 | | | | | | |
| |
|
I'm looking for ways to build this sensor into a studless beam project of
mine, and I was wondering, given it's large size and the comparative
brightness of the LED used to light it, how can I use it to detect if an
object is blocking a path?
Problems I can see right away:
1. The sensor is designed in such a way that the LED cannot be used to
detect objects close to it. IE some of the LED's lightsource will always
reflect back onto it. It can, however, be used to detect changes in
lightsource intensity.
2. I was wondering if the light could be piped uisng some sort of
fiber-optic cable method - this presumably would mean shielding the light
LED in some kind of sheath which would considerably narrow it. I've no idea
how hot the LED can get during operation, so presumably the sheath would
have to be able to withstand a moderate temperature rise without melting or
smoking.
3. If #2 above is possible, how difficult would it be to get an accurate
reading of on/off on the receiving side of the pipe from the LED?
I also would like to know if there are there any 3rd party light sensors out
there that use a setup where a light is transmitted at one part of the
sensor unit, and which can be received at a different sensor unit attached
to the same cable? It probably would be as simple as having a light unit in
one box, which is in turn cabled to the sensing unit in the second box via a
length of cable, and this feeds back to/is powered from the NXT unit.
Cheers ...
Geoffrey Hyde
| | | | | | | | | | | | | | | |
| |
|
Hello Geoffrey,
> 2. I was wondering if the light could be piped uisng some sort of
> fiber-optic cable method - this presumably would mean shielding the light
> LED in some kind of sheath which would considerably narrow it. I've no idea
> how hot the LED can get during operation, so presumably the sheath would
> have to be able to withstand a moderate temperature rise without melting or
> smoking.
First, don't worry about heat - such leds are really cool (thermically wise)
devices! This one dissipates only about 20 mW...
Your fiber optic cable idea is interesting... and it works well!
I used the fiber optic cable included for VLL stuff in RIS 2.0 kit (similar ones
are also available in some ExoForce sets). The fiber is maintained in front of
the LED and the other end is routed to face the sensor.
See http://www.brickshelf.com/gallery/Philo/LightBarrier/lb1.jpg and
http://www.brickshelf.com/gallery/Philo/LightBarrier/lb2.jpg
In order to try this setup, I use it to count soccer balls that are notoriously
hard to detect because of the black dots on white background. Here is the
assembly http://www.brickshelf.com/gallery/Philo/LightBarrier/lb3.jpg, and
simple NXT-G program:
http://www.brickshelf.com/gallery/Philo/LightBarrier/ballcount.rbt
With a slightly improved program
(http://www.brickshelf.com/gallery/Philo/LightBarrier/ballcount2.rbt) it is able
to properly detect and count several nearby balls:
http://www.brickshelf.com/gallery/Philo/LightBarrier/lb4.jpg
Of course, if you use this contraption in bright ambient light, you must shield
it properly!
Cheers,
Philo
| | | | | | | | | | | | | | | | | | |
| |
|
"Philippe Hurbain" <philohome@free.fr> wrote in message
news:JAF8tF.16H@lugnet.com...
> Hello Geoffrey,
>
> > 2. I was wondering if the light could be piped uisng some sort of
> > fiber-optic cable method - this presumably would mean shielding the light
> > LED in some kind of sheath which would considerably narrow it. I've no
> > idea
> > how hot the LED can get during operation, so presumably the sheath would
> > have to be able to withstand a moderate temperature rise without melting
> > or
> > smoking.
>
> First, don't worry about heat - such leds are really cool (thermically
> wise)
> devices! This one dissipates only about 20 mW...
That's good to hear.
> Your fiber optic cable idea is interesting... and it works well!
I'll have to download and check out your program. It looks great! I'll
also have to see if I can get the sensor to detect 1x1 plates - I like
working with big challenges in small sizes! ;-) It probably should detect
them from above at the very least.
But first there's the slight additional challenge of building the model in
studless so I can get 1 or at the most 2 stacked together - they tend to do
that sometimes in the hopper - in front of the sensor so I can tell if it's
going to successfully measure it or not.
> I used the fiber optic cable included for VLL stuff in RIS 2.0 kit
> (similar ones
> are also available in some ExoForce sets). The fiber is maintained in
> front of
> the LED and the other end is routed to face the sensor.
