Cumulus Support

I'm not sure that would work any better.

Take 4 intervals at 50%, 80%, 90%, 65%

Currently that would count 10 minutes of sunshine (0 + 5 + 5 + 0).

Taking the mean would also count 10 minutes (0 + 0 + 5 + 5).

But we know there was *some* sunshine in two five minute blocks, and we rule out one because there was no sunshine in the previous 5 minutes and another in because it was previously "sunny". All we really know is that there was somewhere between 8 seconds and 10 minutes of sun in that 20 minute period.

The problem we cannot get around is that we just do not have the data to do anything other than make guesses at what happened in reality.

Do you have a log with data capture enabled for the last catch-up say, then we try and work out what the effect would be?

I'll set up some logging to see if we can work out what if any options might be possible. This is not critical for me as I said I run live systems 24x7 so I'd only use catchup for short periods of probably no more than 5 or 10 minutes. It would be nice if it could be improved though.

Stuart

Stuart, on the previous solar topic, is it possible you extract the WD theoretical values for day?

I'll then compare with the MX values and see if there is anything appreciable difference.

Mark I have setup a custom log file in WD which will catch the solar data, date time actual reading expected max for time percentage and description.

Stuart

Mark I have started to capture the solar data for today from WD. I have taken an extract from it so far today as I wanted to compare the expected max solar value from WD against my script which is available online to make sure my figures in that are inline with what WD is producing. So far today the script produces figures inline with WD to +/- 2 w/m² for each minute which means the script does largely agree with WD. The script is at https://www.stella-maris.org.uk/maxsolar1.phtml. I will let you have the full data from WD after sunset.

Stuart

Mark I have sent you an email with two spreadsheets containing the data you requested from WD and my CMX database.

Stuart