Major rainfall accumulation error in both Meteobridge Pro and Nano SD - WHEN USING TIMEZONE WITH PARTIAL HOUR GMT OFFSET

[Eucalyptus]

Hi,

Please refer to my earlier post with the same main title earlier this month.

It was suggested that I try a database rebuild to fix this issue.

Unfortunately, the problem still exits. The rebuild had the same effect as a reboot.....a temporary fix to the data issue.

I have since loaded the latest firmware (Meteobridge 5.4 (Jun 28 2022, build 14756), FW 1.4)

Both databases in both devices have been rebuilt again also after the upgrade. Sadly, still no solution.

Large errors in accumulated rainfall are still observed after less than 24 hours!

Once again a reboot appears to restore the data, temporarily.

I would suggest that anyone who has either a Nano SD or a Meteobridge Pro working with a Davis Vantage Pro 2 system check whether their systems actually work correctly! Observe whether the rainfall totals on the Weather 34 screen match those on the Davis Console.

Another easy way to check in the case of either device is to look at max/min data rainfall. Note the values for month, year and total, then perform a reboot and check whether the totals are still the same.

I hope that maybe I am doing something incorrectly with my settings.....surely these products must work correctly?

Any help with this major problem will be greatly appreciated.

Thank you.

Emma

[admin]

When given remote login I might be able to find out what goes wrong with your setup. you have PM.

[Mattk]

I run multiple NanoSD's in Envoys and Vue consoles as well as MBPro's in parallel with Envoys+WLIP over many years and have not struck the issue you describe. For specific reasons I run a variety of FW versions from build 14754 (Jun 11, 2022) back to build 11957 (Feb 12, 2018) and several versions in between.

Do not run Weather 34 so unable to comment on that comparison but I have direct and independent data comparisons with the MBPro's (all direct RF) typically uploading to WoW as well as 24 hour 1 minute data archives and the Envoy+WLIP's uploading to WeatherLink.com and subsequently WeatherLinkPC.

Rainfall totals match exactly between the Envoy and MBPro. There is a bit of an unknown when a NanoSD uploads to WeatherLink.com in regard missing records (applies to all sensor data) due to API workings but downloading a NanoSD directly to WeatherLinkPC works fine.

One question would be, do you have rain0total included in the Station>Settings>Calibration?

[Eucalyptus]

Thank you Mattk for your response.

No data from either the Meteobridge Pro nor the Nano SD is uploaded to any cloud networks in my systems and both devices process and store all data locally.

So maybe your conclusions based upon your scenarios are not a fair comparison in this case, as your instantaneous rain data maybe is uploaded as it falls?

The errors in the stored data that are observable in both the Meteobridge Pro and the Nano SD are generated entirely within each device.

As I highlighted in my post, all settings are default settings as set by the manufacturer, except for settings required to interface with local networks and the Vantage Pro 2 via radio link. There is a calibration offset for the wind direction sensor in the Meteobridge Pro.

Finally, yes, I have the rain parameter rain0total available as a drop down option in Station>Settings>Calibration.

No special functions, script processes or data translations are used in either device.

[Mattk]

No what I am comparing is independent data between the MBPro/NanoSD as stored on the device, Davis uploads from Envoy/WLIP connecting to the same ISS and third party uploads. Rain data that is uploaded, stored locally, downloaded directly or from multiple third party sites agree.

.... Finally, yes, I have the rain parameter rain0total available as a drop down option in Station>Settings>Calibration....

If the rain0total is available as a drop down option then that tends to indicate it has not been added to the sensor calibrations (as such)? Once a sensor is added to the calibration suite then it won't still appear in the drop down option. Not sure of your wording with this one?

[Eucalyptus]

Thank you again for your feedback Mattk.

The data error problem is real and readily observable in the manner I have described in my original post.

What are you suggesting should be done/or I haven't done with the drop down box?

I'm not sure why you think this may cause the accumulating error issue.

Why does the accumulated rainfall data accumulate ever increasing errors after a period of time?

Why does the accumulated rainfall data total for month/year/total change after a reboot or database rebuild after a period of time as previously described?

Thank you.

