OK, you said 'phototransistor', so presumably your data rate is low.

Check out the Stanley LED listings, they have some LED's that are=20
collimated to +/- 2 degrees divergence and, although part of the beam wil=
l=20
be lost, I'd be pretty sure you would have a healthy signal on the far si=
de=20
of the 4 inch tube.

If your signal is not good enough, go to a photodiode receiver, the=20
phototransistor has a dreadfully small active area, which means it's a=20
tough target to hit and much signal is lost. That's why you see them used=
=20
in aps where a sensitive receiver is NOT needed.

TAOS makes a line of integrated amp and photodiode packages that are very=
=20
cheap and easy to use. They are perfect of lower data rates and have much=
=20
larger active areas. Select the bandwidth needed from their online data a=
nd=20
submit a sample request.

Finally, although this is overkill.....consider a laser. Modern lasers ar=
e=20
driven just like LED's and you do not need a sophisticated driver if you=20
select a laser that is designed to be used with a current limiting resist=
or=20
only. The diameter of a cheap laser pointer is probably less that 1.5mm,=20
you probably could transmit over the entire distance without loosing any=20
photons because the beam is so well collimated.

Write me offlist if you need further info.

Art



At 01:42 PM 5/29/2005, you wrote:
>I want to do a bidirectional half duplex communication  over a distance =
of=20
>about 10cm (4") which seems quite trivial at first sight. The difficulty=
=20
>is the fact that I have just narrow tunnel with 2,5mm diameter (1/10") f=
or=20
>the 4" distance so the LED (TX) and phototransistor (RX) for each end ne=
ed=20
>to be very close together to be able to 'see' each other thru this tunne=
l.
>I found a IR-LED and Phototransistor in 0603 packages which could be=20
>arranged close enough - just the beam pattern is about 120=B0 so the=20
>efficiency is about the worst you can get.
>Using an optical fiber unfortunately not possible - the tunnel needs to =
be=20
>necessarily empty.
>I was thinking of using a 45=B0 semi-reflective mirror on each end to ha=
ve a=20
>beam splitter so the LED could look thru straight and the phototransisto=
r=20
>sit at 90=B0 for example - but the mechanical assembly probably is a mes=
s=20
>and how do you get the 45=B0 mirror well aligned on the pcb? Does anybod=
y=20
>know of a device with LED and phototransistor in a single package lookin=
g=20
>thru the same lens? So far I couldn't dig up anything truely useful on G=
oogle.
>The wavelength is not bound to IR - visible light would be acceptable=20
>also. For the receiver both phototransistor or photodiode is ok - but ju=
st=20
>the plain part without any preamplifier or other fancy stuff.
>
>Thanks,
>Tobias Wieler
>
>
>
>
>
>
>
>--
>http://www.piclist.com PIC/SX FAQ & list archive
>View/change your membership options at
>http://mailman.mit.edu/mailman/listinfo/piclist
>



--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist