<HTML><FONT FACE=arial,helvetica><FONT SIZE=2 PTSIZE=10>Andrew asks:
<BR>>What happens at the rcver ?
<BR>
<BR>The receiver breaks up the signal into 1200 chunks every second.
<BR>Each chunk is worth one bit. The receiver looks at the signal
<BR>during that bit's chunk and decides if the signal during that
<BR>bit's chunk/time period looks more like a 1200 Hertz sine wave
<BR>or looks more like a 2200 Hertz sine wave. Typically you run
<BR>a separate "bit clock" that breaks the signal into these one bit
<BR>sections for you. Note that you have to have your bit clock in
<BR>sync with the transmitter's so that you will know where every
<BR>bit starts and ends.
<BR>
<BR>Or you could use a phase locked loop, but that's a different
<BR>technique.
<BR>
<BR>Once you decide if this bit is a zero bit or a one bit, then in AX25
<BR>AFSK you have to un-NRZI and un-bitstuff, frame the bits up into
<BR>characters, find the HDLC flags, get the packet, find the trailing
<BR>HDLC flag, compute your received CRC for the packet and check
<BR>the CRC in the packet to see if the CRC's match.
<BR>
<BR>But I've simplified this a bit. :-)
<BR>
<BR>>Does it not check the distance between zero corssings
<BR>>to determine if the tone is 1200 or 2200 ?
<BR>
<BR>No. Usually not. Zero crossings are not usually used in
<BR>1200 baud Bell 202 AFSK AX.25 packet radio. In other
<BR>modulation schemes, yes, but not usually here.
<BR>
<BR>Now, you *could* use the zero crossing to measure the frequency
<BR>of the sine wave, and use that to decide the bit, but it typically
<BR>doesn't work as well as other DSP techniques. (You still
<BR>need a separate bit clock). The zero crossing technique was used
<BR>years ago by some of the "op-amp connected to a serial port pin"
<BR>programs which generates an interrupt to the PC which measure
<BR>the time between interrupts to determine the frequency. It will
<BR>work, but not very well.
<BR>
<BR>Douglas KA2UPW
<BR></FONT></HTML>