RBS Ball detection notes
Ball detection is essential in checking that the scultpure is running properly. There are two ideas here that I need to look at. 1) if I can get the ball detection to work 100% of the time then I can make the whole electronics of the system much simpler. 2) hmm maybe only 1 issue
Ball detection
Ball detection can be done in several ways. Light, sound, magnetic field, switches. One of the easiest ways is to use IR light but this has the disadvantage that the background like can overpower the IR signal. There are chips that have been design to overcome this but its still not ideal, means that I cannot garantee that it will work in all situations. So I can look at how to improve that and do tests around where the ball is and how the detection system work or I can look at something else.
There are two ways that I want to detect the ball.
1) beam breaking - used for situations where there are a longer stretch of track where the ball might break the beam,or
2) beam bounce - used for single ball
Important points
1) I can use the ball to shade the detector
2) in beam break I can have it very directional. And hence design the layout with the extra light source in question.
So I need to test out whats happening in real life.
1) make a jig that has a ball in the center and a IR sensor that can rotate around with different times of the day. Actually it makes not difference what time of the day it is just how strong the sun is and the angle involved.
2) make the jig so that I can test all the angles relative to the sun. can store the information. Need to see what the relative signal back is and what angles work. Then need to look at how the sun will move. Or I design something that works at all angles that is idea. Yes works at all angle.
notes
1) I could detect when the system is saturated with light. How would I do that? do I need to. Full sun light test out would be ideal.
2) I could have more than one detector in the system if certain angle didn't work. Again not idea as I want something that is fool proof( or close).
3) the shileding needs to be symetrical and should be far enough away that a ball can roll over it. more distance could be good.
construction notes
1) to create IR signal I can either do it vai the attiny85 internals or a 555 timer chip
2) need to think about how this is attached to the sculpture. 3D printing could be cool but simple is good. Arty is even better.
demodulator that looks like I could hand solder - http://www.mouser.com/ds/2/427/vsop584-107262.pdf
Ball detection
Ball detection can be done in several ways. Light, sound, magnetic field, switches. One of the easiest ways is to use IR light but this has the disadvantage that the background like can overpower the IR signal. There are chips that have been design to overcome this but its still not ideal, means that I cannot garantee that it will work in all situations. So I can look at how to improve that and do tests around where the ball is and how the detection system work or I can look at something else.
There are two ways that I want to detect the ball.
1) beam breaking - used for situations where there are a longer stretch of track where the ball might break the beam,or
2) beam bounce - used for single ball
Important points
1) I can use the ball to shade the detector
2) in beam break I can have it very directional. And hence design the layout with the extra light source in question.
So I need to test out whats happening in real life.
1) make a jig that has a ball in the center and a IR sensor that can rotate around with different times of the day. Actually it makes not difference what time of the day it is just how strong the sun is and the angle involved.
2) make the jig so that I can test all the angles relative to the sun. can store the information. Need to see what the relative signal back is and what angles work. Then need to look at how the sun will move. Or I design something that works at all angles that is idea. Yes works at all angle.
notes
1) I could detect when the system is saturated with light. How would I do that? do I need to. Full sun light test out would be ideal.
2) I could have more than one detector in the system if certain angle didn't work. Again not idea as I want something that is fool proof( or close).
3) the shileding needs to be symetrical and should be far enough away that a ball can roll over it. more distance could be good.
construction notes
1) to create IR signal I can either do it vai the attiny85 internals or a 555 timer chip
2) need to think about how this is attached to the sculpture. 3D printing could be cool but simple is good. Arty is even better.
demodulator that looks like I could hand solder - http://www.mouser.com/ds/2/427/vsop584-107262.pdf
Simple RBS system
At present I've got a system that does a clean and has centralised control. If I can get the ball detection working for all light situations then I can make the system simpler as I should be doing as well as I can and henc not need to clean the system. This would dramatically simplify the system and allow me to have attiny85 control the system (excluding the cellphone and the timing unit). This has the advantage that its simpler and I can build components that are smaller. again I would need to see how things here connected. But I could need power (5V) and GND and then connections between the system. I would remove the need for the MUX boards.
Here is the basic idea
1) the system always does the best to clean itself
2) when the system fails I detect this and shut down, then send of the txt message. Not cleaning required.
3) I have a timer and cell phone system that talk to each other. Cell needs to be separate because of power requirement (I think) and the the LCD is i2C anyway so this works. So no big deal.
4) the cell phone is the central i2c board and can talk(i2c) to the timer and maybe to the detection system (the detection system will still send a signal to one of the pins).
5) the cleaning and firing system are controlled from atting85, so its using simple building blocks to create a complex system. The building blocks are ball detection, servo control, group interaction. These are just plug and play units that can be replace easily and tested easily. Don't require a computer to debug and hopefully are small enough.
note - its going to change the look and feel but if the MUX fail then this is a good option, if it works otherwise I will need to have the 328 option. This is also good but more complicated and requires more programming, and I could get away with a simple system. Simple is better and more maintainable.
20140427 - First step
1) get the cell and the time talking to each other. Do I need a i2c shield for this, something generic that allow me to easily connect things together. It looks like everything else is in order. So Shield or no shield that is the question. Shield is nice in that it really clear about what should be happening and the wires won't come out so easily. So yes i2c shield, what are the connections that I want to use. Just want I have at present, don't need anything to fancy here but could have it generic so if I want to put polarized connector in then its cool. Get them talking together. the two lines are pull to Vc. Also ground will be common between the boards. So three line connection? Yes for arduino 3 line connection. GND, SDA and SCL.
2) look at redesigning the counter shield so I have the two A4 & A5 free for i2c
IR LED Unit - QRD1114 - this has both LEDs in a nice package.
20140430 - note - need to be able to connect the LED directly into the board? or just have a 4 connection. have a demodulator
Here is the basic idea
1) the system always does the best to clean itself
2) when the system fails I detect this and shut down, then send of the txt message. Not cleaning required.
3) I have a timer and cell phone system that talk to each other. Cell needs to be separate because of power requirement (I think) and the the LCD is i2C anyway so this works. So no big deal.
4) the cell phone is the central i2c board and can talk(i2c) to the timer and maybe to the detection system (the detection system will still send a signal to one of the pins).
5) the cleaning and firing system are controlled from atting85, so its using simple building blocks to create a complex system. The building blocks are ball detection, servo control, group interaction. These are just plug and play units that can be replace easily and tested easily. Don't require a computer to debug and hopefully are small enough.
note - its going to change the look and feel but if the MUX fail then this is a good option, if it works otherwise I will need to have the 328 option. This is also good but more complicated and requires more programming, and I could get away with a simple system. Simple is better and more maintainable.
20140427 - First step
1) get the cell and the time talking to each other. Do I need a i2c shield for this, something generic that allow me to easily connect things together. It looks like everything else is in order. So Shield or no shield that is the question. Shield is nice in that it really clear about what should be happening and the wires won't come out so easily. So yes i2c shield, what are the connections that I want to use. Just want I have at present, don't need anything to fancy here but could have it generic so if I want to put polarized connector in then its cool. Get them talking together. the two lines are pull to Vc. Also ground will be common between the boards. So three line connection? Yes for arduino 3 line connection. GND, SDA and SCL.
2) look at redesigning the counter shield so I have the two A4 & A5 free for i2c
IR LED Unit - QRD1114 - this has both LEDs in a nice package.
20140430 - note - need to be able to connect the LED directly into the board? or just have a 4 connection. have a demodulator