Searching \ for '[PIC] IR detecting a falling object into a beam (a' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/microchip/ios.htm?key=ir
Search entire site for: 'IR detecting a falling object into a beam (a'.

Exact match. Not showing close matches.
PICList Thread
'[PIC] IR detecting a falling object into a beam (a'
2007\06\05@103554 by alan smith

picon face
HI folks....me again with another one of those...how would you do it questions...
 
 Some time ago, had a short discussion here about getting longer distances with IR LEDs, and my thoughts on this was to pulse them at a higher current to do this.  So picture this setup.  A square "bin" or hopper that gets objects ranging from a several inches square (small toys) to a larger item (stuffed animal) and on each side of the hopper is a row of IR transmitters and across is a row of receivers...8 pairs.  So the original design when I arrived on the scene was just wiring ORing the rx and the tx was on all the time.  However, with the angle of radiation from the tx, you could  block the tx across from the rx (the pair) but the rx would still get IR from one of the adjacent tx.  I think they told me...each tx IR would have enough angle of radiation to trigger the one across from it...plus each one to the side as well.  So, even tho something might have blocked a particular tx beam, the ones next to it would still not allow that one to activate it (ie...turn it on
and pull the signal low).
 
 OK, so the idea...first pulsing the tx with higher current, and second syncronize the tx/rx pairs so that only when the tx.1 was on, rx.1 would be looking for a signal.  Works pretty good, each board has a PIC16F818 on it, and sending a sync from the transmit to the receiver so they lock up.   However....as you pull the boards away....the alignment of the beams become more difficult.  I didnt increase the power on the LED's yet...told them this morning to try that.  But the issue as I see it.....there are now HUGE holes in the sensing pattern.  The LEDs are about 1/2" apart and even tho the radiation pattern is the same, its no longer looking at the IR from the adjacent transmitter.  
 
 So, looking for thoughts on how to get rid of the holes in the sensing field. The answer might be obvious, but you know how it goes...after staring at the problem for so long, you cant see the forest for the trees.


---------------------------------
Sucker-punch spam with award-winning protection.
Try the free Yahoo! Mail Beta.

2007\06\05@105943 by Neil Baylis

picon face
Have you thought about putting the receivers on the same side as the
transmitters, looking in the same direction? They would need to be
looking at a surface that doesn't reflect much IR. When the object
falls through, if any receiver sees an increase in light, you've
detected it.

2007\06\05@111439 by Timothy J. Weber

face picon face
alan smith wrote:
>   So, looking for thoughts on how to get rid of the holes in the sensing field. The answer might be obvious, but you know how it goes...after staring at the problem for so long, you cant see the forest for the trees.

Transmit on 1 LED at a time, receive on ALL LEDs with A/D rather than
on/off, and calibrate to see what profile you get when nothing's in the
way.  Then recognize a suitable departure from that profile as
"interrupted."

Suitable departure might mean, e.g., "Any individual sensor is more than
1 standard deviation away from the mean during calibration."  Or not...
--
Timothy J. Weber
http://timothyweber.org

2007\06\05@112356 by Cristóvão Dalla Costa

picon face
Never mind synchronizing. Just make sure your receiving leds are in the
analog range, if you want you can modulate a wave and filter it in the
receiving end for better interference rejection. Then tie each receiver to a
differentiator to detect changes. When you get a change higher than a set
level that means the leds are being partially obstructed and thus receiving
less light, so you trigger.

On 6/5/07, alan smith <spam_OUTmicro_eng2TakeThisOuTspamyahoo.com> wrote:
{Quote hidden}

> -

2007\06\05@112922 by Vasile Surducan

face picon face
Depends how longer you want to get.
I've got 15m sensing with standard 900nm IR LEDs pulsed at 120mA (with
plastic lense). However the beam was splitted in two on many diodes
I've used, and at the receiver side the beam surfaces where about
30-50cm at 30cm apart.
It's not clear to me if you want just an IR courtain or not. The
simplest way is to manufacture a fresnel plastic lense in front of the
IR transmitter (you may even grab one from a pyroelectric movement
detector, choose the courtain lense or the volume lense).

But you need a lot of IR detectors on the receiver side. IR sensing
perimeters is not a very good ideea as long are using LEDs and not IR
laser beams.

Vasile

On 6/5/07, alan smith <.....micro_eng2KILLspamspam@spam@yahoo.com> wrote:
{Quote hidden}

> -

2007\06\05@113606 by alan smith

picon face
distances are between 12 and 18" and not sure if the sensitivty in the analog domain would detect it well enough.  Right now looking at hard on...hard off of the detectors with the PIC port.

