ROM since it uses gray code on the Y joystick axis, but movement would be really sluggish.

 

I have my doubts about the Driving Controller being more plentiful than the CX-22 and the CX-80. Atari made at least 500,000 of the CX-53 Trak-Ball for the 5200; in actuality, they may have cranked out close to 1 million. There's bound to be twice as many of the CX-22s and CX-80s out there than that. The Driving Controller was only packaged in Indy 500.

[Andromeda Stardust]

 

I have my doubts about the Driving Controller being more plentiful than the CX-22 and the CX-80. Atari made at least 500,000 of the CX-53 Trak-Ball for the 5200; in actuality, they may have cranked out close to 1 million. There's bound to be twice as many of the CX-22s and CX-80s out there than that. The Driving Controller was only packaged in Indy 500.

Well I got a pair of them on eBay in a 2012 BIN for $10. The trackballs go anywhere from $30-$80 depending on the exact model and condition.

[Thomas Jentzsch]

Regarding Tempest/Gyruss style control, I'm aware the driving controller has much lower resolution, but it's a lot more common than either type of Atari trackball. In fact a Driving controller would probably work in a CX-80 mode ROM since it uses gray code on the Y joystick axis, but movement would be really sluggish.

Especially in games like Tempest or Gyruss you would kill your wrist with a DC.

[Lynxpro]

Especially in games like Tempest or Gyruss you would kill your wrist with a DC.

 

It hasn't seemed to kill the wrists of Tempest 2000 fans on the Jaguar. Using the Paddle/DC tops is very common on the custom rotary controllers they build to play that game better.

 

But hey, that's yet another argument in favor of a Trak-Ball mode!

Edited November 23, 2015 by Lynxpro

[Thomas Jentzsch]

 

It hasn't seemed to kill the wrists of Tempest 2000 fans on the Jaguar. Using the Paddle/DC tops is very common on the custom rotary controllers they build to play that game better.

 

But hey, that's yet another argument in favor of a Trak-Ball mode!

It's not about using a rotary controller in general, but about the very low resolution of the Atari DC. The DC is useful for rather slow turns in alternating directions. Play a game of Indy 500 and you know what I mean.

 

Yes, Trak-Ball would most likely be the much better choice.

[BigO]

The DC only generates a maximum of 16 state changes or "counts" per rotation.

By comparison, the Wico trackball I've modified to dump out the raw quadrature encoded signals provides 422* counts per rotation (of the ball).

On top of the low resolution of the DC, it uses mechanical contacts so will be much more prone to failure if spun anything like vigorously or frequently.

 

 

 

*Since I had my Wico opened up on my workbench, I pulled these statistics:

 

Wico stats: Ball diameter = 2.2", driven roller diameter = .5". Ratio = 4.4/1.

Encoder wheel has 24 cycles per rotation.

A cycle is an alternation between 0 and 1. Since there are 24 holes between 24 spokes, that's a total of 48 state changes that can be read per rotation.

There are two sensors reading the wheel at the same time, 90 degrees out of phase. Reading these two simultaneously results in a total of 96 detectable state changes/counts maximum per rotation.

With the drive ratio of 4.4:1, and assuming zero slippage, one rotation of the ball results in 4.4 rotations of the encoder wheel.

(96 counts/rotation) * (4.4 wheel rotations/ball rotation) = 96 * 4.4 counts/ball rotation =

422.4 detectable state changes per ball rotation.

[BigO]

CX-53 has all the same specs as WICO except for the driven shaft is smaller @ .363" (.375" shaft, machined down to .363 over a .5" long section where the ball contacts it.)

 

This yields 581 counts per ball revolution. This higher count corresponds to my observation that the Wico seemed "sluggish" compared to the Atari hardware. Wico appears to have used the same hardware that's in their arcade product.

Edited November 23, 2015 by BigO

[Lynxpro]

It's not about using a rotary controller in general, but about the very low resolution of the Atari DC. The DC is useful for rather slow turns in alternating directions. Play a game of Indy 500 and you know what I mean.

 

Yes, Trak-Ball would most likely be the much better choice.

 

 

Speaking of, you're still working on 2600 Star Wars The Arcade Game, right?

 

I think you and Omega will actually propel the 2600 above the 5200 in terms of available games with Trak-Ball modes.

