[aprssig] ARPS physical layer
oh2mqk at sral.fi
Thu Sep 16 06:14:52 EDT 2010
On Thu, Sep 16, 2010 at 09:44:16AM +0100, Derek Love wrote:
> Thanks for the info guys, maybe I should be a bit more specific.
> I have the APRS 2.0 spec, I know about HDLC, NRZI, byte reversal, bit
> stuffing, CRC/FCS generation and stop/start (7E) flags.
> However, I come from a PMR / Data radio environment, so my first questions
> What's the preamble sequence and length?
> What's the sync sequence and length?
It is 16-32 bits of zero, which after NRZI means alternating tones and
gives maximun change of achieving bit sync. Then starts HDLC flag, and
the transmission ends with a HDLC flag or two.
> The preamble is usually a 1100 sequence of 16 or 24 bits to allow the RF
> sections to settle down and the receiving modem to set its levels up. The
> Sync sequence is usually 16 or 24 bits carefully chosen so they don't
> correlate well with Gaussian noise so that false detects are minimised.
> So far as I can see, there is no mention at all of a preamble and the only
> sync sequence I can see is the 7E start flag, which seems to be a very poor
> I can't see any specification for which state the NRZI coding process should
> be started with either......
Quite so. Continuous transmission landline links push HDLC frames one after
other without preambles. Ergo all specifications written for landline links
omit such a thing.
> I have also found a reference to a data scrambler somewhere, but no
> indication if it's part of the standard or not.
HDLC bit stuffing + NRZI act together as data scrambler. It does not
remove all DC bias creepage, but at least there is no infinitely long
sequences of bits with same modulation signal.
Google translator made scramble of my Finnish AFSK page.. so now in English:
For my Aprx program I have collected 3 byte-wise table lookup "CRC16" algorithms.
One of them is "CCITT-CRC".. Source is available under BSD style license.
> Derek Love
Best Regards, Matti Aarnio, OH2MQK
More information about the aprssig