Datalogger workings?

[George Richardson]

Steve,

This is in no way a complaint or suggestion to change Cumulus, but merely a request for information.

I run a Davis VP2 wireless on a dedicated weather computer running StartWatch, VirtualVP, VPLive, WeatherLink, and Virtual Weather Station. On a remote computer, connected over a wireless LAN via TCP/IP, I run StartWatch, VirtualVP, VPLive, WeatherLink, and Cumulus.

Last night at 3:00AM we had a power failure. My UPS carried my weather computer until 5:00AM (my remote computer is shut down over night). I restarted both computers at 8:30AM (Power had finally been restored) with no problem. At this time my Datalogger is set to record at a 30 minute interval. Cumulus is reporting an overnight low of 20.7 degrees at 7:30AM (Which is what I expected). What is unexpected is that both WeatherLink (both weather computer copy and remote copy) and VWS are reporting 20.6 degrees at 7:18AM. Would you have a layman explanation for this different report?

Thanks Again.

George Richardson

[steve]

From the 0718 timestamp, given that your logger is set to 30 minutes, the others must have got the minimum figure from the station (from its highs and lows section). Cumulus doesn't use that, so strictly speaking they are correct and Cumulus is not.

Cumulus doesn't use the highs and lows from any weather station, it works out all the highs and lows itself from the actual data. This is because of the 0900-0900 option, where the highs and lows in the station will be wrong. I didn't want to have to write and support two completely different sets of code. This does mean that there is a small margin of error when data from the logger is involved, and the longer the logger interval, the greater the potential for this error to occur.

Steve

[George Richardson]

Steve,

So, the information from the ISS passes through the Datalogger to the console then onto the computer if it is on. The Datalogger takes a snapshot of that data stream at whatever interval the Owner sets? The console uses the data that is arriving continously and may easily store data for high/low marks that change before the datalogger slice arrives?

Normally, I don't care how something works, just that it does. I had a problem a few days ago that you diagnosed as a possible conflict between datalogger snapshot and Cumulus retrieval. I had DL at 30 mins and Cumulus at 5 mins. You recommended setting both to 5 minutes. My concern was that a change of the datalogger interval might affect 2 working programs (WL & VWS) unless I could figure out how to change the interval in those programs. At this time I have not been able to discover how to change the Datalogger timer much less whether it would affect WL & VWS. Therefore I'll continue to have all set at the 30 minute interval, not because I want that time frame but because changing it might adversly affect something else.

Thanks for your continued support.

George Richardson

[steve]

So, the information from the ISS passes through the Datalogger to the console then onto the computer if it is on. The Datalogger takes a snapshot of that data stream at whatever interval the Owner sets? The console uses the data that is arriving continously and may easily store data for high/low marks that change before the datalogger slice arrives?

That's how I understand it - except that I think the data goes to the console which monitors and displays the data, storing highs and lows etc, and then snapshots the data to the logger at the required interval.

My concern was that a change of the datalogger interval might affect 2 working programs (WL & VWS) unless I could figure out how to change the interval in those programs. At this time I have not been able to discover how to change the Datalogger timer much less whether it would affect WL & VWS. Therefore I'll continue to have all set at the 30 minute interval, not because I want that time frame but because changing it might adversly affect something else.

I know almost nothing about VWS, but I know that the logging interval in WL is the same as the datalogger interval, because the way that it works is that periodically, WL downloads the data from the logger and copies it to its own log files. But it must also be downloading the highs and lows somehow. Cumulus is different in that it logs its own data in its own format independently of the data logger - it doesn't even use the data logger data after the initial download to catch up at start up. The main reason for this is that Cumulus has to cater for different types of weather station, so can't do it the simple way that WL does it. I imagine that VWS works in a similar way to Cumulus (in respect of its data logs).

Steve