Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

art-circuit-design question: IR pulse filtering and ADC?

Status
Not open for further replies.

AweLucid

New Member
Hey all. I am begining an art/electronics project and have some questions that i would love input on.

The basic idea of the project is to attach discrete IR emittering circuits to a performers hands or fingers and encircle their stage with an array of sensers each of which triggers a particular event when they are activated. The 1st prototype aims to simply trigger an associated cluster of LEDs to turn on above/near an activated senser. The final vision is more complex and involves many more sensors and outputs but for now I will feel great getting a this first version prototype in action.

In my first version i am aiming to to attach all sensors to different analog input pins of a single PIC chip, and then use analog to digital comparison (on the MCU) to determine which sensor has the highest voltage so I can infer where the performer is pointing. Then on the same chip I intend to use different output pins to go to MOSFET transistors that will power the desired cluster of LEDs to fire.
This method assumes that the spread of the IR LED will create the most bright point at its center, is this generally the case?

to avoid false triggering from unintended light, I am going to try to use pulsing ultrabright IR emmitter circuit, and filter it on the recieving end. Is this the reasonable meathod, and how is it best done for this situation? How will the analog comparison on the recieving side be affected, and can something be done to combat this?
I have never done this before and i do not know whether to generate the IR pulses via code on an MCU or a 555 timer and then listen for the pulse via code on each pin of the recieving MCU or to use a filtering IC (wich one i do not know)? i have heard reference of oscilation filtering ICs that can listen for such specific IR patterns...What are they and will their output reflect the amplitude of the input they recieve? and are their prebuilt IC oscilators that work with my IR LEDs and the recieving filter (that could simplify my life). can i simply put one filter inbetween each sensor and input pin so that the mcu can compare voltage levels?

thank you very much for reading this and for any help you have to offer, I am comparatively a new to circuit design, and still have have much to learn.
Many thanks.
 
new idea

perhaps it is possible to skip the pulsing alltogether and simply use a sudden change in voltage from a particular sensor to signal that it is being triggered. I could possibly do this by periodically creating a baseline analog reading on a given sensor and determine which sensor shows the greatest increase from its baseline reading to infer where the performer is pointing. Does this sound like a better/more simple plan since i am not actually sending data through the pulses? I think it could work well if i had a bright enough IR emmiter. However since i wanted multiple emitters, (one or more for each hand) the coding might be tricky. any thoughts on this revision vs the last.
 
There are are a lot of variables here.

Your best bet is to use commercial IR receivers as used in TVs and VCRs. They expect to see a fixed frequency (often 40KHz) and reject anything but light modulated at that frequency. They have a digital output (on or off).

IR emitters come in different beam widths. How wide the beam is will determine resolution (how far apart the receivers are placed).

Signal strength may not be the way to go. If the emitter is wide enough to activate more then on receiver you pick the middle one.
 
Thanks for your advice. I ordered some 38khz IR remote recievers to test out this method. Will the reciever continually output high as long as it sees 38khz IR? is it pretty straight forward to set up right? Also what is the best way to emit the correct pulse: a 555 timer, a PIC chip, an RC oscilating ciruit? do you know of any good links for a schematic to generate an IR pulse? and does should the emmission be a square wave or sinusoid?

Choosing the middle activated sensor does make sense. however, I would like to make the resolution of the motion sensor system expandale so that their could eventually be many sensors encircling the stage. thus I forsee some tricky (but possible) coding to determine the proper triggered sensor especially since their will be seperate IR emmiters for each hand that will move near and far from the target array of sensors. that was my draw to the analog level comparison idea. I see pros and cons to both methods do you have any suggestions?

many thanks again for all help
 
My Junebugs PIC18F tutor includes a typical 38kHz IR demodulator, although it may be possible to get a solid active low out when it's getting a solid 38kHz IR source the AGC may try to severely limit the gain (range). They're designed for pulse operation.
 
I've used the **broken link removed**on my IR projects

Special Features
• Enhanced data rate of 4000 bit/s
• Operation with short bursts possible (≥ 6 cycles/burst)

As for generating an IR pulse, well I had issues modifying the UART library to control the PWM module on board the PIC in Swordfish, but I found a much easier solution...

Simply setup a 38Khz PWM signal, route that signal too one input on an **broken link removed**, and the output of your UART TX pin from your PIC to the other AND gate input. Wire up the output of the AND gate to a driving circuit, eg, ULN2003, to control the **broken link removed**. The result, 38Khz modulated UART.

Works best with low speed bauds (eg, 300 b/s to 600 b/s)
 
Last edited:
You will want to use a PIC to generate the IR. Even though you have multiple emitters only one will be active at a time. Each emitter will in turn emit a short but unique encoded IR data. That will allow the computer watching the IR sensors to identify which finger activated it. The advantage is that the emitters do not interfere with each other. Also the receiving computer only has to decode data from one sensor at a time. Each finger will transmit several times per second so it will seem as if they were all working at the same time from the human perspective.

There are several ways to generate the 38KHz and modulate it to encode the unique finger ID. I personally would use a strictly software approach.
 
thanks! I cant wait to get the parts and see if it works.
I think i will try the software approach because of its versatility. When i modulate the signal, am i sending burst o 38k to be on for a duration and off for a duration, and then timing the signals of each on and off burst at the receiving circuit's software to determine the identification of the sending emitter? so do i start a clock as soon as the oscilation is detected and assume the the ID info will be read within the next particular time period? but what if it hears the signal midway through the ID name is sent, how does it know the begining and end of the ID sequence? are there any references for a good method on coding to generate and interpret such signals both the on the emitting and receiving ends? Thanks.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top