Cumulus Support

Hello again, I've been reading the FAQ but can't find this answer: namely my console is showing a max gust of 42mph but Cumulus is saying 36mph. I know the wind recordings depend on configuration, but I thought the max speed might be the same regardless of what options are ticked. Thanks again.

The maximum for what period? Where exactly are you looking in Cumulus? When were the the two figures recorded? Do you run Cumulus 24/7, or do you turn it off for periods, e.g. overnight?

Hello Steve, I'm referring to the All time max; so on Cumulus I click 'view' on the display (36.5mph 19.14 on 19 March) and on the console I touch 'Wind' 5 times (42.5mph 15.43 on March 18).

Yes, I turn Cumulus off overnight (and also on occasions during the day). Looking at the data logs for March 18 I see the gusts either side of the 'record' gust were 12mph and 19mph so it's maybe a station misread...

Cumulus won't necessarily pick up the all-time max gust if it occurred while it wasn't running, because it doesn't necessarily appear in the console logger data, the entries in which are snapshots of the current values at the time of the entry.

If Cumulus was actually running at 15:43 on March 18th, then it should have picked up that figure, assuming that it did actually receive the figure from the console, and you haven't got a 'spike removal' setting which would have caused it to be ignored (as it does appear to be a large sudden increase).

If you zip up the diags folder and attach it, I can take a look.

Well I've attached the diags folder Steve, I actually recall that Cumulus wasn't running that March 18 afternoon and there were some heavy squalls at the time, so the console figure might be correct.

I was under the impression that Cumulus retrieved the data from the station logger whenever it restarted, and that any max values would be recovered in this way but I may need to look over that re configuration etc. And I can confirm that I don't have any spike removal settings at this point.

Mortello1 wrote:Well I've attached the diags folder Steve, I actually recall that Cumulus wasn't running that March 18 afternoon and there were some heavy squalls at the time, so the console figure might be correct.

Yes, you closed Cumulus at 14:54, and started it again at 22:45, so the data for the period in between came from the logger when it started up again.

I was under the impression that Cumulus retrieved the data from the station logger whenever it restarted, and that any max values would be recovered in this way

Yes, it does download the data from the station logger when it starts up. And the data from the station logger didn't contain that high value:

timestamp = 18/03/2014 15:38:20
Data: 0A 2A D6 00 43 75 00 36 27 29 58 00 08 7C 00 00
Gust = 8.8 m/s = 19.7 mph


timestamp = 18/03/2014 15:48:20
Data: 0A 2A D5 00 43 75 00 37 27 1B 41 00 08 7C 00 00
Gust = 6.5 m/s = 14.5 mph

As I said, the figures in the logger data are snapshots of the values at that moment in time. If the peak gust happens not to be in there - and the longer the logger interval, the more likely that is to happen - then Cumulus won't see the figure.

Using the station's logger data is a compromise; it only contains a sample of the available data points.

Apologies Steve, I now see that the time on the console said 15.23 and not 15.43 I think I'm just trying to clarify how an entry can be stored on the console but not appear in the logs, even if it is an error. Sorry for the trouble.

The exact time doesn't make any difference, the same explanation applies.

timestamp = 18/03/2014 15:18:20
Data: 0A 2C DA 00 49 6C 00 35 27 0A 1F 00 0E 7C 00 00
Gust = 3.1 m/s = 6.9 mph

timestamp = 18/03/2014 15:28:20
Data: 0A 2B D8 00 46 70 00 36 27 0E 36 00 0A 7C 00 00
Gust = 5.4 m/s = 12.1 mph

Mortello1 wrote:Apologies Steve, I now see that the time on the console said 15.23 and not 15.43 I think I'm just trying to clarify how an entry can be stored on the console but not appear in the logs, even if it is an error. Sorry for the trouble.

Basically the station shows you data every 48 seconds and transmits it to Cumulus if it's connected in the same period. However, it also records it every 10 minutes to the station's internal memory. The station's internal memory only stores the data it was showing at the minute it was recorded. For instance, if when downloading the data to Cumulus you had two entries, one at 15:30 and other at 15:40, if the maximum gust had been at 15:34, it wouldn't appear on the Cumulus records because it would only have downloaded those two entries (given the station stores data in the internal memory every 10 minutes. That memory is that one that is downloaded to Cumulus when you start it after being off overnight).
If you want all the data to show, you could reduce the logger interval do one minute, however, with the compromise that it would only hold two days of data.

