Hi Charles..

Some comments on your article.

- Using a voice repeater is in a sense like using a bit regenerative
device albeit with no intelligence. The great thing about it is that it
avoid the hidden transmitter problem as repeaters are often built in
high places. Straight digi's or popular nodes on top of hills get messed
up really quickly when one station cant hear another and transmits on
top of them. We use to have a system in NSW that was continually getting
into packet crash problems because it was badly designed. With one user
at each end of the circuit it was great but anymore than that and it was
an abject failure.
- I would suggest that the repeater stays on during both sides of the
data exchange. This means that you only need to have a large txdelay
once and then almost none at all while the exchange is going on. You
will of course have to have a CTS squelch that works not on channel
quieting but on audio or a partial decode of the start framing sequence
- Long PACLENs will always help provided there is little or no dropout
(ie a good strong signal). Large MAXFRAMEs wont really help at all
though if you keep the repeater running.
- If you want to keep keying the repeater on and off think about a
PACLEN way in excess of what is common. I use to run 768 bytes in kiss
mode on a TNC2 box (Using KA9Q NOS) with MAXFRAME 7. At 1200bps I use to
get about 120cps with the TX being keyed up for 30-40 seconds at a time.
- Obviously one has to allow for the repeater timeout and restart. I
dont know the regulations over there but perhaps you can keep it
transmitting data as long as you like. I also wonder what a mess the
repeater access tone will make of the data.
- Using a voice repeater is also something that would work well for
RDFT/Wyman multicast you have already mentioned. I keep hearing that a
faster implementation (wider b/w) is due and since the protocol is being
written even now it may be worthwhile figuring a way to use repeaters.
- If you want to use NewPSK it is worthwhile making the Tunelength close
to zero. Since you are using FM there is no real frequency error to
worry about. I'd also fiddle with increasing the bit rate on an FM
channel (2500 works for an SSB one) (I assume that the W32 soundmodem
version of Tom Sailor's code has that adjustable. It is on the Linux
version)

Tnxs for your article

Bob VK2YQA

Charles Brabham wrote:
http://www.uspacket.org/mixmode.htm

Charles Brabham, N5PVL