Lichen Wang wrote :
> Jean Tourrilhes said:
>
> > Ultra is much more power efficient. In fact, if we could do
> > Ultra at 115 kb/s, we could save even more power...
>
> Ultra uses a single data rate so that you do not need
> to go through the lengthy negotiation to agree upon
> what data rate to use. It chooses 9.6 Kbps as this
> single data rate so that the majority of micro-
> controllers can handle it. I think introducing a
> 115.2 Kbps Ultra is counterproductive.
I know. That's why I said "if we could".
In fact the main reason is for having both normal IrDA and
Ultra operating at the same time (Our Linux box can listen to both
IrDA nd Ultra devices). The basic rate of IrDA is 9.6 kb/s, and as no
IrDA transceiver can auto-detect the rate of incomming data, Ultra
need to be the same.
> Actually, 115.2 Kbps needs more power than 9.6 Kbps.
> In both speeds, you only need to turn on the IR for
> 1.6 usec to transmit a bit of 0. Thus the energy per
> bit is the same. Higher data rate needs to dispense
> the same amount of energy at a higher rate, and thus
> needs higher power.
Depend how you see the problem...
For us, the problem is to send a specific amount of data
(let's say, 100 bytes). It's not a continuous flow of data, just a
single transaction.
For speeds between 9.6 kb/s and 115 kb/s, as you mention it,
the energy per bit is constant, so it we need the same amount of
energy to send those 100 bytes of data, regardless of the speed.
Now, that was only for the energy sucked by the Ir
transmitter. If we transmit things slower, our processor need to be up
a longer period of time, whereas if it transmit faster, we can shut it
down earlier. That's why transmitting faster can save us power ;-)
> Lichen Wang
By the way, any comments about the quality of Ir tranceiver ?
We had trouble with every Ir transceiver we use in presence of
fluorescent light and outside. The medium is detected as busy, packet
are corrupted and all that jazz... Your dongles tend to be slightly
better (but still so easy to confuse), the absolute worse being the Ir
transceiver fitted on a popular WinCE device just released by a
manufacturer that I don't want to name (a bit of sun, and gone).
With radio stuff (802.11, GSM), most of the cost of the
hardware in in the saw filter and other active filters, which reject
frequencies from adjacent bands and keep only the energy present in
the desired band. Those filters tend to be pretty tight and efficient.
With Ir, it seem that all manufacturers don't use any
filtering of the incomming signal (the receiving led is connected
directly to the digital) and don't really care about this...
Anyway, have fun...
Jean
_______________________________________________
Linux-IrDA mailing list - [EMAIL PROTECTED]
http://www4.pasta.cs.UiT.No/mailman/listinfo/linux-irda