 

I'd really like to see Crossbow - on all Atari platforms that have the game ported to it - with Trak-Ball modes added to them. If only I could code and had the two of your talents in the field!

[Lynxpro]

The DC only generates a maximum of 16 state changes or "counts" per rotation.

By comparison, the Wico trackball I've modified to dump out the raw quadrature encoded signals provides 422* counts per rotation (of the ball).

On top of the low resolution of the DC, it uses mechanical contacts so will be much more prone to failure if spun anything like vigorously or frequently.

 

 

 

*Since I had my Wico opened up on my workbench, I pulled these statistics:

 

Wico stats: Ball diameter = 2.2", driven roller diameter = .5". Ratio = 4.4/1.

Encoder wheel has 24 cycles per rotation.

A cycle is an alternation between 0 and 1. Since there are 24 holes between 24 spokes, that's a total of 48 state changes that can be read per rotation.

There are two sensors reading the wheel at the same time, 90 degrees out of phase. Reading these two simultaneously results in a total of 96 detectable state changes/counts maximum per rotation.

With the drive ratio of 4.4:1, and assuming zero slippage, one rotation of the ball results in 4.4 rotations of the encoder wheel.

(96 counts/rotation) * (4.4 wheel rotations/ball rotation) = 96 * 4.4 counts/ball rotation =

422.4 detectable state changes per ball rotation.

 

So BigO, I know you have these mad Trak-Ball modding skills at your disposal, but have you considered hacking together a rotary controller for the 2600/5200/7800 that performs better than the DC? Seems like you might be the ideal person for such a project. But don't let that get the your way of your current Trak-hackery. After all, it would be interesting to see what other trackballs could be modded to output "CX-80" gray code...

[BigO]

 

So BigO, I know you have these mad Trak-Ball modding skills at your disposal, but have you considered hacking together a rotary controller for the 2600/5200/7800 that performs better than the DC? Seems like you might be the ideal person for such a project. But don't let that get the your way of your current Trak-hackery. After all, it would be interesting to see what other trackballs could be modded to output "CX-80" gray code...

I have given the subject some thought once or twice.

 

In your estimation, what does "performs better" entail?

Edited November 24, 2015 by BigO

[Lynxpro]

I have given the subject some thought once or twice.

 

In your estimation, what does "performs better" entail?

 

Closer to the performance of the arcade Tempest "spinner"… or the homebrew hacked rotary controllers on the Jaguar for Tempest 2000.

 

On the 2600, that would at least help a Gyruss hack and the Tempest prototype. I can't remember some of the other homebrew hacks that have added DC support. There might only be one so far.

 

On A8/5200, it would aid [potential] hacked versions of Tempest, Gyruss, and possibly Tempest Xtreem.

 

On the 7800, potential new homebrews. There's been talk over the years in the threads about a homebrew Tempest, or two.

[BigO]

I don't know what the resolution of a decent Tempest, Gyruss friendly spinner would be.

 

I suspect it's considerably higher than what the driving controller supplies (and probably less resolution than a trackball supplies). The only reason the resolution disparity would be a concern is if you wanted one controller to work with both Indy 500 and the spinner games.

 

A spinner would obviously need to be more mechanically robust than the DC. That's just a matter of time and money.

 

The market is tiny, so producing a purpose built, commercial product is probably not something for anyone to pursue.

Are the original arcade spinners still available, at least as a price comparison?

What are people using for spinners with MAME? All homebuilt stuff?

Edited November 24, 2015 by BigO

[BigO]

Another option:

 

There appear to be some USB spinners out there on the market. They may use an encoding scheme that could be tapped into. Somebody who owns one could tear into it and see what's in there.

 

Depending on the manufacturing setup, size of company, etc., somebody may be able to talk them into making a product that just dumps the raw signals instead of converting to USB (assuming it's quadrature encoded).

[Lynxpro]

I don't know what the resolution of a decent Tempest, Gyruss friendly spinner would be.

 

I suspect it's considerably higher than what the driving controller supplies (and probably less resolution than a trackball supplies). The only reason the resolution disparity would be a concern is if you wanted one controller to work with both Indy 500 and the spinner games.

 

A spinner would obviously need to be more mechanically robust than the DC. That's just a matter of time and money.

 

The market is tiny, so producing a purpose built, commercial product is probably not something for anyone to pursue.

Are the original arcade spinners still available, at least as a price comparison?

