RTK/GPS Injection via NTRIP Issues

I am having problems using RTK/GPS Injection via NTRIP. If I supply credentials to our local service I seem to get a connection. The input data rate shows ~1Kb/s but the output data rate remains zero. Valid messages are seen but very intermittently and obviously the vehicle does not achieve an RTK Fix.
The interesting thing is, if I connect using rtklib’s str2str using the same credentials and have it setup a TCP server forwarding the same data over a local port, I can connect quite happily in MP using the ‘TCP Client’ option. In this case, everything seems to work as it should with reasonable in/out data rates, the expected message counters ticking over and ~30 odd satellites show up. Most importantly, I obtain an RTK fix on the vehicle.

It seems that there may be an issue with MPs NTRIP implementation?’

Here is my connection string: ntrip://xxx:xxx@positionz-rt.linz.govt.nz:2101/TRNG00NZL0

NTRIP connection:

TCP Client connection using the same source (via str2str):

Here is approx 30s of RTCM data from this source (RTCM 3.2):
LINZ_RTCM_TRNG00NZL0.log (50.3 KB)

tick NTRIP protocol v1.0 if you are using rtklib !

Umm - rtklib seems to work fine - the problem is that it doesn’t work unless it is connected thought rtklib over tcp… My understanding of str2str is that it doesn’t even parse the data but rather simply passes on the raw bytes between streams. In theory, MP should be seeing the same thing either way…

Just for completeness, I can confirm that checking “NTRIP protocol v1.0” makes no obvious difference and the behavior I have described remains the same.

I have a friend that needs a Windows only solution so have got around this issue in the mean time by using the Mavproxy ntrip module