Re: Overlong RTCM3 packets from EUREF-IP

[Greg Troxel]

Marek Szuba <scriptkiddie@wp.pl> writes:

> Having accidentally found out during my struggle with the GNSS time
> source that gpsd supports NTRIP, I decided to try integrating it into my
> poor man's WiFi signal-strength mapper in place of the separate client I
> used to use earlier. My NTRIP caster (euref-ip.net) uses RTCM 3.3 but no
> problem here, a quick look at gpsd-3.21 source code showed that it
> should be able to handle this version. What HAS turned out to be the
> problem is that I immediately began to get "overlong RTCM3 packet"
> errors... The largest packet I have seen so far was 590 bytes long,
> which indeed is larger than the hard-coded limit of 512 bytes.
>
> Unfortunately I do not own a copy of RTCM 10403.3 so I do not know what,
> if anything, the specifications say about maximum packet length. That
> said, having randomly sampled the source code of various NTRIP clients
> and servers listed at https://igs.bkg.bund.de/ntrip/download (under the
> assumption that BKG of all people should know how to handle NTRIP
> correctly) it seems 2048 bytes is the most common value for this.
> Indeed, replacing 512 with 2048 in the source code seems to make
> gpsd-3.21 handle EUREF RTCM data just fine.

By 'handle', do you mean

   gpsd 3.21 with that change connects to the EUREF caster at some RTCM3
   mountpoint

   gpsd sends the RTCM3 to some receiver (F9P?)

   the receiver is then able to do RTK successfully

> What do you think about about increasing RTCM3_MAX to 2048, guys?

I'd say if we don't have a footnote justifying 512 :-), and BKG uses
2048, and you are seeing 590, it's entirely reasonable to flip to 2048.

signature.asc
Description: PGP signature