@dsj my UV seems to be out by quite a lot still. Reading today was 5.2 and should have been around 7. We have had plenty of fine clear days so that shouldn’t be the issue.
Gary
1 Like
We’ll have a look. It can be more than just the number of clear days - sometimes the automatic analysis doesn’t have enough confidence to suggest a change, though often a human can cut through the noise.
@dsj we are cooking here today. UV reading 16 
2 Likes
Wow! What should it be reading?
Put up advertising! “Come get your suntan. Only takes 5 minutes!” 
Hmm, it looks like the humans (I was one of them) who manually set your UV cal a few weeks ago apparently didn’t do a very good job… Good news though: I just checked and the CL process has finally collected enough data to confidently suggest a change, and that was applied this morning. Please report what you find!
3 Likes
Hi @WFmarketing
I would like to know if there is a unit value being used to calculate the UV index. Australian BOM information shows 1index=25mW/m2:
Their forecast for my area today is 11:
and 10 here
http://www.bom.gov.au/uv/
But my Sky has never recorded over 9
The weather station down my street is currently recording maximums of 11 which match the BOM special UV forecast map today.
If the units are the same then my Sky is reading quite low with max values currently around 7.
cheers Ian
2 Likes
Interesting information. Has the sensor been identified here? I don’t remember seeing which one WeatherFlow used.
I found it. The Broadcom APDS-9200.
Thanks, Ian. Yes, 1 UV index is equivalent to 25 mW/m^2 of UV radiation in our model as well. The CL process for UV uses data from several sources, and is not perfect (yet!). The more data the system collects, the better the calibration gets. I can see in the logs that your UV sensor has received a total of four calibration updates just in the past two weeks, and each time it has led to higher readings from your station. It should continue to get better as it collects more data.
2 Likes
I seem to be having problems with the UV reading being too high. Readings seem to be almost twice what is forecast and what is reading on my Atlas a few feet away.
Is there a way to tell if the auto-calibration system is working for me?
George, yours is another where the CL system doesn’t have enough data to make an automatic adjustment to your UV calibration, but a human inspection does suggest a significant reduction is in order. We’ll apply a tweak tonight and you should see better data starting tomorrow.
Thank you very much!
1 Like
Looks like my Sky are under predicting by a fair bit. Is the CL process still working?
Thanks, Peter
How do the Skys compare with Solar Radiation and illuminationm
Peaks respectively for the 3.
UVI: 5.6, 5.5, 5.6
Lux: 93400, 90100, 93800
w/m2: 779, 751 , 781
Predicted UVI for my location about 12.
Measured Solar radiation at closest calibrated site was about 1000w/m2 (they quote values in Mj/m2 per hour so not sure of exact peak). Another private site reported. 987w/m2 peak.
Still reporting under here. Currently showing 6.8 when it should be around 8.8
Hi Peter. Yes, the CL process is still working but always improving. We’re analyzing your data to find and hopefully fix the issue. Thanks for your diligent reporting - please let us know how it looks later this week.
Hmmm. Your latest UV data shows a pretty strong shift from earlier data. Has anything about the environment changed recently? Have you moved the device lately?
@dsj Hi, All the stations around me were recording the same UV as me. I have not moved the sky unit but we have had a lot of cloudy days lately.
UV index at my station seems a bit high still after around 10 days. Station ID 5854, reading around 5-6 when local stations including NWS ones read 1-2. Except for the cloudy days it’s consistently a bit high.
Hi @vtyler98 . Appreciate the report! The CL system for UV requires a number of days with clear skies, especially during solar noon where we can compare daily max UV versus your SKY’s reported UV values in order to calibrate in-situ. We looked at your data. So far, only two days have qualified as valid clear days : 11 Dec and 17 Dec. Once the system can evaluate a few more days, it will automatically calibrate your UV sensor. Then it will rerun the data every day to ensure all data is precisely calibrated forever.