The issue is not only at midnight. It occurs throughout the day whenever 
the wind speed reaches a new daily high.

Bob

On Thursday, March 9, 2017 at 10:08:15 PM UTC-8, gjr80 wrote:
>
> That would give you (in most cases) the same result. The issue is that 
> 'weewxDailyHigh' does not reset until weeWX saves the 00:00 archive record 
> to database which (by default) occurs at 00:00:15. After that 15 second 
> period everything is as it should be. The issue can certainly be fixed but 
> for a small issue that presents itself for 15 seconds at midnight I am 
> loathed to spend more time altering an already complex system of loop 
> stats. The new approach I am working on should do away with the 
> (continuous) reliance on the the weeWX daily summaries and I will wait 
> until I have that nailed down before dealing with this issue.
>
> Gary
>
> On Friday, 10 March 2017 15:55:52 UTC+10, tempus wrote:
>>
>> It seems that merely setting "wgustTM" in 'gauge-data.txt' to the highest 
>> of the daily-high-gust-value from weewx and the current rtgd "wgust" value 
>> would be an easy fix.  I am not fluent in Python, so this may not be 
>> technically correct, but in principle like this:
>>
>> IndicatedDailyHigh = max(weewxDailyHigh, rtgdTenMinuteHigh)
>>
>> -Bob
>>
>> On Thursday, March 9, 2017 at 2:49:06 AM UTC-8, gjr80 wrote:
>>>
>>> I observed again at midnight just gone, yes the daily high gust pointer 
>>> does take a number of seconds before resetting, this is due to the 
>>> (default) 15 second delay weeWX has from an archive record being generated 
>>> until the data is saved to archive and the daily summaries updated (rtgd 
>>> uses a combination of the daily summaries and the loop data since the last 
>>> archive record to determine daily highs). I am working on a better method 
>>> of keeping various stats (highs/lows/avg etc) in a loop environment for a 
>>> separate project, when I get that working I may roll the new approach into 
>>> rtgd which may improve this. In the meantime the delay can be lessened by 
>>> reducing the archive delay using the archive_delay option in 
>>> [StdArchive] in weewx.conf. Whilst the option can be safely reduced 
>>> from 15 how far it can be safely reduced will, I expect, be station and 
>>> system dependent.
>>>
>>> Gary 
>>>
>>> On Thursday, 9 March 2017 17:19:06 UTC+10, tempus wrote:
>>>>
>>>> Each time the wind reaches a higher value than previously reached since 
>>>> midnight, both the gauge-hand and the 10-minute-gust-indicator move 
>>>> simultaneously to the new high, but the daily-high-wind-gust-pointer does 
>>>> not. It remains at the previous daily-high until expiration of the current 
>>>> weewx 300-second archive interval when it then also moves to the new high.
>>>>
>>>> I will be traveling the next 24-hours and not able to respond quickly.
>>>>
>>>> -Bob
>>>>
>>>> On Tuesday, March 7, 2017 at 4:04:37 PM UTC-8, gjr80 wrote:
>>>>>
>>>>> I would say that this will quite often be the case.
>>>>>
>>>>> The daily high wind gust pointer should always indicate the highest 
>>>>> wind speed seen after midnight at the start of the day and up to and 
>>>>> including midnight at the end of the day. In other words it will reflect 
>>>>> the highest wind speed indicated by the gauge hand throughout the day. At 
>>>>> midnight the daily high wind gust pointer is reset to 0 and it will again 
>>>>> start tracking the highest wind speed seen throughout the new day. On the 
>>>>> other hand, the 10 minute gust value is the highest wind speed seen over 
>>>>> the last 10 minutes, it is not constrained by day boundaries. So consider 
>>>>> the first packet that arrives after midnight, say at 2 seconds after 
>>>>> midnight. Assuming its windSpeed value is >0, the daily high wind gust 
>>>>> pointer will move the this new value. Now if the 10 minute wind gust 
>>>>> value 
>>>>> is currently higher than this latest windSpeed (in other words the 10 
>>>>> minute wind gust occurred some time in the last 10 minutes of the 
>>>>> previous 
>>>>> day), the daily high wind gust pointer will not be affected by the 10 
>>>>> minute wind gust and nor should it (because it occurred yesterday). If 
>>>>> the 
>>>>> 2 seconds after midnight windSpeed value is higher than the 10 minute 
>>>>> wind 
>>>>> gust value, then the 10 minute wind gust value will be updated to this 
>>>>> new 
>>>>> windSpeed value and the daily high wind gust pointer will indicate the 
>>>>> same 
>>>>> value (as will the wind speed gauge hand).