Can I just quote the relevant part of the faq:

"Fine Offset sensors measure wind over 48-second intervals, divided into 24 2-second intervals. At the end of each 48-second period, they transmit the highest of those 24 values as 'Gust', and the average of those 24 values as 'Wind'. The console keeps writing the newly received Gust and wind values to the same area of console memory every 48 seconds until the station logging interval time is reached, then it moves on to the next memory location, leaving the last values it wrote in the previous console memory location.

If Cumulus is not running all the time, on restart it works through the console memory locations and extracts the Fine Offset Wind and Gust in each memory location, i.e. at the station logging interval."

So if I'm reading this correctly, by only extracting the data from the logging interval Cumulus misses out on the remaining areas of the console memory. This seems like a drawback but I suppose any record data will still be available through the console.

Yes, that is how it works because it is what the console saves. You can reduce the logging interval to a minimum of one minute, using an utility also provided in the Downloads page of this site (SetLogger), with the compromise of having memory for only two days but with the advantage of not losing anything.

Mortello1 wrote:So if I'm reading this correctly, by only extracting the data from the logging interval Cumulus misses out on the remaining areas of the console memory.

Which "remaining areas" do you mean?

Spider-Vice wrote:Yes, that is how it works because it is what the console saves.

Yes, but the console did save the max gust which Cumulus did not retrieve when it restarted. The timing of the gust was given as 15.43 even though the station timestamps are 15.38, 15.48 etc. It's that disparity which confuses me.

steve wrote: Which "remaining areas" do you mean?

It's a reference to the faq above, which mentions "the same area of console memory every 48 seconds" and then talks about
"each memory location, i.e. at the station logging interval." Sorry for dragging this out but I'm basically trying to clarify why the 10-minute logging interval on the console is not mirrored exactly by the data Cumulus extracts when it restarts.

The memory used for the 48-second logging interval is the same memory used for the 10-minute logging interval. It's all the same data. That's what the FAQ is trying to explain, and clearly not succeeding.

The console writes a new set of data (consisting of 16 bytes) to the logger memory every 48 seconds. It overwrites that same data every 48 seconds. When Cumulus is running, it reads that data before the console overwrites it. No data is therefore missed.

After a period equal to the station logging interval (10 minutes in your case), the console increments its current memory pointer by 16 bytes and starts writing new data every 48 seconds to the new location. The last set of data that it wrote the previous location - and only that data - is thus preserved for posterity. All the other individual values it had been writing for the last 10 minutes have gone for ever, as they have been overwritten.

Your max gust figure at 15:23 was in one of the sets of data which had been overwritten. If Cumulus had been running at the time, it would have read that data before it got overwritten. But Cumulus was not running at that time, so the only data available to it when it started up was the data left behind after the console moved on to the next location. This happened at about 15:28, so that was the data available to Cumulus, and that was the data that it read.

I'm basically trying to clarify why the 10-minute logging interval on the console is not mirrored exactly by the data Cumulus extracts when it restarts.

Sorry, I don't understand that. The data read by Cumulus when it started up was at 10-minute intervals, you can see that in the examples I quoted:

timestamp = 18/03/2014 15:18:20
Data: 0A 2C DA 00 49 6C 00 35 27 0A 1F 00 0E 7C 00 00
Gust = 3.1 m/s = 6.9 mph

timestamp = 18/03/2014 15:28:20
Data: 0A 2B D8 00 46 70 00 36 27 0E 36 00 0A 7C 00 00
Gust = 5.4 m/s = 12.1 mph

timestamp = 18/03/2014 15:38:20
Data: 0A 2A D6 00 43 75 00 36 27 29 58 00 08 7C 00 00
Gust = 8.8 m/s = 19.7 mph

timestamp = 18/03/2014 15:48:20
Data: 0A 2A D5 00 43 75 00 37 27 1B 41 00 08 7C 00 00
Gust = 6.5 m/s = 14.5 mph

That hex 0A in the first byte of the entry is the logging interval, i.e. 10 minutes. Cumulus calculates the timestamps from those values.