Inexpensive

I'm building a little toy, and need to read a hand-encoded storage
medium something like punched tape. Right now I'm looking at the
Panasonic CNB1002 reflective photosensors, which integrate an IR LED
and phototransistor in a small gull-wing SMD package. In my small
quantities they're about $0.55 each. I need three - a clock bit, and
two data channels. Apart from the cost they're perfect - they read
back well at a distance up to about 0.5mm (claimed 1mm, but my
materials aren't ideal).

Is there a cheaper alternative? Discrete LEDs and phototransistors
seem to be even more expensive.

I looked at CdS cells but they're also expensive-ish, and also don't
respond fast enough - I want to read 256 clocks in a 1-second run, and
that means about 1.5ms response time or better. Plus the varieties
I've found are not well shaped for the application, as I only have one
mounting surface for the illuminator and reader (but I'd be happy to
be pointed at other flavors of cell, since I was fighting Digi-Key's
search engine.

You know better than I about these things.  But can you explain why
you cannot consider a pair of very cheap smt LEDs or else an LED and a
glass-encapsulated diode (which will respond to light?)  Is it the
conditioning circuitry?

Jon

On Jul 13, 3:53=A0pm, Jon Kirwan <j...@infinitefactors.org> wrote:
> You know better than I about these things. =A0But can you explain why

I doubt that!

> you cannot consider a pair of very cheap smt LEDs or else an LED and a
> glass-encapsulated diode (which will respond to light?) =A0Is it the

Hmm, I haven't tried using an LED as a photodetector, you know - I've
only read about it. I have dozens of reels of 0805 LEDs lying around
that would suffice for production for the foreseeable lifespan of this
thing... I'll read up on that. Note that I need to read three at once,
so it's a minimum of four, probably six LEDs. But still cheaper than a
packaged device.

Thanks.

On Mon, 13 Jul 2009 13:00:24 -0700 (PDT), larwe <zwsdotcom@gmail.com>
wrote:

>On Jul 13, 3:53&#4294967295;pm, Jon Kirwan <j...@infinitefactors.org> wrote:
>> You know better than I about these things. &#4294967295;But can you explain why
>
>I doubt that!
>
>> you cannot consider a pair of very cheap smt LEDs or else an LED and a
>> glass-encapsulated diode (which will respond to light?) &#4294967295;Is it the
>
>Hmm, I haven't tried using an LED as a photodetector, you know - I've
>only read about it. I have dozens of reels of 0805 LEDs lying around
>that would suffice for production for the foreseeable lifespan of this
>thing... I'll read up on that. Note that I need to read three at once,
>so it's a minimum of four, probably six LEDs. But still cheaper than a
>packaged device.

   These might not all be cheaper (especially not as cheap as using
LED's as sensors!), but they're in the same ballpark as your original
item.
   For a phototransistor you may be interested in a Sharp PT481, it
has a nice small "narrow-acceptance" characteristic so it only sees
through a narrowish cone in front if it, though it's thru-hole and I
haven't looked for an SMT version.

   The Rohm RPM-075 (found on a Freescale demo board) is SMT, but as I
recall has a much wider acceptance characteristic, though you may have
some "walls" between the three LED/sensor pairs to stop optical
crosstalk.

   The Vishay TCRT5000 IR LED/photoransistor pair (thruhole) works
well too. I have only small quantities of these things, as I don't see
a huge demand for line following robots...

>
>Thanks.