-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Wed, Jun 20, 2007 at 07:36:11PM +0200, wouter van ooijen wrote:
> Your prices seems too high. For 100+ I would calculate:
> 
> $0.50 - 10f200 SOT23-6 (maybe DIP is better..)
> $0.10 - led of some sort (a low-cost LED is more like 0.05)
> $0.01 - diode
> $0.10 - capacitor
> $1.40 - 1"x1" pcb, qty 100 (Olimex) (seems a bit large for just these
> components, and it is the largest cost factor)
> $?.?? - conformal coat of some sort
> $0.00 - student labour to assemble it all...

True, I was just using my usual suppliers, digikey and APCircuits.

> There is no transistor to short the power, so I assume this is one-way
> communication. If the voltage drop over the wire is no problem (let's
> say 5.2V at the source, no less than 3.0V at the destination) the above
> list should be enough. 

Why use a transister to short the power? Why not just use a IO pin with
a limiting resistor? (or without even if I'm brave)

> For this kind of application you should not worry about reliability of
> the message itself - reliability is by repetition. Who cares if one of
> the LEDs is on or off a little bit longer.

Exactly. Heck, I was thinking that given my reliability needs I could
leave out any sort of preample and just have the nodes constantly look
for their node id, once found, do whatever the following data bytes say.

The chances of the 32 bit node id appearing in the data bytes is small
enough that I'm not worried about it.

> Do not feed the LED from the capacitor, feed it directly from the power,
> so the capacitor can be smaller.

Good point, I'll do that.


Wait, no that won't work, when the soft 5k pull up is enabled on the
line, to allow the slaves to talk back, the LED's will effectively short
the line, no good. They'll have to be behind the diode for sure.

Then again I could rely on the LEDs being off when communication needs
to happen... but that would require the cooperation of all devices on
the line.

> For the brave: leave out the diode, feed the PIC via one of its IO pins,
> put the capacitor at the power pins.

Oh... like that note on the piclist website? Sneaky, anyone else
actually implemented that?

Then again, probably not a good idea for 20ma given that the LED can't
be directly on the line.

> But back to your purpose: you want to control a bunch of LEDs, connected
> in a star configuration, right? So you have 2 wires to each LED/slave?
> So you not connect them all to a big matrix and do without the local
> intelligence?

Er... yeah, I have kinda lost sight of that...

Initially in my discussions with the artists involved they wanted to
have distributed sensors as well, stuff like temperature, emf from cell
phones and sound. The idea is that the viewers in the space would
interact with the sensors, for instance from their body heat subtly
raising the local temperature or their cell phones emitting detectable
emf.

But that's not what they asked for the report on... So looks like I'll
give them two options, one a good deal cheaper than the other. :)

- -- 
http://petertodd.org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFGeXM33bMhDbI9xWQRAhgIAJ441IJpKIgcISeWf/04NXGFihSCEQCeJ5el
SawxWzEFS4PBjZerOoWzBuU=
=QNxa
-----END PGP SIGNATURE-----
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist