720 "Spinner Joystick"- I'll let you know

[tommyinajar]

Found this on Ebay to add to my controller collection crate-

 I Emailed him a few times mentioning this site, if he didn't know about it..

IDK if he's making more or if there are similar others, but I got mine, I'll post the results (did someone else already post it?)
if I can find my Tornado USB headder



"720 DEGREE ARCADE CONTROLLER KNOB FOR ALTERNATE SPINNER-HAND CANDY BALL" on Ebay

[bollwerk]

Is this a knob I could swap out on my TurboTwist 2?
(although I have no idea how to swap the knobs)

[KenToad]

With a Turbo Twist 2 or Spintrak, the correct sensitivity settings in MAME for 720 should be 12. The big issue will be the lack of auto-calibration disc. You should be able to map the auto calibration or autocenter button if you set the machine configuration to real spinner, but unfortunately, the settings don't work very well for me because the spinner will not turn exactly 1:1 with the skater. It's close, but, if you place your finger on the edge of the spinner knob in the direction that the skater is facing, it is always out of alignment after after a few spins.

Anyway, the idea of having an angled joystick on a spinner looks cool, but I question the usability of such a thing. It's especially worrying because the guy in the video says it works well as a driving wheel or in mad planets, which it doesn't seem like it would, being almost always pointed in a different direction than the way you're facing, again because of lack of perfect 1:1 movement.

Please report back if you get the spinner joystick to work decently well. It might be worthwhile to see if you can find better settings in MAME to get the spinner to the level of precision that would require less active calibration. Maybe you can do finer adjustments in the cfg/ini files, like 11.49 for analog sensitivity. 

[PL1]

The big issue will be the lack of auto-calibration disc.

Very true.

For anyone who hasn't seen a real 720 controller, it has two encoder discs.
- The one on the left is the position disc.
- The one on the right is the calibration disc.



When assembled, the discs are parallel.



Every time the joystick handle passes through the 12 o'clock position, the calibration disc slots trigger the calibration.




Scott

[J_K_M_A_N]

So, in theory, if the position disc was missing teeth, would the guy on the screen jump to the up position when the calibration disc hit its position? I am just wondering how it would actually calibrate if the position disc is off for whatever reason. Or does it only calibrate on a specific screen? I would think, once it knows which way it up, it would not need to calibrate again as it would be programmed for that encoder wheel. (obviously I don't really know what I am talking about or how it works)

J_K_M_A_N

[PL1]

So, in theory, if the position disc was missing teeth, would the guy on the screen jump to the up position when the calibration disc hit its position? I am just wondering how it would actually calibrate if the position disc is off for whatever reason.

Yes, the skater will align to "up" when the calibration signal is triggered.
- IIRC, you can confirm this in MAME v0.142 or newer versions where the "Machine Configuration" menu option of controller type includes "real".  Not sure what other settings to adjust to get the snap alignment, but there was a related thread about it here.

The position disc works just like a spinner.
- It outputs relative position changes. i.e. two steps clockwise, 5 steps counter-clockwise
- When the game board boots up, it doesn't know what direction the joystick is pointing until the calibration disc triggers.

The calibration disc is mounted so the tooth is aligned with the calibration optos when the handle is in the 12 o'clock position.
- This provides an absolute position.  i.e. When the calibration optos trigger, the joystick is passing through the 12 o'clock position.  Now that it knows where "up" is, any changes from the position wheel are relative to that "up".

Or does it only calibrate on a specific screen?

It calibrates throughout gameplay.

Every time the joystick handle passes through the 12 o'clock position, the slots on the calibration disc pass through the calibration optos, triggering a calibration signal that tells the game PCB to turn the skater to face 12 o'clock.

I would think, once it knows which way it up, it would not need to calibrate again as it would be programmed for that encoder wheel.

Remember that the encoder wheel outputs relative position changes, not an absolute position.
- If the spinner alone could hold calibration, there would be no need for the calibration disc.

I suspect that part of the problem also has to do with how the game logic applies the directional angle and speed (a vector) to the grid-like layout of the playfield.
- Add in the possibility of jitter or backlash on the position encoder wheel and it makes sense that Atari would add the calibration disc.


Scott

[J_K_M_A_N]

So, if it calibrates every time, what would happen if your position disc was messed up. Like if it was missing teeth or you replaced it with one with a different TPI. A 90% turn could register as a 50% turn and when you hit 100%, it would snap to straight up.

I guess what I am thinking is, the position disc itself must be somewhat close. Otherwise, if you don't hit straight up, it could keep getting off more and more (thinking about the downhill part...it has been a long time since I played).

Obviously, in real game play, you are going to hit straight up quite a bit doing all the spins. I am thinking more in technical terms. Completely pointless other than my own curiosity of how things work. Feel free to ignore me. 

[PL1]

So, if it calibrates every time, what would happen if your position disc was messed up. Like if it was missing teeth or you replaced it with one with a different TPI. A 90% turn could register as a 50% turn and when you hit 100%, it would snap to straight up.

If the encoder wheel is damaged/broken, repair it using the correct part.   

