Cumulus Support

With TS PAM working it's way down our East Coast (we have been relatively unaffected by this - not our strongest wind speed/gust nor our highest rain fall/rate), I decided to move the test VUE I'm using with MX from inside to outside.
This is the first time I have had the full suite of incoming live data to MX having always copied over the data from C1 (VP2) and then not had wind or rain live data (VUE inside).

This has shown up a couple of anomalies in MX that are not in C1.
1. <#r24hour> is always 0.0
2. <#wchill> is always exactly the same as outside temperature.
The console shows different values.
Realtime shows different values
There is no difference if the Davis value is used or MX calculates the value.
The dashboard changes both temp and wchill at the same time to the same value
The trends chart tooltip shows the same value at the same time for both.

None of this makes a lot of sense to me but is what I am seeing. The temps are in the high teens / low 20's and the wind is not strong so windchill will not be a lot different but this difference does show up on the console and in realtime.

The wind chill web tag is working fine for me - it's on the standard index.htm page. Wind chill only applies below 10C, so that's why you're seeing the same value as the current temperature. I can't see any way that the realtime.txt value could be any different, so I assume you have simply made a mistake in that respect. Note that Cumulus MX always calculates the wind chill for Davis stations as the LOOP data does not have the wind chill. The LOOP2 data does have wind chill, so it's possible that I could write the code to use that eventually, but there's not much point as Cumulus does it correctly anyway.

The code which calculates the rain in the last 24 hours (which is then supplied on the <#r24hour> web tag) looks OK to me. I've faked some rain on my indoor WMR200 test system, and the <#r24hour> web tag reflects it correctly. We're expecting some real rain today, so I'll look at my real web pages from my VP2 when it rains. Perhaps in moving your Vue you have caused some anomaly with the rain counter which will need to pass through the system. Your VP2 page here - http://silveracorn.nz/cumulusmx/ - shows the figure, and the code is exactly the same.

Thanks for looking into this and for your response.
Wind Chill;
I was working from the Davis description (AN_28) where it appears to indicate that it will always = temp above 34degC and may = temp between -18C through 34C if the wind speed is low enough.
I did not realise it was not reading from the station which was showing wind chill or that MX has a cut off at 10degC rather than 34C. So I presume it is working as intended.

r24hour
Have just come back to the PC and it is now over 24 hours since the rain started. r24hour is now showing data. So it appears MX does not begin showing data until it was 24 hrs since rain began (after how long a dry spell)?
This is not the same as C1 which was showing r24hour data this morning when MX was not. It began raining about 28 hrs ago and stopped about 12 hrs ago after 6 days without rain.

BCJKiwi wrote:r24hour
Have just come back to the PC and it is now over 24 hours since the rain started. r24hour is now showing data. So it appears MX does not begin showing data until it was 24 hrs since rain began (after how long a dry spell)?

MX shows the figure as soon as it rains, as I indicated above, as long as it is running at the time of the rain. If there really is a problem in MX in this area, perhaps it is that it doesn't initialise the last 24 hours of rain counter figures when it starts up. I'll look into that. I'm only guessing, though, based on the information you've given about the circumstances.

Edit: I restarted my test system and it still shows the correct 'rain in last 24 hours' figure. So I suspect that as I suggested above, you have had an issue with the progression of the rain counter, perhaps due to moving the station. You could perhaps look at the rain counter figures in the log file to see what they have been doing. The code (in both C1 and MX) works by comparing the current rain counter with the value 24 hours ago.

I'll keep the VUE outside and see how it goes on the next rain session - expecting more over the next few days.

MX and the VUE were shut down and the VUE moved outside. However had also just copied over the data from C1 to get the latest historical data into MX.

The rain began as I was doing this and C1 may have measured the first tip while this was going on and a first tip may have been in the data copied across but I thought not.

Will confirm behaviour on the next rain cycle.

Can confirm MX is reducing the r24 hour number as time passes in sync with C1.

Still puzzling over the wind chill;

Wind chill only applies below 10C

C1 is happily graphing wind chill at 25 deg C so am wondering why MX now has a cut-off at 10 deg C when Davis (via NWS) does not cut off until 34 deg C?

The standard formula which Cumulus uses only applies below 10C (as the Davis document mentions), and both MX and Cumulus 1 apply this. If you are seeing wind chill in C1 with temperatures above this, then either you are using the Davis figure (from the DLL, which presumably does something similar to the console firmware) or you have applied the ini-file override UseWindChillCutoff=0.

Perhaps I will implement the override in MX at some point, even if it is 'wrong'. And the concept of 'wind chill' at 25C or 34C doesn't make sense to me.

Interesting.
I was quoting Davis AN-28. Can find no reference in any Davis documentation to 10C cut off. Davis quote the NWS formula and so I went to NWS who have a table and a calculator. The table stops at 40deg F and 5mph.
Their calculator says temps need to above -50F and below +50F, or above -50 C and below +10 C.
NWS also say wind chill is only defined for temps at or below 50F and above 3mph.

So I have no idea why Davis has significantly different ranges and the consoles (both VUE and VP2) show wind chill different from temp well above the NWS 10C cut off.

I can't see anything in C1 regarding wind chill settings in cumulus.ini so do not have a "UseWindChillCutoff=0". In C1 configuration/station, calculate wind chill is NOT ticked, so I assume it is getting it from Davis DLL. The C1 graphs have routinely shown wind chill at our normal mild temps as they do right now.

I have been back over the C1 and MX setups.
MX realtime.txt is definitely showing a small amount of wind chill but the tags are not. Some times wind chill is actually higher than temp which makes no sense either.

