using blackberry track balls in controllers, ps3,360 etc.

[1TONpete]

Here is the part I want to use ...

here is a sample code for changing the leds color / intensity to directional movement...
https://docs.google.com/Doc?docid=0ASu71eTOMQ9zZDl3a3J4N182MmY0YzJtc2Nj&hl=en
I would like to see this used as red/blue instead . Everyone complains about green brightness. The guy who wrote this said swapping red for green was best.
Also if we do something like bellow , i would like white on button press.
more sample code ... think this is for reading as a variable when pressed on button... my thought... have it read like a stick unpressed and have it read like a dpad pressed or vsv.
This code is posted on the product page by a user...

"save the previous value to a variable, then read the new state and compare, works like a charm for me on a PIC: For below, my ball is connected to PortA pins 0-3 (button on 4)"

" Port A --> Pressed (variable) If(pressed AND 16 = 0){ //button pressed } If(pressed AND 8 up){

//up has changed up = (pressed AND 8) } If(pressed AND 4 down){ //down has changed down = (pressed AND 4) } If(pressed AND 2 left){ //left has changed left = (pressed AND 2) } If(pressed AND 1 right){ //right has changed right = (pressed AND 1) }

someone please step in and help?

[stan4]

Nice work buddy.

Where did you get that "trackball" breakout board?

I'm working on unificating the comfort of the controller and the accuracy of KB/mouse.

[hyper999]

Nice work buddy.

Where did you get that "trackball" breakout board?

I'm working on unificating the comfort of the controller and the accuracy of KB/mouse.

Sparkfun

[1TONpete]

The picture is a link to the product.

[stan4]

Hahahahaha, of course, the good old spark.

well, somehow I missed this beauty in sparkfun; now that I read the specs, I'm kinda sad it's resolution is 9 pulses per revolution.

I never used a blackberry, I will look some videos on youtube to see what it does on it, if it's used to control a "mouse" cursor, that resolution would be enough, if it's just menu browsing it might be too slow.

** Indeed, it's used for menu browsing, it was designed for that purpose, the resolution is definitely low for replacing an analog stick on a controller.

*** One thing that comes to my mind is to scrap from it the magnetic rollers and use them with a bigger ball, that'd be a sick trackball.

[hyper999]

just thought, 555 or 556 timers could be used in order to avoid using a micro-controller...

[1TONpete]

One thing that comes to my mind is to scrap from it the magnetic rollers and use them with a bigger ball, that'd be a sick trackball.- i like this idea.

[1TONpete]

I have in my possession one of these track ball breakouts now. I did some early testing with dpad out put. interesting. Will try hook directly to ps3 controller analogs and see what happens.... While I did get response out of the dpad track ball , I got nothing out of the sticks... was expected though. The track ball in menu was wicked sensitive. lil movements where hard to predict.

[FOOKz™]

sooo.... basically if you want to use a 555 timer build a one-shot with the timer.

When a pulse goes to the 555 timer after someone moves the cursor the chip will make the output pin High for a certain amount of time that's dependent on the Resistor-capacitor circuit.

http://ecelab.com/circuit-monostable-555.htm
C1 and R1 will trigger the output.

So if you want to have the output trigger for 1 second then R1 should be about 200K Ohms and then C1 should be 4.7 Micro farads

[hyper999]

If you use a pot for the resistor you can tweak the sensitivity of each direction

[1TONpete]

Wow Fookz you never fail to amaze me. Thanks for stopping in and having a look.
I will try that w a pot for the resistor and when I get a setting I like , I'll measure the resistance on the pot to get the proper resistor.
If i want to use a 10k pot , i still need a resistor in there or can I just use a different cap to match?

[FOOKz™]

i'd personally use a small 1uF cap and a large value resistor (keeps the size of the circuit very tiny).

If i want to use a 10k pot , i still need a resistor in there or can I just use a different cap to match?

just replace R1 with your pot *plus a 100 Ohm resistor in series so it doesn't short out.

You can play around with the cap values any way you like its fundamentally the balance between cap size and how long you want the output pulse to be.

[1TONpete]

Sweet. I'm going to the shack l8r it looks like, I was thinking about picking up some 555's to try one on sharpshooter.
Edit: After looking at the diagram , 555 will only give me 1 input and 1 out put. I think programming is the way to go. All on one chip. For 2 track balls I would need 8 "555" timers.

[FOOKz™]

We're going to build a DCA also called a Digital to Analog Converter.

You were saying that the trackball outputs are PWM, as the trackball moves faster the PWM signal pulses faster. The trackball can only move in two vectors simultaneously at once.

