Cumulus Support

Hi Steve,

Why is there a difference between my Davis Vantage Vue Console and Cumulus in terms of high wind gust?

For example:

DVVC: 40 mph
CU: 38 mph

They both read the data at 2.5 seconds, so why is there a difference?

William

Cumulus will inevitably miss a few readings because it has to request other data from the console, and the DLL will only let it do one thing at a time. If you're unlucky, the high gust speed will be in one of those missed readings. What firmware version does your console have?

steve wrote:Cumulus will inevitably miss a few readings because it has to request other data from the console, and the DLL will only let it do one thing at a time. If you're unlucky, the high gust speed will be in one of those missed readings. What firmware version does your console have?

That's annoying because it was quite windy today. How do I find my firmware version?

William Grimsley wrote:That's annoying because it was quite windy today.

Sorry, it's the way Cumulus works. Other software is available.

How do I find my firmware version?

Cumulus writes it to the diags file, if the console supplies it. Some firmware versions seem not to. Other than that, your console will be able to display it.

steve wrote:

William Grimsley wrote:That's annoying because it was quite windy today.

Sorry, it's the way Cumulus works. Other software is available.

How do I find my firmware version?

Cumulus writes it to the diags file, if the console supplies it. Some firmware versions seem not to. Other than that, your console will be able to display it.

Ok, thanks.

Here we go:

30/04/2013 21:04:24.292 : Firmware Version: 2.14

William Grimsley wrote:30/04/2013 21:04:24.292 : Firmware Version: 2.14

I'm not aware of any issues with that firmware, so Cumulus will be reading the data as fast as the DLL allows it.

steve wrote:

William Grimsley wrote:30/04/2013 21:04:24.292 : Firmware Version: 2.14

I'm not aware of any issues with that firmware, so Cumulus will be reading the data as fast as the DLL allows it.

Ok, is that the latest firmware?

William Grimsley wrote:Ok, is that the latest firmware?

The latest is 3.00. It appears that there may be issues with that version.

steve wrote:

William Grimsley wrote:Ok, is that the latest firmware?

The latest is 3.00. It appears that there may be issues with that version.

Ok, how do I update to that? Via http://www.davis.net?

EDIT: Just read the readme.txt file and found that adding: 'DavisCalcAltPress=0' to the Cumulus.ini file stops Cumulus from missing data.

William Grimsley wrote:

steve wrote:

William Grimsley wrote:Ok, is that the latest firmware?

The latest is 3.00. It appears that there may be issues with that version.

Ok, how do I update to that? Via http://www.davis.net?

EDIT: Just read the readme.txt file and found that adding: 'DavisCalcAltPress=0' to the Cumulus.ini file stops Cumulus from missing data.

Hi Steve,

It looks like adding 'DavisCalcAltPress=0' didn't do the trick and instead still missed the high gust today of 22 mph. Not impressed. I added 'DavisCalcAltPress=0' to my Cumulus.ini file under the [Station] section of the Cumulus.ini file but that didn't work, even though the readme.txt file DOES say this:

This can take several seconds, so it means
that a 'normal' data reading may be missed, so you might miss
a high gust value, for example.

So, it must be something else that made CU miss the gust...

William

William Grimsley wrote:So, it must be something else that made CU miss the gust...

Well deduced.

Not impressed

I know the feeling.

Try using Weatherlink, that was written specifically for Davis stations, and doesn't cater for 0900-0900 meteorological days, so can use the highs and lows from the console, unlike Cumulus. Or do a bit of work yourself and work out when and why the value was apparently missed. At least give me something to go on instead of just complaining. Are you missing an unexpectedly high number of readings, for example?

The Hi's and Lo's screen on the console should show the time and date of the max gust (going back a day or two if necessary), without necessarily needing to run Weatherlink. That should make it clear which day a particular Hi gust fell into. (Remember the console is midnight to midnight and only stores one max gust per day, so it won' tell you much about the gust pattern - you'll need WL to get a more detailed view.)