> See http://www.brickshelf.com/gallery/Philo/LightBarrier/lb1.jpg and
> http://www.brickshelf.com/gallery/Philo/LightBarrier/lb2.jpg
Your BrickShelf gallery has plenty of good building techniques. ;-)
> In order to try this setup, I use it to count soccer balls that are
> notoriously
> hard to detect because of the black dots on white background. Here is the
> assembly http://www.brickshelf.com/gallery/Philo/LightBarrier/lb3.jpg, and
> simple NXT-G program:
> http://www.brickshelf.com/gallery/Philo/LightBarrier/ballcount.rbt
By the way, is the reddish-pink hue just a camera side-effect, or did that
come about because the light doesn't travel perfectly through the fiber
optic cable? Or is it because of the orange background of the light sensor
itself?
> With a slightly improved program
> (http://www.brickshelf.com/gallery/Philo/LightBarrier/ballcount2.rbt) it
> is able
> to properly detect and count several nearby balls:
> http://www.brickshelf.com/gallery/Philo/LightBarrier/lb4.jpg
>
> Of course, if you use this contraption in bright ambient light, you must
> shield
> it properly!
Yup, I figure it's going to need quite a bit of shielding for it to get
working fairly well.
Many thanks for the contraption, and for some ideas I might be able to use
from your BrickShelf folder.
I'll have to give you a credit line! Please let me know if I forget! ;-)
Cheers ...
Geoffrey Hyde
| | | | | | | | | | | | | | | | | | | | | |
| |
|
Hi Geoffrey,
> I'll have to download and check out your program. It looks great! I'll
> also have to see if I can get the sensor to detect 1x1 plates - I like
> working with big challenges in small sizes! ;-) It probably should detect
> them from above at the very least.
Detecting 1x1 plates should not be hard if they pass close to fiber optic end to
efficiently block light.
> By the way, is the reddish-pink hue just a camera side-effect, or did that
> come about because the light doesn't travel perfectly through the fiber
> optic cable? Or is it because of the orange background of the light sensor
> itself?
I believe it's caused by the blue background I used that somewhat fooled my
camera white balance. That said there is a lot of light loss along the fiber -
These are not the high grade ones that transmit light for hundreds of km without
repeaters!!!
> Many thanks for the contraption, and for some ideas I might be able to use
> from your BrickShelf folder.
And many thanks for the original idea ;o)
Philo
| | | | | | | | | | | | | | | | | | |
In lugnet.robotics.nxt, Philippe Hurbain wrote:
> Your fiber optic cable idea is interesting... and it works well!
I've also used this the other way around, using a fiber optic cable to route
light to the phototransistor itself. Do this with more than one light sensor,
and you can have much better spatial resolution (the ends of two fiber optic
cables can be right next to each other; not so the light sensors themselves).
--
Brian Davis
| | | | | | | | | | | | | | | | |
In lugnet.robotics.nxt, Brian Davis wrote:
> In lugnet.robotics.nxt, Philippe Hurbain wrote:
>
> > Your fiber optic cable idea is interesting... and it works well!
>
> I've also used this the other way around, using a fiber optic cable to route
> light to the phototransistor itself. Do this with more than one light sensor,
> and you can have much better spatial resolution (the ends of two fiber optic
> cables can be right next to each other; not so the light sensors themselves).
Hi Brian,
What kind of application needed this precision? Line follower?
Did you need to shield the fiber to avoid stray coupling of light to the
phototransistor?
Philo
| | | | | | | | | | | | | | | | |
In lugnet.robotics.nxt, Philippe Hurbain wrote:
> What kind of application needed this precision? Line follower?
Yes. I didn't end up using it because speed ended up being more critical than
precision, but for narrow lines I may go back to it. I suspect I might be able
to use it to follow lines drawn by a black marker.
> Did you need to shield the fiber to avoid stray coupling of
> light to the phototransistor?
No. Well, kind of. I used a 1x2 Technic beam with a "stud pin" in the hole to
securely hold the fiber, and placed this right over the front face of the
olod-style light sensor. The first time I did this, it didn't work at all...
because there was almost no ambient illumination to get a reflected signal from.
When I tried it, I can't remember but I think I used a LEGO lamp as a light
sensor for surface illumination. You should be able to do this with the LED
light as well I'd guess.
--
Brian Davis
| | | | | | |
