I have been using the wonderful service from 86.82.68.237:10000 that streams I/NAV data as NavMonMessage protobufs and have noticed some interesting problems. The first issue is that the TOW for word type 16 is off by 14 seconds 50 % of the time, this is described in #162.
The other observation is that the stream is missing a large number of page pairs. Digging into this a little I found some interesting patterns and it appears that the pages that are missing are highly consistent and are a function of:
Whether or not the satellite has the new I/NAV improvements
Whether or not the satellite is transmitting OSNMA
From my analysis it seems the following table holds:
OSNMA?
I/NAV Improv?
Missing WTs
N
N
1, 3, 2, 4
N
Y
1, 3, 2, 4, 17, 18, 19, 20
Y
N
-
Y
N
-
I have also observed that PRN E01 appears to transmit WT5 at SF index 7 and 14, when we expect WT16. But I'm not sure if E01 was undergoing some testing/upgrades at that time.
In short, it appears that when the satellites are not transmitting OSNMA then whatever process is driving the live stream on 86.82.68.237:10000 filters out their messages for WTs 1, 2, 3, 4, 17, 18, 19 and 20.
Note that WTs 17, 18, 19 and 20 appear only when OSNMA is also present.
Note also that when I run ubxtool locally (built from current master) I do not see any missing page pairs - so the issue is presumably with whatever filtering is applied to consolidate all the source streams.
Here are some plots of the received word types at each subframe index for the four rows of the table. These were received from 86.82.68.237:10000 over a short three minute interval on 2023/03/29.
I have been using the wonderful service from
86.82.68.237:10000that streams I/NAV data asNavMonMessageprotobufs and have noticed some interesting problems. The first issue is that the TOW for word type 16 is off by 14 seconds 50 % of the time, this is described in #162.The other observation is that the stream is missing a large number of page pairs. Digging into this a little I found some interesting patterns and it appears that the pages that are missing are highly consistent and are a function of:
From my analysis it seems the following table holds:
I have also observed that PRN E01 appears to transmit WT5 at SF index 7 and 14, when we expect WT16. But I'm not sure if E01 was undergoing some testing/upgrades at that time.
In short, it appears that when the satellites are not transmitting OSNMA then whatever process is driving the live stream on
86.82.68.237:10000filters out their messages for WTs 1, 2, 3, 4, 17, 18, 19 and 20.Note that WTs 17, 18, 19 and 20 appear only when OSNMA is also present.
Note also that when I run
ubxtoollocally (built from current master) I do not see any missing page pairs - so the issue is presumably with whatever filtering is applied to consolidate all the source streams.Here are some plots of the received word types at each subframe index for the four rows of the table. These were received from
86.82.68.237:10000over a short three minute interval on 2023/03/29.