Could Cumulus read the currdat.1st file

[Axelvold]

I'm using WUHU to create the currdat.1st file and now I'm wondering if it's posible to have Cumulus to read this file like WD and VWS?

[steve]

I had a look at the WUHU page, and I think the file is actually called 'currdat.lst' (letter l rather than digit 1)? But I can't find a specification of the contents of the file - could you tell me where I can find one, and what it is that you'd like Cumulus to do with it?

[Axelvold]

I have attached a copy of the currdat.lst file.

What I want Cumulus to do is read that file and show the values like if It whas conected directly to my station.

[steve]

Thanks, it looks reasonably self-explanatory, and it's in an 'ini file' format which makes it very easy for Cumulus to read. It would still be useful to see a specification of the format of the file. How often does it get updated? Is it just for La Crosse stations - it mentions Heavy Weather? I'm a bit confused as searching the web seems to suggest that the currdat.lst file is actually produced by Heavy Weather, and WUHU reads it?

Would you like to create an enhancement request for this, so it doesn't get forgotten?

[Axelvold]

The currdat.lst file is created by HeavyWeather as well but if you use WUHU you dont need to run HW.

It is created in realtime, if there is an update in the station then the file is updated as well.

Yes it's just for LaCrosse stations

Would you like to create an enhancement request for this, so it doesn't get forgotten?

I have done this.

[casacota]

Just in case this is useful, I've put online the currdat.lst generated by my station (WS-3600 from La Crosse), updated every minute:

http://www.casacota.net/currdat.lst

The La Crosse stations have often a problem with data spikes with are making they almost unusable in some cases. I've written a small utility wich does filter the spikes, taking the currdat.lst generated by Heavyweather and creating another one with corrected values:

http://www.casacota.net/teranyina?num=1213088369

Reading currdat.lst would make possible to run WS-3600 with Cumulus also. If you want, I can give the (very poorly written) source code from the filter for Cumulus.

[steve]

Reading currdat.lst would make possible to run WS-3600 with Cumulus also. If you want, I can give the (very poorly written) source code from the filter for Cumulus.

That sounds interesting - perhaps you could tell me the algorithm you're using, rather than giving me the source code?

[casacota]

Reading currdat.lst would make possible to run WS-3600 with Cumulus also. If you want, I can give the (very poorly written) source code from the filter for Cumulus.

That sounds interesting - perhaps you could tell me the algorithm you're using, rather than giving me the source code?

Allready PM'd.

For those interested: the utility reads currdat.lst every minute, and keeps for each parameter the las 5 values in memory. Each parameter has its own threshold set for the maximum change expected in that amount of time (1 minute). If a reading goes over that threshold it is ignored and replaced by the last good one. Only if the value persists more than 5 minutes it is taken as valid.

Also, the utility generates a data2.csv in VWS format, wich can be used to feed data to Awekas.

[steve]

For those interested: the utility reads currdat.lst every minute, and keeps for each parameter the las 5 values in memory. Each parameter has its own threshold set for the maximum change expected in that amount of time (1 minute). If a reading goes over that threshold it is ignored and replaced by the last good one. Only if the value persists more than 5 minutes it is taken as valid.

Right - this is basically what Cumulus does with temperature, except that it doesn't have the "accept after 5 minutes" part. I want to extend this to other readings eventually, and introduce some sort of averaging perhaps, but doing it for wind speed is quite a challenge...

[casacota]

For those interested: the utility reads currdat.lst every minute, and keeps for each parameter the las 5 values in memory. Each parameter has its own threshold set for the maximum change expected in that amount of time (1 minute). If a reading goes over that threshold it is ignored and replaced by the last good one. Only if the value persists more than 5 minutes it is taken as valid.

Right - this is basically what Cumulus does with temperature, except that it doesn't have the "accept after 5 minutes" part. I want to extend this to other readings eventually, and introduce some sort of averaging perhaps, but doing it for wind speed is quite a challenge...

Yes, but on the other side: wind readings on Home Weather Stations (also Davis Pro) are not very accurate, due to the construction and sampling rate. There are some comparisons being made, eg. at the Servei Meteorològic de Catalunya (Meteorological Office of Catalonia, http://www.meteo.cat) a Davis Pro anemometer is running together with (much more expensive) professional units, and the values are quite different, specially the gusts are randomly too high. The cup (Savonius) anemometers are good for windrun measurements over relatively long periods, but not for instant speed nor gust measurements. I'm sorry but the wind values from such stations are relatively irrelevant.

Also, the La Crosse (and other) units tend to produce high wind spikes, eg. 92 km/h. Such a wind spike makes the wind "measuring" completely useless. It is better to have a delay in wind values (after 5 readings sudden high values will be accepted anyway) than unsure and completely wrong readings.

I have myself a WS-3600 running since 2004 here http://wx.casacota.net with that software spike filter and the wind readings differ not very much from the official station. The main difference is that the official anemometer is on an 10 meter pole and that from the WS-3600 only 2 meter high (for agrometeorological calculations)

Here you can see the real problem: apparently much 0 readings, this comes from the La Crosse sampling rate - only every 90 seconds or so - totally useless anyway: