Maiana and SignalK - position in future, incorrect NMEA0183 checksum

[mrstas]

I've got a really weird behavior of how SignalK detects navigation.position from Maiana source. After powering raspberry pi up I am receiving proper position but after a while some weird stuff is happening. In DataBrowser navigation.position freezes for few minutes without any updates but then hour counter starts incrementing values and moving the date into the future. After some time position timestamp going back to current time and it starting to work correctly. But after few minutes process repeats again. During this glitch all connected devices stops receiving position and going crazy.
My setup:
RPi5 (openplotter4, Maiana plugin 4.1.0-stable, SignalK 2.11.0)
McArthurHAT
Maiana(hw 11.9.1 fw 4412)
few displays on NMEA2000 network

I could remember it was working just fine before.
I have updated SignalK and few other plugins recently to get Seatalk1 working.

Appreciate any help

[mrstas]

Is it possible that data received from maiana have incorrect checksum?
For example first line on my screenshot is "$GPGSV,4,4,14,3020,45,,,,0*54" where "*54" is a checksum,
but online calculator https://nmeachecksum.eqth.net/ gives different checksum "*4C"