Wrong size teeth will cause the optos to not work.

I guess what I am thinking is, the position disc itself must be somewhat close.

It is very close, but isn't perfect so the calibration disc gives a little nudge when needed.

The only time you're likely to see a large calibration adjustment is on the very first time after booting the system.


Scott

[J_K_M_A_N]

I find that pretty interesting. Thanks for taking all that time to explain it.

J_K_M_A_N

[KenToad]

It seems like no one has gotten active autocentering to work properly in 720 on MAME and it looks like the controls are mapped incorrectly by the MAMEDEVS, at least in MAME .237. You cannot map a non-analog control to be an autocenter button and, when you map buttons to autocenter inc or dec, that just makes your skater rotate towards the 12 o'clock position when you press those buttons, rather than snapping to that position like it seems like it should.

I don't know. Maybe it is correct and the skater just slowly corrected toward the 12 o'clock position whenever the autocentering notches were triggered. Are we really even certain that it is an active calibration? I watched the game get calibrated on John's Arcade Youtube channel and it seems like the calibration disc is just telling the system where 12 o'clock is during calibration. If there is active calibration during gameplay, it's not obvious.

There's nothing in the 720 arcade manual talking about active calibration. Does anyone have a definitive answer to this question, maybe a good explanation of how it works in game and how MAME is emulating that?

[tommyinajar]

LOL i'M STILL ON IT, NEW USB headder on way, I COULD scavenge one from a finshed project, but I know It will just be broken then...

It looks like he did a fair amount of thinking going  into it, I hope it would work, I had lots of "ideas" I'd never sell on Ebay...

[tommyinajar]

It seems like no one has gotten active autocentering to work properly in 720 on MAME and it looks like the controls are mapped incorrectly by the MAMEDEVS, at least in MAME .237. You cannot map a non-analog control to be an autocenter button and, when you map buttons to autocenter inc or dec, that just makes your skater rotate towards the 12 o'clock position when you press those buttons, rather than snapping to that position like it seems like it should.

I don't know. Maybe it is correct and the skater just slowly corrected toward the 12 o'clock position whenever the autocentering notches were triggered. Are we really even certain that it is an active calibration? I watched the game get calibrated on John's Arcade Youtube channel and it seems like the calibration disc is just telling the system where 12 o'clock is during calibration. If there is active calibration during gameplay, it's not obvious.

There's nothing in the 720 arcade manual talking about active calibration. Does anyone have a definitive answer to this question, maybe a good explanation of how it works in game and how MAME is emulating that?

Wasn't 720 a semi big hit? Surprised it wasn't perfect. Of course I remember turbo Robotron, so theres THAT!

[KenToad]

This thread has a lot of interesting information.

http://forum.arcadecontrols.com/index.php?topic=5015.0

Especially reply #13 by Jerry Janis.

Hey, I think I got it!!!

Aaron Giles' code always passes a 0 for the value of the calibration spinner disc.  In order to find out if the code for the spinner disc COULD register, I added a looping integer variable that pretended to spin the second spinner disc really really fast.  Each time my fake calibration spinner registered a change, the skateboarder would turn a minute (my-noot) notch towards 12 o'clock.  (requiring 128 'notches' to actually turn the skateboarder from south all the way to north).

What this tells me is that the calibration spinner with it's two notches will only account for up to two missed notches per complete turn of the main spinner disc assuming that the track x and track y sensitivities are the same.

When I first start 720 in Mame with these changes, the FIRST time that I register movement on the calibration spinner, the skateboarder immediately jumps to 12 o'clock, thus initially calibrating it when you spin it the first time.

I set the default sensitivity for the calibration stick to 1% (darnitall, a value of 0 caused a divide by 0) and the key/joy speed to 0.  This is so that it's affect on spinner/mouse users is minimal - it will do the initial calibration and jump the skateboarder to the 12 o'clock position, but besides that I don't think it has much of a noticeable affect (avid 720' mouse users may beg to differ with me on that statement).

If you use it with the 720 joystick, then you will have to boost up the track_y sensitivity for the calibration spinner.  Also, it's likely that the track_x sensitivity will need to be adjusted (I am unable to try it with the 720 arcade joystick right now).

I changed the dial control to IPT_TRACKBALL_X and added another input port for IPT_TRACKBALL_Y and processed it as the calibration spinner, and I think it would work like the arcade machine with the 720 joystick.  You'd just have to make sure that the sensors for the spinner discs were in the 12 o'clock position and the main spinner disc was mapped to the x mouse axis and the calibration spinner disc was mapped to the y mouse axis.

The only code that I changed was 720 specific code for atarisy2.c, so it shouldn't affect any other games.  (That makes me very happy!)

Somebody with an optipac (or some kind of hack that involves both spinner discs) and a 720 controller, please try it out!  I can't wait to play it myself!

I also modified my source code for using an 8 way joystick.  It's a little bit more complicated to set up, but it will not affect any other games anymore.

Get both modified files at:
http://www26.brinkster.com/jstookey/joystick/720mame.html

Please try them out and let me know what you think!