I've done some timing on my VP2 (fw 3.12), looking at the time it takes for the DLL to read the data over a minute. The time between each 'Load current data' and 'Data loaded' is all time spent in the DLL fetching the data, typically around 3.2 seconds. The time between 'Data loaded' and 'Load current data' is the time taken for Cumulus to process it, normally less than 0.001 seconds.

There are 19 reads over a period of a just under a minute; you would expect the ISS to send 23 packets in the same period. So, as I said, it's inevitable that there will be missed packets, and this is almost entirely due to the time taken for the DLL to get the data from the console. I have to say that I'm surprised that as many packets as that are lost, it's about 17%. I've done a similar exercise some time ago and didn't get such poor results; I suspect this is down to the firmware changes.

0341.518 : 11:08:58.163 Load current data
0344.747 : 11:09:01.382 Data loaded
0344.747 : 11:09:01.383 Load current data
0347.961 : 11:09:04.601 Data loaded
0347.961 : 11:09:04.601 Load current data
0351.190 : 11:09:07.820 Data loaded
0351.190 : 11:09:07.820 Load current data
0354.404 : 11:09:11.040 Data loaded
0354.404 : 11:09:11.040 Load current data
0357.617 : 11:09:14.259 Data loaded
0357.617 : 11:09:14.259 Load current data
0360.846 : 11:09:17.477 Data loaded
0360.846 : 11:09:17.477 Load current data
0364.060 : 11:09:20.697 Data loaded
0364.060 : 11:09:20.697 Load current data
0367.414 : 11:09:24.052 Data loaded
0367.414 : 11:09:24.052 Load current data
0368.397 : 11:09:25.033 Data loaded
0368.397 : 11:09:25.033 Load current data
0371.610 : 11:09:28.254 Data loaded
0371.610 : 11:09:28.254 Load current data
0374.840 : 11:09:31.473 Data loaded
0374.840 : 11:09:31.473 Load current data
0378.053 : 11:09:34.692 Data loaded
0378.053 : 11:09:34.692 Load current data
0381.267 : 11:09:37.912 Data loaded
0381.267 : 11:09:37.912 Load current data
0384.496 : 11:09:41.130 Data loaded
0384.496 : 11:09:41.130 Load current data
0387.710 : 11:09:44.348 Data loaded
0387.710 : 11:09:44.348 Load current data
0390.923 : 11:09:47.567 Data loaded
0390.923 : 11:09:47.567 Load current data
0394.153 : 11:09:50.786 Data loaded
0394.153 : 11:09:50.786 Load current data
0397.366 : 11:09:54.005 Data loaded
0397.366 : 11:09:54.005 Load current data
0400.580 : 11:09:57.225 Data loaded
0400.580 : 11:09:57.225 Load current data

Steve, a question I haven't asked as I'm sure it is been covered the past somewhere before I got a Davis... the Davis protocol has the ability (the "LPS" command) to send LOOP and/or LOOP2 packets every 2.5 seconds rather than requesting them individually - or send the LOOP packet every 2(?) seconds. Is this feature not available via the DLL? It would seem that requesting a minutes worth at once would reduce the number of missed readings?

The recently introduced LPS command is just the equivalent of the LOOP command, but you can request the newer LOOP2 packets on more recent firmware with it, as well as the older LOOP packet. The LOOP command itself also has a parameter to request the number of packets, and the station sends that number of packets without requiring further intervention. But the DLL always sends a "LOOP 1" command, and there's a delay before the data comes back. Of course, as we've seen, this is firmware-dependent, and some of them return the data straight away, so there has to be a delay in Cumulus. Similarly with VVP. Presumably fewer packets are missed in those cases.

I've added code to read today's high wind figure from the console once a minute and use that if it's higher than the figure that Cumulus got from the LOOP data. But only for midnight start of day users, of course; hopefully those users, including "not impressed", will find it an improvement.