[Mattk]

Well I can't explain what you are seeing because it's an issue I have not observed.

What I am doing is asking queries re what I have previously observed with rainfall totals following the upgrade from the original interface to the 2-tab version. The question still remains in regard the rain0toal calibration inclusion, I'm really not sure from your comments if you have it in the active calibration or not.

Could you post a screen shot of both the Station>Setting pages?

[Eucalyptus]

The only calibration offset in the Meteobridge Pro is for wind direction as per my last post.

No other calibrations are used. That is why rain0total is still observable in the drop down list.

Are you suggesting rain0total needs a calibration factor?

Not sure how any calibration factor on any variable would cause data to change after a reboot or rebuild.

I'm sorry you cannot duplicate the problems I have illustrated with my both detailed descriptions and screen shots of the errors.

Not quite sure where to go from here, other than to wait for the manufacturer to provide a solution.

Thank you for your suggestions so far.

[admin]

I don't question your observations, but I currently lack any idea why it is fine after a reboot (at least I understood you that way) and then starts getting bad again.

I had a situation once, where a client did receive rain data from a far away ISS here and there and that messed up rain data as the ISS is reporting bucket counter values that of course differ a lot between stations. But having this will just lead to spiking your data not to have it fine after reboot and then again wrong after some time.

You seem to be very focussed on that the Meteobridge is generally simply doing wrong on rain totals, but as Mattk also explained to you it is not with the thousand's other users.

As already offered, I can inspect via remote login and testing things on your Meteobridge might give me a clue. Link explains how to setup remote login very easily: viewtopic.php?t=9885

[admin]

Another question... in your screen dumps in the initial post you compared data from the dashboard and the database edit tab. Did you experience inconsistencies in the "hourly data edit" tab (which would then also reflect in exports and charts) or just in the dashboard?

[Eucalyptus]

Hi admin,

The data is definitely not being corrupted by reception of other weather stations. The locations of both systems are remote from any interference and well separated from each other and as you convinced yourself in your recent post the data error would not be as I am observing.

As I described in my original post I have shown the data error becomes more obvious as either time passes or more rain falls.

Please refer to my original post for the explanation again as follows;

These screenshots are all from a Meteobridge Pro.

Please note the first screen shot which shows accumulated rainfall for the 4th June at 11.54pm as about 7mm.

However, please look at the next screenshot, which shows the total for 4th June at 11.54pm as 10mm under Max/Min data for the day Saturday 4th June.

Next Image 3 which from the time stamp shows the Max/Min data on 5th June at 12.15am. Please note the rainfall totals for the "month", "year" and "total" columns.
The system has rolled over to a new day (5th June) and no rain has fallen as yet.

As I described the "all in one chart" in the graph generator section, shows the rainfall just before midnight on 4th June as 10mm.

Here is a screenshot of the graph for 4th June, just before midnight.

D2225D4E-9D10-48C1-B115-C9AB1725720B.jpeg (129.81 KiB) Viewed 1787 times

So at this stage rainfall total data for the 4th June just before midnight varies between 7mm and 10mm.

Next a reboot is performed.

Please look at screenshot 4 (12.58am on 5th June) which shows the hour data column now at 10mm for 4th June. The data now agrees with the graph total for the day of 4th June, 10mm.

Please look at the next screenshot, which shows the max/min data at 12.59am on 5th june. This is after the reboot as indicated above. Note the rainfall totals for "month", "year" and "total" have changed several mm although only 0.4mm has been recorded as a daily total so far for 5th June.

As you can hopefully observe the error is real.

I can't make any comment as to the assertion that thousands of other users don't have a problem. This information is not available to me.

However, I do note that most users appear to use a cloud service. If the data sent to the cloud service is in error, maybe its quite possible that locally recorded database information is in agreement with that data.

Surely it does not necessarily mean that that data is correct?

This problem was first observed by comparing the accumulated monthly total rainfall on a Davis Console which contained the Nano SD. It was noted that the monthly totals on the console and the Nano SD did not agree with each other.

That is why I am at this point now.

In the absence of any other reference other than the Meteobridge data collection platforms, maybe no one sees the issue?

Maybe this issue does not occur if data is sent to a cloud server?