What are people using for spinners with MAME? All homebuilt stuff?

 

Here's one of many threads about making homebrew rotary controllers for Tempest 2000 on the Jag:

 

http://atariage.com/forums/topic/202166-tempest-2000-and-rotary-encoders/

Edited November 24, 2015 by Lynxpro

[Thomas Jentzsch]

Speaking of, you're still working on 2600 Star Wars The Arcade Game, right?

Not me, but Omega is at it.

[Lynxpro]

Not me, but Omega is at it.

 

My apologies. You guys are seriously cranking them out…

[BigO]

Wandering minds want to know...

 

This recent Trak-Ball mania got me to thinking about the poor unfortunate souls who are stuck with a trackball that has only joystick mode.

 

The JS emulation from a trackball toggles the standard joystick signals on and off at slow enough rolling speeds. This finer control of the joystick signals ought to be able to make it look like the player is moving slowly.

 

Taking the finer pulsed control into account, I wonder if hacking the game so that the player moves more quickly in response to joystick input could make a standard joystick game work more like a real trackball game. If so, it would probably be borderline unplayable with a real joystick.

 

I have a cheap feeling, lightweight Accuball js mode only trackball around somewhere that I haven't hacked up yet. Maybe I'll do some experimenting with my basic premise of being able to fake slower movement.

Edited November 25, 2015 by BigO

[Thomas Jentzsch]

Are you sure a Trak-Ball in joystick mode generates pulses?

 

That would mean the single check per frame would miss a pulse by 50%. And then the cursor (or whatever) would move slower than with a real joystick.

[+SpiceWare]

the poor unfortunate souls who are stuck with a trackball that has only joystick mode.

 

Why would they be stuck with it? I see a bunch of trackballs available on ebay.

[BigO]

Are you sure a Trak-Ball in joystick mode generates pulses?

 

That would mean the single check per frame would miss a pulse by 50%. And then the cursor (or whatever) would move slower than with a real joystick.

I'm confident that the behavior must exist due to how the circuitry works. But, it might only manifest when moving the ball at an impractically slow rate. I have to play test and put an oscilloscope to it.

 

It might be a fun side-effect to play with if it is usable. Just a stray thought for the moment. If I find something meaningful, I'll share in case somebody else wants to tinker.

[Lynxpro]

 

Why would they be stuck with it? I see a bunch of trackballs available on ebay.

 

It's still a compelling idea to mod those unfortunate CX-22s w/o the switch to actually support native CX-22 Trak-Ball mode.

 

Plus, we need as many possible working and available since those of us enthusiasts will want 2 or 4 of 'em each.

Edited November 25, 2015 by Lynxpro

[BigO]

 

Why would they be stuck with it? I see a bunch of trackballs available on ebay.

I dunno. Proper financial priorities or other such troublesome burdens.

 

Had I not found my $2.99 Wico which turned out to be dead and ultimately turned into a "real" trackball, I probably wouldn't see what all the hype was about. I'd just have a joystick emulating trackball stuck in my closet somewhere gathering dust.

 

So, maybe you'd be more sympathetic with the poor unused joystick emulating trackballs sitting idle, lonely, dusty and forlorn in dark closets around the world just wishing that they could go someplace as wonderful as the Island of Misfit Toys.

[BigO]

 

It's still a compelling idea to mod those unfortunate CX-22s w/o the switch to actually support native CX-22 Trak-Ball mode.

 

For whatever reason, that statement caused me to look at this from a different perspective than I have before. (Actually I have sorta looked at the problem from this perspective but never applied it to this device)

 

It occurs to me that it may be possible to mod a JS only CX-22 to be an "real" TB only device without any soldering or permanent changes. I'll have to look closer. That might be useful to some people out there.

 

Yeah...it might be possible to make it be a Dir+Clock type in the same manner if someone so chooses.

 

We'll see if I come up with any spare time over the holidays to tinker with this. Unfortunately, it will have to be theoretical as I don't have a CX-22 JS only to play with.

 

(Am I the only one that thinks of Pinnochio saying "I'm a real boy" when we say "real trackball"?)

Edited November 25, 2015 by BigO

[Andromeda Stardust]

Aren't the JS only trackballs rarer than the normal ones? They might be collectible. (ducks to avoid the oncoming onslaught of thrown bricks)