I think this may be due to a rounding issue.
MX realtime.txt is always reporting wind chill as whole numbers (i.e. with no decimal component).
Accordingly it may therefore be plus or minus on the temp if rounded to the nearest whole number - i.e. rounding up or down.
C1 reports to 1 dec for both tags and realtime.txt.

I built a routine in the PCs to log both realtime.txt and a tag version of realtime.txt to a log file for both C1 and MX all at our realtime interval of 15 secs.
Have included a sample from the files to make it a little easier to see what is happening.
The MX is reading off the VUE and is lower and not seeing much wind. The temp is slightly different and the chill much less than for the VP2.

BCJKiwi wrote:I think this may be due to a rounding issue.
MX realtime.txt is always reporting wind chill as whole numbers (i.e. with no decimal component).

Yes - I've used the decimal places for wind rather than temperature; I'll fix that in the next build.

Just to confirm that r24hour has been working as expected with the station outside and no data copy from C1.

Also confirm that the windchill in realtime.txt is now correct in b 3022 Another item to cross off the list

Thanks.

Hi Steve,
Sorry to open the thread again, however I am experiencing issues with CumulusMX r24hour (rainfall in last 24 hours).

I have used this tag for a long time now with Cumulus1 without issue.

I run my Meteo day (midnight to midnight) though as I'm a registered Australian BoMeterology Manual Rain Gauge Station, I need to capture Rain Fall over last 24 hours to 9am each day for comparison purposes against the Manual Readings.

I have scripts run by cron which grab r24hour from CUtags.php (originally fed by C1 (now CMX).

The mm rain being recorded by r24hour in MX is weird and I can't figure out where it is obtaining it.

Info: I am running DavisPro2(+) and VVP is allowing me to run both C1 and MX simultaneously. (running latest MX (3.0.0-b3029))

I have recently (with help of another forum user) upgraded my site to PHP and using 90 odd percent data from MX.

The balance of data is from C1 (purely for MySql old type as I have not had time to convert over to MX / MySql).

I run cross checks between C1 and MX data uploaded to server and this r24hour seems to be the main issue.

Currently I'm having to upload r24hour from C1 specifically for the purpose mentioned above. My intent is hopefully once everything seems OK to shut down C1.

You can see the dif in r24hour in following live updates.
c1 : http://hrvistaweather.com/Cumuluswebtags.php?sce=view
MX : http://hrvistaweather.com/weather/CUtags.php?sce=dump

At time of post the MX was 1.8mm and C1 was 0.6mm (C1 0.6mm is physically correct).
If needed, this link displays Rainfall from MySql (C1) for last seven days in 1 minute intervals. http://hrvistaweather.com/graphs_recent_highslide.php
Not sure why or what I can do about this issue.

One last question , can Cumulus Toolbox be used with MX once I finally shut down C1.

Kind regards,

Hi Steve,

I'm not sure but I may have stumbled upon either a bug or an issue with my CMX set up.

With ref to the issue in my previous post (r24hour incorrect in MX).

I could not logically see why it was returning what it was, so time to reverse logic.

The fact it must be looking at a different time period than last 24hours being apparent. The only thing I could not logically think of that may affect that was the Graph Hours. Since initial earlier versions installation of CMX I have had the Graph/Graph Hours in stationsettings.html set to 48 (rather than default 24) because I like to get a better feel of recent short term trends than 24hrs.

So I changed from 48 to 24 (hrs) and found the Graphs/Charts changed immediately to display 24 (hrs) Note: without shutting down CMX and restarting. Also note Cumulus.ini reflected same change.

So then checked r24hour (uploaded to server in CUtags.php) and it now displayed ( 0.6 mm) which is correct.

So in order to prove that possibly (Graph/Hours) was the cause, I changed Graph/Graph Hours in stationsettings.html back to 48 and confirmed change in Cumulus.ini though Graph/Charts did not change to 48hr period this time, so Ctrl C shut down CMX and restarted (no change. Cumulus.ini still showing GraphHour=48 (but Graph/Charts only displaying 24hrs), rebooted CMX again no change, tried increasing to 72hrs (no change) then back to 48 and again still no Graph/Chart change but Cumulus.ini changed correctly each time.
Note: r24hour still showing correct (0.6) each time.

Now I am unsure why I can't change it back and also cannot confirm/validate that this was the cause of incorrect r24hour reported. or a pure coincidence in timing, but I'm sure there is a definite link !

In case you need it, system info;
1: Running latest C1 and CMX versions;
2: Operating System (Laptop running Win7).
3: Using VVP with C1 and MX
4: Station Type VPro2(+)
5: DavisFirmwareVersion 3.15
6: MX ( Cumulus.ini ) [station] No entry for ( UseDavisLoop2=0 ) as felt not relevant by your specs because firmware = 3.15 (even though using VVP)..
7: MX ( Cumulus.ini ) set to "VP2ConnectionType=2

I have left Graph/Graph Hours set to 24 for now in order to determine if it continues to report r24hour correctly. Though really would like to display 48hrs in Graph/charts.
Can't help but be curious to know what the underlying code is that determines r24hour ?
If you need additional files please advise.

Kind regards;

Forgot to mention above, on each occasion cleared Browser cache to eliminate that as a factor too.

regards,

I have a data structure which holds all of the data points for the last 24 hours. I use this for the graphs, and also for the rain in the last 24 hours - by subtracting the last figure from the first. When I changed the code to allow more than 24 hours for the graphs, I didn't take account of this for the rain calculation - I'll fix it in the next build.