>>>>>
>>>>> The real check for correct operation of the daily high wind gust 
>>>>> pointer just after midnight is to verify that it tracks the highest speed 
>>>>> indicated by the gauge hand after midnight. If the gauge hand indicates 
>>>>> say 
>>>>> 5mph and the daily high wind gust pointer indicates 4 mph then we have a 
>>>>> problem.
>>>>>
>>>>> Gary
>>>>>
>>>>> On Wednesday, 8 March 2017 02:40:39 UTC+10, tempus wrote:
>>>>>>
>>>>>> The only issue I have noticed since upgrading to v0.2.9 is that for a 
>>>>>> short time period following midnight the daily high wind gust pointer 
>>>>>> was 
>>>>>> below the 10-minute high gust indication.  However, I went to bed after 
>>>>>> that and haven't watched since.  Based on that single observation it 
>>>>>> appears that periodic daily high wind speed values from the weather 
>>>>>> station 
>>>>>> aren't being updated to include latest wind speeds.
>>>>>>
>>>>>> -Bob
>>>>>>
>>>>>> On Tuesday, March 7, 2017 at 4:20:34 AM UTC-8, gjr80 wrote:
>>>>>>>
>>>>>>> Well I observed your wind speed gauge over a couple of continuous 
>>>>>>> 10+ minute periods this afternoon, there were 5 occasions where the 
>>>>>>> gauge 
>>>>>>> indicated a speed higher than the present 10 minute gust and in each 
>>>>>>> case 
>>>>>>> the 10 minute gust immediately jumped up to the indicated speed. I 
>>>>>>> guess 
>>>>>>> this supports your observations under v0.2.9. I noticed a number of 
>>>>>>> changes 
>>>>>>> (daily max, gauges rescaling, windrun going to 0) at 6PM my time and 
>>>>>>> thought what? Then I realised it was midnight and change of day (I am 
>>>>>>> +10 
>>>>>>> UTC). So the chnage of day appears to works fine.
>>>>>>>
>>>>>>> Periods of lower wind speeds are more likely to include some 0 wind 
>>>>>>> speed packets than periods of higher wind speed. It's possible the old 
>>>>>>> method of calculating the 10 minute gust could have been upset by 0, I 
>>>>>>> don't believe that was the case but who knows. 
>>>>>>>
>>>>>>> Staring at a gauge for 10+ minutes is incredibly boring so I will 
>>>>>>> leave that for now, let me know if you notice anything else that does 
>>>>>>> not 
>>>>>>> appear right.
>>>>>>>
>>>>>>> Gary
>>>>>>>
>>>>>>> On Tuesday, 7 March 2017 17:24:58 UTC+10, tempus wrote:
>>>>>>>>
>>>>>>>> Sorry about the 403 Forbidden error. That is a test domain and 
>>>>>>>> public access is often blocked. I opened access a few hours ago when I 
>>>>>>>> saw 
>>>>>>>> your message, but didn't have time right then to post a response or 
>>>>>>>> upgrade. I have since upgraded to v0.2.9. You should now be able to 
>>>>>>>> see it 
>>>>>>>> working here:
>>>>>>>>
>>>>>>>> http://www.lablibrary.com/ss/newport.php
>>>>>>>>
>>>>>>>> This area recently had a several-day period of unusually light 
>>>>>>>> winds. The 'wgust" value anomalies occurred at those unusually low 
>>>>>>>> wind 
>>>>>>>> speeds. I don't know why functionality would be different at different 
>>>>>>>> speeds, but whatever the reason, v0.2.8 'wgust' anomalies occurred far 
>>>>>>>> less 
>>>>>>>> frequently today at higher speeds and I haven't seen them at all since 
>>>>>>>> upgrading to v0.2.9.
>>>>>>>>
>>>>>>>> Weather forecasts are not very reliable in this area, but there may 
>>>>>>>> be an opportunity to check higher-speed functionality tomorrow, 
>>>>>>>> because the 
>>>>>>>> National Weather Service has issued this High Wind Watch: 
>>>>>>>>
>>>>>>>> "*South wind 30-40 mph with gusts to 65 mph, strongest near 
>>>>>>>> beaches and headlands but also possibly affecting coastal communities. 