Realistically in a PS3/Xbox360 controller both of the X and Y axis pots have four outputs, because the center pin is usually common and then the chip recognizes the voltage drop on ether side of the pot therefore it knows what direction to go. (there are multiple ways to look at this).

In this case we have to just convert the speed of a pulse to a voltage output as simple as that. So now we have to do is figure out how to program a chip to convert our signal.

[stan4]

Sweet. I'm going to the shack l8r it looks like, I was thinking about picking up some 555's to try one on sharpshooter.
Edit: After looking at the diagram , 555 will only give me 1 input and 1 out put. I think programming is the way to go. All on one chip. For 2 track balls I would need 8 "555" timers.

Programming is always the best choice, with such a low price on these marvels, you can't go wrong.

USB capable, i2c, spi, ADC, PWM, timers, interrupts and you can change everything by code ten thousand times if you want.

[stan4]

We're going to build a DCA also called a Digital to Analog Converter.

You were saying that the trackball outputs are PWM, as the trackball moves faster the PWM signal pulses faster. The trackball can only move in two vectors simultaneously at once.

Realistically in a PS3/Xbox360 controller both of the X and Y axis pots have four outputs, because the center pin is usually common and then the chip recognizes the voltage drop on ether side of the pot therefore it knows what direction to go. (there are multiple ways to look at this).

In this case we have to just convert the speed of a pulse to a voltage output as simple as that. So now we have to do is figure out how to program a chip to convert our signal.

FYI, the trackball itself is a ball assembly with 4 rollers, each roller has small magnets ( I don't know how many ) that's it, it's mechanical. The rest is on the PCB, the outputs come from the 4 hall effect sensors that pickup those magnetic signals, the SMD sensors are located below the magnetic rollers.

[hyper999]

Tell me what you want the code to do and I'll try my best to do it.

[1TONpete]

in a message to fookz...
"1 360 of the ball = 9 high lows in one direction. So every 40 * rotation is a pulse.
only 2 halls are ever triggered at once. x,y (+or-) That part is mechanical, way they designed the ball.
so to make an equivalent ,1 pulse could equal small resistance (analog) for a certain amount of time .  multiple pulses in a given amount of time could equal more resistance (analog) for a certain amount of time. (sensitivity) I could be backwards on more or less resistance. possible?
or would it have to be more like 1 pulse is a given amount of resistance (analog) and multiple pulses in that same direction would be more/less resistance (analog)?"

[hyper999]

Well ok I want to use an atmega168 but I would need someone else to design the DAC circuitry...

[FOOKz™]

Well ok I want to use an atmega168 but I would need someone else to design the DAC circuitry...

use a resistor ladder.

[hyper999]

Yeah i thought we would use a R2R but not quite sure how we would implement it in this instance...

[1TONpete]

Thank you guys for stepping in. Its getting past the point of me helping now.

[stan4]

Yeah, sort of integrating the number of pulses per unit of sample time to get the resistance desired.

I'm guessing you want a DAC to generate the voltage for the Microsoft chip, I was thinking that as the sticks are 10 kilo ohm potentiometers, you could use electronic digitally controlled variable resistors. They behave like potentiometers, the cool feature is that the position of the wiper is set by you digitally.

Using this you just have to connect your digital pot where the original was (in theory, lol).

http://en.wikipedia.org/wiki/Digital_potentiometer

[1TONpete]

I did some messing with the PS3 version of this mod.
The Ps3 has a direction input instead of Just x,y on analogs. The pins for the PS3 hall sensors go like this ...

at rest everything zero'd .

X sensor
pin 1 =ground
pin 2 =VCC 2.88
pin3 =1.40
pin4= 1.40
Y sensor
pin 1 =g
pin 2 =vcc 2.88
pin 3 =1.40
pin 4= 1.40
 

movement to the left max
X sensor
pin 1 g
pin2 = v2.88
pin3 1.44v
pin4 =1.36v
Y sensor
pin 1 =g
pin 2 =vcc 2.88
pin 3 =1.40
pin 4= 1.40

Movement to the right
X sensor
pin 1 g
pin2 = v2.88
pin3 1.36v
pin4 =1.44v
Y sensor
pin 1 =g
pin 2 =vcc 2.88
pin 3 =1.40
pin 4= 1.40


Movement up
X sensor
pin 1 =ground
pin 2 =VCC 2.88
pin3 =1.40
pin4= 1.40
Y sensor
pin 1 g
pin2 = v2.88
pin3 1.36v
pin4 =1.44v


Movement down
X sensor
pin 1 =ground
pin 2 =VCC 2.88
pin3 =1.40
pin4= 1.40
Y sensor
pin 1 g
pin2 = v2.88
pin3 1.44v
pin4 =1.36v
 
The vales for the 3,4 pins change in exact opposite proportion to each other.


