Weird wind problem again

[Mapantz]

Screenshot 2024-12-31 214431.png

WeatherLink shows 36mph as the highest gust in the last 10 minutes.

I can turn the logging on, but it's one of those things that you spot randomly, then you don't see it again for ages.

[mcrossley]

Since I made the change to not add missed wind gusts to the recent data - due to it potentially extending the 10-minute period by another 10 minutes - if CMX misses a broadcast packet with the high value in it, it will not be recorded as the current gust, but the gust will be included for daily record purposes.

Can't have it both ways I'm afraid.

What are your broadcast reception stats from the WLL like?

[Mapantz]

Another odd one today; I recorded a gust of 50mph, which CMX shows in the highs of today, but it was never shown on the graphs. That's the first time I have seen that.

As the graphs use #wgust and data is added every minute, you'd have thought that it would have been logged, as it holds the high for 10 minutes.

Screenshot 2025-01-06 121814.png

The high gust was 50mph at 03:36

[mcrossley]

Same answer as above, if MX missed the multicast packet with the gust in it, then it will now no longer appear in the graph data, or the current 10-minute gust value. If it exceeded the current daily high gust, then it will be recorded as the new daily high.

What are your broadcast reception stats from the WLL like?

[Mapantz]

I'm confused

CMX shows the 50mph gust as today's max, if the packet wasn't received, wouldn't it show 48?

Reception stats have always been good. Around -50 to -55db. Just the odd packet missed here & there.

It just seems odd that in all the years I have ran CMX, I have never seen the max gust of the day not get logged in the graphs.

[mcrossley]

Its because of the change I made a few releases ago to "fix" the problem you reported of the 10-minute gust value lasting up to 20 minutes.

The "current" data from the WLL can anything but - it is just a snapshot of the last packet that the WLL uploaded to weatherlink. Unfortunately the timestamp it send with the data is the time of the request, not the time the data was generated - not very helpful, I know when I made the request!

Originally if that "current" data contained a higher gust value, I added that current gust into the recent wind data - on the assumption that MX missed the broadcast packet (note that I mean the local network broadcast packet sent by the WLL, not the radio packets received by the WLL).

But if the "current" data from the WLL is slightly stale, then the gust could have legitimately timed out from the 10 minute window, and we just added it back in for a futher 10 minutes.

So, now the "current" value is not added into recent wind, but is compared against the daily values.

I may be able to tweak this some more, back dating the current data for instance, or using another value from the data set.

I miss broadcast packets from my WLL when I reboot my router, and very occassionally the WLL seems to stop for a minute or so, the nresume.

[mcrossley]

The reason I keep asking about your multicast packets is because all this processing only occurs if CMX misses multicast packets, and the current gust value is higher than today's highest so far.

I'm never seeing these wind "funnies", but my network reception is virtually 100%. It implies to me your local network may be dropping quite a lot of the multicasts.

[Mapantz]

My multicast stats are always 100%.

Before I upgraded today, there were zero bad packets after 24 days of CMX uptime

Just looked in the previous log:

2025-01-06 14:31:29.247 WLL: Percentage good packets received from WLL 100.00% - (0,888246)

[freddie]

How do you measure receipt of multicast network packets? They are UDP, hence will be "fire and forget". So you have no way of knowing if one has been missed.

[Mapantz]

How do you measure receipt of multicast network packets? They are UDP, hence will be "fire and forget". So you have no way of knowing if one has been missed.

The WLL does the logging, and provides health updates freddie