>>>>>>>> Winds spreading northward late Tuesday morning, with peak winds from 
>>>>>>>> about 
>>>>>>>> noon to 4:00pm or 5:00pm. These winds could cause tree damage that 
>>>>>>>> could 
>>>>>>>> lead to power outages, and will cause hazardous driving conditions.*
>>>>>>>> "
>>>>>>>>
>>>>>>>> My anemometer is 33-feet (*10 meters*) above ground, 204 feet (*62 
>>>>>>>> meters*) above mean-sea-level, and line-of-sight to the sea, so it 
>>>>>>>> should be spinning rather quickly if the forecast is correct. The time 
>>>>>>>> here 
>>>>>>>> is 8 hours behind UTC.
>>>>>>>>
>>>>>>>> -Bob
>>>>>>>>
>>>>>>>> On Monday, March 6, 2017 at 2:26:22 PM UTC-8, gjr80 wrote:
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On Sunday, 5 March 2017 06:11:53 UTC+10, tempus wrote:
>>>>>>>>>>
>>>>>>>>>> I have been traveling for a few days with only a SmartPhone to 
>>>>>>>>>> watch data. However, based merely on wind-speed-gauge indications 
>>>>>>>>>> there 
>>>>>>>>>> seem to be "wgust" value anomalies. The pink gust-area sometimes 
>>>>>>>>>> disappears 
>>>>>>>>>> even though there have been recent gusts into that speed-range. 
>>>>>>>>>> Furthermore, gusts above the indicated upper-gust-limit often do not 
>>>>>>>>>> cause 
>>>>>>>>>> the indicated gust-range to expand.
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>> I have been closely observing my wind speed gauge over the last 
>>>>>>>>> few days during various windy periods. I did notice periods where the 
>>>>>>>>> red 
>>>>>>>>> gust 'wedge' would disappear for a short period. I observed several 
>>>>>>>>> occasions where the indicated wind speed went higher than the 
>>>>>>>>> existing 10 
>>>>>>>>> minute gust value, on all occasions the ten minute gust value was 
>>>>>>>>> immediately increased and the red gust wedge similarly increased in 
>>>>>>>>> size 
>>>>>>>>> immediately to match the indicated wind speed.
>>>>>>>>>  
>>>>>>>>>
>>>>>>>>>> It is my understanding that the value of "wgust" is supposed to 
>>>>>>>>>> be the speed of the highest gust over the last ten-minutes.  If so, 
>>>>>>>>>> the 
>>>>>>>>>> ten-minute time-window should include recent "wlatest" values.
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>> That is my understanding as well and that is exactly what rtgd 
>>>>>>>>> does. The issue in this case was the method in which the ten minute 
>>>>>>>>> wind 
>>>>>>>>> gust value was derived from the ten minute wind speed list. This has 
>>>>>>>>> been 
>>>>>>>>> fixed in v0.2.9.
>>>>>>>>>
>>>>>>>>> I have no solution for the gusts you observed that did not update 
>>>>>>>>> the ten minute gust value. I have not been able to replicate this 
>>>>>>>>> issue and 
>>>>>>>>> I have reviewed the code and am confident it is correctly capturing 
>>>>>>>>> windSpeed data and calculating 10 minute gust values. I suggest you 
>>>>>>>>> upgrade 
>>>>>>>>> to v0.2.9 and see if the issue remains, if so we need to determine 
>>>>>>>>> the 
>>>>>>>>> conditions under which the issue occurs. If this is not possible 
>>>>>>>>> visually 
>>>>>>>>> then I have some wind speed debug code that will dump the windSpeed 
>>>>>>>>> history 
>>>>>>>>> list and other key values to log and we will need to wade through the 
>>>>>>>>> figures to determine the issue.
>>>>>>>>>
>>>>>>>>> I did go to your previously posted link to try to observe the 
>>>>>>>>> issues you describe myself but the link now gives me a 403 Forbidden 
>>>>>>>>> error.
>>>>>>>>>
>>>>>>>>> Gary
>>>>>>>>>
>>>>>>>>

-- 
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 [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to