WMR-200 Data skipping

Hello

I recently purchased a WMR-200 and I’m using it with the free version of WD 10.37i.

After a steep learning curve with the program everything seems to work pretty fine now. However I do get a lot of skipped entries when I download the data from the logger to WD.
I have attached my wmr200history.txt file and would appreciate if somebody could have a look at it and tell me if this is normal behaviour.

Thanks!
Andy


wmr200history.txt (94.3 KB)

Andy - welcome to the forum! :wink: Windy the developer and engineer, is working on this issue I believe right now. Tghere muszt be a thread on it somewhere already - have a search “wmr200” and “history” or “graphs”

TokKiwi

Hi TokKiwi

Thanks for the welcome! I did search for related WMR200 problems before I posted and if I understood correctly most of the problems are related to the live data feed to WD.
Maybe Windy can confirm if that’s the case and if my problem with the history extraction is different?

Thanks
Andy

Lots of skipped data in the ‘missed data’ phase is certainly something I have come to expect. :frowning:

And due to an oddity in the sequencing of skipped records in the log, pretty much every day I find myself doing a ‘sort’ of the log files to get the graphs looking sensible (i.e. outside of WD I re-sort the file, then use Action => Convert… to correct the graphs).

David

I just checked the log file and noticed that my log entries are messed up as well (see attached). They are not in chronological order on many occasions. There seems to be a major problem with this and I hope Windy can come up with a fix for it…

Regards
Andy


42008lg.txt (18.6 KB)

The problem I see and which I’ve described more than once is that if there is a skip, the extra records designed to fill the gap:
a) Start with a record for a time that already has a record (the time just before the gap).
b) Are inserted into the log after the record that marks the end of the gap.

Hence the need to sort.
Initially I used to edit out the duplicate records too, but it was just too much to do!

Hopefully it will be addressed next time Brian is working on that area of code.

In my opinion not there are record skipped in the internal memory of the wmr200:

It would of course be very nice if the data were read successfully in the first place - then we also wouldn’t have the sequencing problems caused by the positioning of the records that are provided by WD to fill the gaps. :frowning:

Maybe there is some technical problem with the skipped records in the data that follows the time stamps you have found, e.g. a checksum? Positioning of the fill-in records however ought to be a straight forward issue to solve.

make sure you are using the most recent version of WD
as the skipped data should only occur occasionaly (in testing here)
yes I do need to spend more time on trying to eliminate the problem even more (I think it occurs when the data is split over more than 1 line of code)
how much data are you downloading ?
i

Typically around 9-10 hours. Very rarely more than 14 hours. The logging interval is 1 minute.

Skipped data sections are far from rare here. I’d not be surprised if there were a dozen such sections in a 10hr catch up.

I was actualy asking barpm that question…

that is relatively rare, 12 out of 720

I have the 10.37i build 11, because the problem of the chronological order of the log files in the next versions

how much data are you downloading ?

usually around 18-20 hours

Sorry. I misunderstood.

If it was only that sort of volume I’d not be bothered (honest!) - but ‘sections’ of skipped records are usually multiple minutes long, and it seems I was being conservative in my estimate.

I’ve just done a check on two actual wmr200history.txt files picked at random, in one there were 134 skipped minutes in 660 minutes of missed data, in the other there were 277 skipped minutes in 630 minutes of missed data - that’s pretty substantial amounts of data.

You are welcome to copies of the wmr200history.txt files if you wish.

Brian, please have a look at my history file in the first post. You can see that the skipped records are far from rare.
I would hope that if I am forced to use your software - because it’s the only one supported - it would work in a reliable manner.

There are other annoying issues with this software, e.g. it has rained yesterday, it shows it the total amount, but not in the “Yesterdays rainfall” section. In addition I have a very strange reading of 37.9 mm/hour for the last six hours which is bogus as it hasn’t rained at all during that period.

Andy, I assume you are shutting the PC down every day or evening, have you considered leaving it running to stabilise your data, or do your problems still exist while connected and running?

Often, with a new style of weather station model, the problem is with parsing the data stream accurately and assuming the communications protocols are clearly defined. It may be the comms and parsing that is the problem and not the WD display issues.

Is your weather pc a stand alone PC or are you using it for other purposes while receiving data?

Have you run Task Manager to see what cpu load you are incurring? What pc details (performance specs etc) can you share?

TokKiwi

Running till a little after midnight (at least 00:15), ought to sort out the ‘Yesterday’ figures because (as I understand it) those are only calculated if the ‘change of day’ happens while WD is running with ‘live’ data.

To achieve that with my WMR200, I use some free software (http://tinnes.org.uk/) to automatically shutdown the WD computer at 00:25.

Maybe my usage of the software differs from most people here. I don’t have a pc connected all the time. I connect my laptop from time to time to extract the history data and then disconnect it from the WMR200. I then publish the files locally to analyze them.

I might consider setting up a pc to do real-time logging, but for the moment I would be happy to have accurate history data.

CPU load is ok, no anomalies there.

Thanks for the replies btw!
Cheers
Andy

the daily reset should occur as long as the change in day/time was found in the extracted history data
(you would then see that daily reset being logged under view, program event log)
I will try and find some time to do more work on the wmr200 history soon

I would hope that if I am forced to use your software - because it's the only one supported - it would work in a reliable manner.

I think that is a pretty rude comment
VWS now supports the WMR200, so I would try that

There are other annoying issues with this software, e.g. it has rained yesterday, it shows it the total amount, but not in the "Yesterdays rainfall" section. In addition I have a very strange reading of 37.9 mm/hour for the last six hours which is bogus as it hasn't rained at all during that period.

if WD was running all the time then you would not have these problems
but the daily reset should still be occuring from the history data, I will do some testing of that

Maybe it was rude, sorry for that. I’m just a bit frustrated that such a renowned company like OS does not provide an open interface for the software.
I certainly appreciate the efforts you put into WD, but I was not prepared for spending so much time in getting familiar with the software. While WD is very powerful for the advanced user, it can be overwhelming for someone new to the topic, especially since it is not the most user friendly piece of software (again, this is my opinion and I hope it is taken as such). By not being user friendly I mean that WD looks “overloaded” with options and the menu structure does not always seem logical.
Maybe it might be worth thinking about the ability to toggle between a “beginner” mode with only the necessary basic options and an “advanced” mode for expert users.
Again, much of my criticism about the lack of reliability with data extraction goes to OS as I think they should have made sure it’s working before going to market with this bundle.

Cheers
Andy

the latest update has the skipped data problem much improved (nearly perfect now)