> Karel Kulhavy <[EMAIL PROTECTED]> writes:
>
> > I have heard that IrDA is not capable of sending full-duplex. My channel is
> > full-duplex (four separate simple optics heads) so it would be a great
> > pity to lose half of the precious bandwidth.
> >
> > Is there any possibility to force IrDA chips to work full-duplex?
>
> Yes, most of them can do that, but the problem is that the transmitter is
> placed so close to the receiver, that it "blinds" it by its own
> transmission. That makes the receiver less sensitive and can even make it
> recieve the nodes own transmissions. So in most chips, you can set a bit
> that makes it turn off the receiver while it's transmitting. IrDA devices
> negotiate whats called the minimum turn around time which is the time the
> receiver needs to get its sensitivity back after being blinded by the
> transmitter. I don't think it would help to separate the transmitting and
> receiving diode for such row range communication, since the signal will
> bounce back from the peer device (or other objects) anyway.
My optics is relatively well focused -- the beam has only about 1 degree
divergence angle and there are no obstacles in the path -- it's through free
air. I am also able to place the two tubes about one meter away fro each other
to prevent crosstalks. Do you think that in this case atmospheric dispersion
would make full duplex unavailable?
Which analog-to-digital chip (without optic diodes) and which packet chip would
you recommend as being widely available and reasonably cheap? I need of course
4Mbit-capable chip. :-)
As I understand to your explanation, did you mean that the anoalog-to-digital IrDA
chips are capable of full-duplex, but the packet chips not?
I have FIC VA503+ motherboard with "IrDA connector". Is the packet chip already
on my board?
Best regards
Karel Kulhavy
_______________________________________________
Linux-IrDA mailing list - [EMAIL PROTECTED]
http://www4.pasta.cs.UiT.No/mailman/listinfo/linux-irda
