Rain Rate Query

[Bugs]

Not sure how the rain rates are determined?

Recent rain event the display showed the rain rate at 48.4mm/hr at 01:50 yet the record all time high rain rate shows 122.2mm/hr at 2:00 (same day)?

The software or PC was not running at the time but started at about 7am (same day). The following day "Yesterday" rolled over the 48.4mm/hr figure, All time rain rate remained 122.2mm/hr.

WeatherLink recorded 122.2mm/hr at 2am (archive interval weatherlink 1 hour), Cumulus archive 10 minutes

Also the "All time high hourly rain" is 0.4mm back on Oct 9 but WeatherLink recorded 6.6mm for the hour ending 2am

Version is 1.8.1

[steve]

For the VP/VP2, The rain rate figure is read from the station, both during startup, from the logger, and during normal running.

I can't explain how Cumulus could have got the all-time high correct but somehow got a 48.4 for the daily high. Each time it reads a new value from the station it first checks if it is a new all time high, and then immediately afterwards checks if it is new daily high. I'll have a look at the code to see how it might be going wrong. It would be more understandable if it hadn't logged the 48.4!

The hourly rain is calculated by Cumulus by comparing the current total with the total an hour ago (by recording the current value once a minute). During startup there is no 'once a minute' for this to happen, so I think the hourly calculation doesn't work during startup. I'll have to think about how to do the hourly rain calculation during startup.

[steve]

Recent rain event the display showed the rain rate at 48.4mm/hr at 01:50 yet the record all time high rain rate shows 122.2mm/hr at 2:00 (same day)?

The software or PC was not running at the time but started at about 7am (same day). The following day "Yesterday" rolled over the 48.4mm/hr figure, All time rain rate remained 122.2mm/hr.

Thinking about this, there's something odd about those times. If your logger records only once an hour, I wonder where the two figures 10 minutes apart came from?

Could you have a look in dayfile.txt and see what your high rain rates and times were for the last few days?

[steve]

Thinking some more (it's amazing what comes to you when you're trying to fix the central heating) I'm wondering if your hourly logging interval is causing me a few problems. I can't say for certain that it is, and in theory it should be OK, I just didn't envisage anyone using such a long interval. Anyone using 2 hours is definitely going to have problems, I'm afraid

Could you send me the files in your Diags folder, please?

[dane]

I can confirm that the display field: Rainfall Last Hour is incorrect when restarting Cumulus.
I shut down Cumulus at 23:00 last night and just restarted it.
At start-up Cumulus says: Rainfall Last Hour: 3.9 mm - and I have had no rain since 3 in the morning.
I shut down Cumulus immediately (before it added a new record to the oct08.txt file) and restarted.
2nd time round Cumulus said: Rainfall Last Hour: 0.0 mm
Can supply screenshot and Diags files if required.
Cumulus 1.8.2, build 453.
Ib

[steve]

The difference between what you're seeing and what a VP/VP2 user sees is that for the VP/VP2, I deliberately set the 'last hour' rainfall to zero just before switching to normal running but I forgot to do that for Fine Offset users.

With the way I currently work out the last hour's rainfall, it's just too hard to work out what it was in the hour before startup, and continue to keep it correct for the next hour (i.e. until I have an hour's worth of real, current data). The same applies to the temperature and pressure trend values, which use the same mechanism.

I will have to think of a different mechanism for working out the three values. Currently I have an array of 60 values, and each minute I write the current value into the location indexed by the current minute past the hour, having first compared it with the value that was there already. Hopefully you can see why this doesn't work when reading the history data. Anyone with any bright ideas, don't be shy!

[dane]

Steve, thanks for the explanation.
No: no bright ideas
I am happy with my circumvention: just shut down immediately and restart. That fixes it.

[Bugs]

First one correction. The logger is configured for an interval of 10 minutes using WeatherLink 5.8.0 to a standard serial logger in VP2 Wirless Envoy. There just happens to be 2 wireless VP2 Envoys (ISS direct) and one console through a repeater. One Envoy is logging 1 hour, the other which is connected to Cumulus is logging 10 minutes and the console is 1 hour.

There would appear 2 separate issues here 1) Rain rate 2) All time record high hourly rain

1) Main points

- Cumulus and WeatherLink both configured for 10 minute archive interval
- Cumulus display showing a high rain rate of 48.4mm/hr at 01:50
- dayfile.txt for 17/10 also 48.4mm/hr at 01:50
- Cumulus "Record" All time high rain rate is 122.2mm/hr at 02:00
- WeatherLink high rain rate is 48.4mm for the 10 minute period ending 01:50 Oct 17
- WeatherLink high rain rate is 122.2mm for the period ending 02:00 Oct 17

- PC (Cumulus) was shutdown during the night and restarted at approx 7am (Oct 17)
- Rain occured around 2am Oct 17.
- High rain rate for Oct 16 matches WeatherLink at 27.6mm/hr at 22:00

NOW what the Oct08log.txt shows is there are NO records for the period 01:40 to 01:10 inclusive ???? Which probably might explain the difference with the high rain rate BUT these "missing" records do exist in the WeatherLink download?? So this does not appear a reception problem as the WeatherLink records appear complete for all three loggers and especially the one connected to Cumulus and configured for an archive of 10 minutes for both Cumulus and WeatherLink.

If of any use can provide the WeatherLink and Cumulus data files and a print of the display from Oct 17 (late in the day)

[Bugs]

Splitting the issues

2) All Time Records window, All time high hourly rain

- displays 0.4mm/hr on Oct 9
- Between 02:00 and 02:10 (Oct 17) 5.4mm was recorded
- There have been many other periods since Oct 9 of more than 0.4mm during an one hour

[steve]

First of all, the easy bit: the hourly rain is a known deficiency, see above. I have come up with a revised scheme for working it out (and the temp and pressure trends) which should fix this, probably in 1.8.3 rather than this release as it's quite a big change.

The other problem I will have to investigate, but sadly it's now Monday so real work intervenes.

Edit: But please send me the contents of the Diags folder, as requested above. Hopefully the relevant files haven't been overwritten yet. And yes, the Oct08log.txt file will be useful, please. What you say about the missing entries is interesting and is clearly the cause of the missed high value. For some reason Cumulus didn't get those entries from the logger.

[steve]

First of all, the easy bit: the hourly rain is a known deficiency, see above. I have come up with a revised scheme for working it out (and the temp and pressure trends) which should fix this, probably in 1.8.3 rather than this release as it's quite a big change.

Sadly I've realised that my new scheme isn't going to work due to the way I store the rainfall in the log files. I think this is all going to have to wait until the change to The Great Database Storage System.