Wind direction quantized to only compass points

[Mike Ramsey]

I'm running Weewx 3.8.2 on a Raspberry Pi with Rasbian Jessie.  Weather station is a Davis Vantage Pro2 wired.

I'm getting the wind direction reported at only the major compass points: (0,22.5,45,67.5,90,112.5,135,157.5,180,202.5,225,247.5,270,292.5,315,337.5).  This behavior started on Oct 14 but I can see periods in the archive where this also occurred in the past.  For example, on May 5, 2018 it did this for a 6 hour period.

What I expect, and what normally happens, is to report the full range from 0 to 360.  

A dump of the May 2018 period is below.

The Vantage Pro displays wind direction down to the degree even when weewx is recording the quantized direction.  This means the weather vane on the roof can measure and the data is getting down to the console.  Why is weewx reporting less accuracy than is being measured?

| 2018-05-05 00:00:00     | 26.8534531239994 |

| 2018-05-05 00:15:00     | 40.1762746405119 |

| 2018-05-05 00:30:00     | 31.8391669072265 |

| 2018-05-05 00:45:00     | 27.6576423981335 |

| 2018-05-05 01:00:00     |               45 |

| 2018-05-05 01:15:00     |                0 |

| 2018-05-05 01:30:00     |                0 |

| 2018-05-05 01:45:00     |                0 |

| 2018-05-05 02:00:00     |                0 |

| 2018-05-05 02:15:00     |            337.5 |

| 2018-05-05 02:30:00     |              315 |

| 2018-05-05 02:45:00     |              315 |

| 2018-05-05 03:00:00     |              315 |

| 2018-05-05 03:15:00     |              315 |

| 2018-05-05 03:30:00     |              315 |

| 2018-05-05 03:45:00     |                0 |

| 2018-05-05 04:00:00     |              315 |

| 2018-05-05 04:15:00     |              315 |

| 2018-05-05 04:30:00     |              315 |

| 2018-05-05 04:45:00     |                0 |

| 2018-05-05 05:00:00     |                0 |

| 2018-05-05 05:15:00     |                0 |

| 2018-05-05 05:30:00     |                0 |

| 2018-05-05 05:45:00     |             22.5 |

| 2018-05-05 06:00:00     |               45 |

| 2018-05-05 06:15:00     |               45 |

| 2018-05-05 06:30:00     |             67.5 |

| 2018-05-05 06:45:00     |             67.5 |

| 2018-05-05 07:00:00     |             67.5 |

| 2018-05-05 07:15:00     | 54.2728729207129 |

| 2018-05-05 07:30:00     | 59.2775430647446 |

| 2018-05-05 07:45:00     | 62.1555767325972 |

[Thomas Keffer]

Hello, Mike

Wind direction downloaded from the logger is limited to 22.5 degrees of precision.

Previous to V3.8.2, wind direction was always calculated in software, even when hardware values were available, which could result in higher precision (although not necessarily accuracy). See issue #336. 

Did you by any chance upgrade to V3.8.2 on 14 October?

-tk

--
You received this message because you are subscribed to the Google Groups "weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to weewx-user+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Mike Ramsey]

The station is at a vacation home and I was there on Oct 14 so it is possible that I upgraded to 3.8.2 at that time.  But it's curious that I see periods as short as 6 hours of similar behavior before the upgrade.  Maybe the software algorithm couldn't calculate higher precision with the data available at the time.

I plot my wind vector with a 90 degree East offset.  I noticed the new behavior when there was a prolonged period of steady wind made the plot look like a picket fence.

Really disappointed at Davis for dropping resolution within their hardware.

Thanks for the prompt answer!