As I have described, my settings for these devices are essentially the factory default ones. There is no scripting, converting, uploading nor creating of any special databases so duplication of this problem should be very simple with a production model at the point of manufacture. This problem exists in two physical variants of what must be essentially the same product. Both devices have been operating for just over one year.

I look forward to your comments.

Thank you.

[Eucalyptus]

Hi admin,

The following clearly shows the rainfall error on a Nano SD fitted to a Davis Vantage pro 2 Console.

This device operates with factory default settings except for those that are set up for network connection to the local network. No data is sent from this unit to cloud services via the internet. This device and the console has been operating for about one year. The internal SD memory is only 20% used.
The Nano SD is running the latest firmware from your webpage and was rebooted at about 6am on the day of the attached screenshots.

The first screenshot shows 1.8mm rainfall so far for the day. The totals for the month and year are 82.6mm and 214.2mm respectively.

0D965067-4237-42C1-8587-8CA52E7051B1.jpeg (237.53 KiB) Viewed 1758 times

The next screenshot is following a reboot. Please note that no rain has fallen in the few minutes in time between the two images. Note however that the rainfall totals for the month and year are now 83.4mm and 215mm respectively.

F8484EBF-DA9B-43CA-AC5C-473140DFA2CE.jpeg (233.7 KiB) Viewed 1758 times

These screenshots clearly show the rainfall data error. please note that the max/min data for the Nano SD reflected this data at the time of the above screenshots.

The change in data between screenshots following the reboot is small. This is because a previous reboot was performed at about 6am of the same day which corrected the rainfall totals at that time, hence the correction is not as large as the 1.8mm of rainfall for that day.

This rainfall error and the effect of a reboot is observed in exactly the same manner with the Meteobridge Pro.

Please note that the rainfall error continues to grow larger each day as the rain falls, only correcting after a reboot.

I look forward to your feedback.

Thank you.

[Gyvate]

Nobody doubts that there may be an error - but in order to figure out what might have caused it, such descriptions only won't help.
Therefore - why don't you give Boris (@admin, the author of the software) access to your system (as he already offered) so he can inspect your setup and your database to get more insights. Just continuing the discussion at this point won't give any added value.
Why don't you (seem to) accept the offer and let him investigate further ? You won't get a more competent person to do so.
A doctor also won't go for a diagnosis based on a phone call but will want to see the patient in person and perform direct investigation.
Same here with your Meteobridge install(s).
Please consider.

[Mattk]

Yes trying to find issues without information and detail is rather difficult and there does appear a reluctance to provide that detail/screen shots etc for some reason? Ignoring the irrelevance if anything is cloud based (as such) or not any comparisons from HWS locally being a third party application is quite separate to the underlying raw data every Weather Network relies upon. One could assume any reboot issue would be reflected through the entire WeatherNet list?

Concentrating on discussion in regard Meteobridge changing data values with reboots etc I have no idea but I have observed rain issues in various contexts but without providing requested feedback then not a lot can be gained.

As there has conveniently been some rain this month the following raw data cuts are from two(2) Nano's, two(2) MBPro's and an Envoy/WLIP. The first 2 value columns are of primary importance, the third data value can be deemed relevant depending on the systems's uptime, ignore the fourth column. Firmware versions are as noted. All MB and Davis units receiving from the same ISS either directly or via a repeater

NanoSD61 (Apr 5, 2022)


Nano63 (Apr 5, 2022)


MBPro75 (Feb 9,2021)


MBPro78 (Jun 11, 2022)


Davis Envoy/WLIP for comparison




The question the above screen shots highlight is, Rain Error? What rainfall error?

Lets have some feedback, initial screen shots of the following would be a start
Station>Primary tab
Station>Settings tab

[admin]

@Eucalyptus:

Did you experience inconsistencies in the "hourly data edit" tab (which would then also reflect in exports and charts) or just in the dashboard?

Thanks for your explanation, but I don't really got an answer to my question repeated above. This is important as exports and graphs make direct use of permanently stored data while the Dashboard and reports to networks is based on data derived from that. If the stored data is in sync with your console data, this helps me to narrow the field to be inspected.