Here is a design I drew up. Inside the blue box I would need a chip or something to take the pulse in from the track ball and drop the v with in range of the sticks and what i would like to see happen is the pulse converted into a pulse /time = V swing. So lets say 3 pulses per sec = max movement or 1.44+ on 1 pin and 1.36 on the other pin. 1 pulse per sec can equal min movement or 1.41v on 1 pin and 1.39 on the other pin.
For a dpad adapter, I am thinking about just dropping the v with in limits of the analog signal for dpad or the dpad comm and see what happens. But the dpad works by connecting an analog signal to the direction. So I would need to flop from ground to vcc instead of high low.

[Rodent]

Seems to be  coming along pretty good there , 1TONpete nice work

[magic_man185]

So when the track ball pulses to signify movement, what does the pulse consist of? At idle, is the wire from the trackball high or low?  If high, does it pulse to ground (or if low, pulses high) however many times per second? I am kinda formulating an idea in my head, but i still need a little more information.

[1TONpete]

The way this thing works is the magnets spin , the flip flop of the magnets poles causes the pulse. Each direction has its own magnet. IF the flip flop lands on a flip and flip = on pulse then if left in that position it will stay on. To bad I couldn't just put  a lil tiny wire coil around the spinning magnets and generate a variable V out depending on how fast i spin the track ball . 
Here is a demo video of the track ball in use . the output for the track ball is hooked up to the four color leds under the track ball. I want a pic to read the pulses coming out and count how many pulses per .5secs and output those pulses as a given max output v of 1.44 and a min of 1.40.

[1TONpete]

VIDEO DEMO 1
http://s1226.photobucket.com/albums/ee417/Erin_Schnare/trackball360/?action=view&current=DSCF5591.mp4
VIDEO DEMO 2
http://s1226.photobucket.com/albums/ee417/Erin_Schnare/trackball360/?action=view&current=DSCF5592.mp4

[Rodent]

Upload videos here  http://www.youtube.com/

Make an account will get a better video , you got talent why not show it off

[1TONpete]

Yeah I dont have a login for youtube anymore

[wetfusion]

dude i was just discussing fitting a blackberry trackpad in the controllers with snowcolt a few days ago. :faceplam:

[1TONpete]

This is what I would like to see happen in the code.

count pulse over time coming into the chip from track ball

Output pins to controller.
Output variable V depending on # of pulses per sec.
The output would also have to swing opposite directional pin the opposite V so if .01 is added .01 needs to be taken away from the other pin.
directional pins get there own pinout each starts with
v constant at 1.40
I would like to have 3 modes. Have the mode switchable by a tact.

Mode 1

3pulses per 0.5sec = 1.43v
2pulses per 0.5sec =1.42v
1pulse per 0.5sec =1.41v

Mode 2
6pulses per 0.5sec = 1.43v
3pulses per 0.5sec =1.42v
1pulse per 0.5sec =1.41v

mode 3
7pulses per 0.5sec = 1.43v
5pulses per 0.5sec =1.42v
1pulse per 0.5sec =1.41v

These different modes would give the user sensitivity and we could make it so the deadzone is adjustable . by warying the code for each directional output.
All outputs would have to pulse for a given amount of time say .25 sec.

edit from below...
Found this interesting 55 timer gif thought id share here. IF i can convert that pulse to a variable vcc like in the image that would be perfect. Use 2 556 timers to create 4 directional variable vcc outputs.

 Something similar to this circuit
(anyone care to take this and modify it to work with in the tolerances needed?)
                                                        Car Tachometer

The timer receives pulses from the distributor points. Meter M receives a calibrated current thru R 6 when the timer output is high. After time-out, the meter receives no current for that part of the duty cycle. Integration of the variable duty cycle by the meter movement provides a visible indication of engine speed.
here is the signal output on graph.

[1TONpete]

I purchased 2 556 timers (dual555) .
My plan is to build a breadboard 1 direction circuit using 1 555 timer as a proof of concept.
I will then need help designing a circuit to do all 4 directions using 2the 2 556 timers. Please, I want to repeat the design of the tach circuit but made for the controller analog voltages.  I need help making the adjustments to the design to work on controller voltages and tolerances.