Neil Baylis <neil.baylisspamKILLspamgmail.com> wrote:  Have you thought about putting the receivers on the same side as the
transmitters, looking in the same direction? They would need to be
looking at a surface that doesn't reflect much IR. When the object
falls through, if any receiver sees an increase in light, you've
detected it.

2007\06\05@141045 by Brooke Clarke

flavicon
face
Hi Alan:

I followed you up to:
"However....as you pull the boards away....the alignment of the beams become
more difficult."

Does this mean that even through timing the Tx with it's associated Rx you
still want a narrower look angle?

Have you done anything to prevent sun or bright light desense of the Rx sensor?
--
Have Fun,

Brooke Clarke
http://www.PRC68.com
http://www.precisionclock.com

2007\06\06@104818 by alan smith

picon face
I dont have to worry about sunlight, this is indoors and somewhat shaded....so no worries about sunlight causing issues.  
 
 I dont really want or need a narrow angle on the rx side, or at least I dont think I do.  In fact I'm pretty sure I do not want to do this, as it would end up leaving holes in the field.

Brooke Clarke <.....brookeKILLspamspam.....pacific.net> wrote:
 Hi Alan:

I followed you up to:
"However....as you pull the boards away....the alignment of the beams become
more difficult."

Does this mean that even through timing the Tx with it's associated Rx you
still want a narrower look angle?

Have you done anything to prevent sun or bright light desense of the Rx sensor?
--
Have Fun,

Brooke Clarke
http://www.PRC68.com
http://www.precisionclock.com

2007\06\06@124101 by Rolf

face picon face
So, being a hobbyist, I have hare-brained ideas.

Why use narrow-beamed IR TX LED's? How about this as an idea.

Set up the TX Led's to have a "chasing light" system where they
sequentially emit from one side to the other. Never have 2 LED's on at
once. Like the lights on KIT from Night-Rider. Use broad-emmitting
diodes so that all IRX Rx's can "see" the emmitter. If any one RX
doesn't get a signal, then there must be a beam break. By chasing the TX
led's you cover the whole chute. You just have to make the chase
sequence "really fast" so that it sweeps the whole chute in the time it
takes a small object to fall through.... ;-)

Actually, thinking about it, I would put alternate TX LED's and Rx's all
the way around the chute, and just identify which RX's are in
line-of-sight of which TX's.  Then chase the TX's around the chute. If
any RX does not have a signal when it is supposed to, you could actually
calculate pretty much exactly where in the chute the part fell through....

There is probably a good reason why this idea would not work...

Rolf

alan smith wrote:
{Quote hidden}

2007\06\07@162224 by alan smith
picon face
Intersting idea actually...sorta doing it that way.  
 
 Just left the client....had to visit because nothing was working...grrr
 
 Oh...helps if you install the LED's correctly.....
 
 No call from him for the last 2 hours.....maybe...just maybe.....I have him happy.
 
 naaaaaa.....never happens that easily!!!

Rolf <learrspamspam_OUTrogers.com> wrote:
 So, being a hobbyist, I have hare-brained ideas.

Why use narrow-beamed IR TX LED's? How about this as an idea.

Set up the TX Led's to have a "chasing light" system where they
sequentially emit from one side to the other. Never have 2 LED's on at
once. Like the lights on KIT from Night-Rider. Use broad-emmitting
diodes so that all IRX Rx's can "see" the emmitter. If any one RX
doesn't get a signal, then there must be a beam break. By chasing the TX
led's you cover the whole chute. You just have to make the chase
sequence "really fast" so that it sweeps the whole chute in the time it
takes a small object to fall through.... ;-)

Actually, thinking about it, I would put alternate TX LED's and Rx's all
the way around the chute, and just identify which RX's are in
line-of-sight of which TX's. Then chase the TX's around the chute. If
any RX does not have a signal when it is supposed to, you could actually
calculate pretty much exactly where in the chute the part fell through....

There is probably a good reason why this idea would not work...

Rolf

alan smith wrote:
> I dont have to worry about sunlight, this is indoors and somewhat shaded....so no worries about sunlight causing issues.
>
> I dont really want or need a narrow angle on the rx side, or at least I dont think I do. In fact I'm pretty sure I do not want to do this, as it would end up leaving holes in the field.
>
> Brooke Clarke
wrote:
{Quote hidden}

2007\06\11@143555 by alan smith

picon face
Just a quick followup.....
 
 Isnt it nice when the client is happy and everything works and the code released and he can assemble it and it works fine and the gerbers are clean and the docs are done....
 
 alls well...that ends well.  

     
---------------------------------
Yahoo! oneSearch: Finally,  mobile search that gives answers, not web links.

More... (looser matching)
- Last day of these posts
- In 2007 , 2008 only
- Today
- New search...