Solar anomalies in WD

it looks like WD is getting out of sync in reading the WH3081 solar data
(which is a different memory read address than the normal data), and so causing false data

so that explains that
that is not related in any way to the occasional data spike you (and I have not seen other reports of that) get with your Davis station

Thanks, all. I much appreciate your help.

I had one glitch at 00:57 this morning on the WH system, the Davis being OK. Couldn’t have been my fault because I was snoring at that moment #-o. I’m waiting for the Davis system to glitch, but sod’s law applies.

re the WH3081, what I need to do is that if WD knows its night time, then to say, well, that cant be above zero → set to zero

Brian

Thanks, I see where you’re coming from, but the anomalies happen on both the Davis and WH systems at any time of night or day. They may be a single time or they may last for many minutes. As it happens, there have been no anomalies on either system for a couple of days; the timing of them seems totally random. I can’t be categorical, but I’m reasonably sure that if there are glitches on the graphs on one system, they are not at the same times on the other.

I’ve not run the Cumulus software long enough to be sure, but I’ve not seen any glitch on WH (haven’t tried it yet on Davis). Its default interface is not graphical and they have a user-selectable choice of separate graphics for each variable. Not as clever as WD for user interfacing! However, it does tell you everything with enough clicks! I’m not sure yet, but it seems to have a downside that the interval on the data logger may be 10 minutes, pending investigation. My general feeling is that it is a rather clumsy interface but fairly complete if you know the quirks.

One negative glitch this morning on the Davis/WD system. Attachment explains all. No glitch seen on the WH3083/WD. AFAIK, there is no explanation known to me at that time; it was just chugging along, minding its own business!

(negative glitch is down to zero, positive glitch is where it goes to much higher-than-normal levels)


The more I think about this the more I feel that in order to get somewhere we need some way to log the raw data coming in, not just as currently is available where only the latest is captured but we need it running to a file which is appended to, ideally not as currently formatted but as one line per record from the logger and time stamped. That way it would be clear whether or not the raw data was bad.

Stuart

Stuart

I agree with your thoughts, but programming that is something way beyond my capabilities now (gave it up 25 years ago, using a scientific dialect of Forth!).

I’ve just discovered a partial negative glitch on the Davis logs at 11:50 - 11:51, where the value dropped from 719 W/m

Now that I’m looking for anomalies, I’m finding them. This one seems to be less serious in its consequences. The gramophone needle got stuck for about 30 minutes at 14:34 with the same value. Unlike the previous one on Davis, it is on the WH3083 system. No problem on the Davis system during this time. During this same period, another glitch, the UV index has dropped from 6.0 to 0.0 to resume correctly at 15:05 at 5.0 (I don’t have UV on the Davis system). Curioser and curioser!


One thing I have had trouble with is a black bird who liked to land on my solar sensor dropping the sunny value to overcast while he was on it. So glitches like that can happen depending on where your sensor is mounted.

Stuart

Blackbirds are rare here, have seen a male twice this winter. Pigeons are more likely and have had poo on the Davis solar sensor on one occasion with obvious results - actually quite difficult to clean off. However, I can’t see that being the cause of the glitches I’ve catalogued.

Both systems are behaving themselves just now.

Several glitches on the WH3083 system. The attachment shows them clearly. At 19:12 there was a positive to 1662.88 with UV index rising to 1. From 22:34 to 23:03, there was a smaller positive, constant for


as explained in my post above, for the WH3081 station, WD has to read the data from a different memory address on the console, to normal weather data
sometimes that will get out of sequence and so bad data will be returned
(what I need to do is check for the data being bad, e.g check on date or times in the data), so as to ignore that data

so trying to use a WH3081 station with the WD software to show that the davis solar glitches are not real etc is not really going to work
I would instead be using the cumulus software for both

I have though made WD ignore glitches in the data at night for the WH3081 station in the latest update of WD

I’m sorry, Brian, we seem to be at cross-purposes. Please allow me to try and clarify the position.

Several years ago, I added the Davis solar device to my VP2 installation. From the very start, I was getting anomalous glitches, which I blamed on the device as everything else worked with WD. Davis could not understand the anomalies, no matter what we tried. I gave up on solar. Since then, I’ve tried it again on 2 occasions when I changed computers and operating systems, always with the same kind of glitches, but only on solar. I just cussed Davis!

A couple of months ago, my Davis anemometer gave up. As my Davis system was old, I decided to retire it and I bought the WH3083 because my pension would not run to a new Davis system and, at 85, my mind is not as active, so I decided to run it on the WD I was familiar with. I was flabbergasted to find that the solar system on it, with freshly installed WD software (no legacy errors), did exactly what I experienced for years with the Davis solar; it glitched totally unpredictably but only on solar. IOW, I have problems with WD with two independent sets of weather systems on two independently installed versions of WD and have had so (with Davis only) for many years.

In your last post, you suggest I use Cumulus for both. This is NOT what I want to do. If you were in the habit of driving a Rolls Royce, would you want to use a Mini as your main car? In any case, I have not been able to configure Cumulus as a second system with Davis AND WH3083. Also, I use other features in WD that Cumulus does not have. I’ll continue to use WD with both systems, knowing that it’s not perfect for solar/UV, at least here.

Out of curiosity, I did start Cumulus to see how it behaved at the time of a known glitch in WH3083. The attached graph shows the curve was smooth. The red is a freehand rendering of the same time data from WD as in the table.

Finally, the thing that flummoxes me the most is that there are no ‘me too’ responses from other WD users. How can it be that I have had the problem for years and no one else has seen it?


I have just realised something about your solar sensor readings shown in your log files earlier in this thread, they all seem to show 2 decimal places. This is simply not possible with a Davis solar sensor as it only passes integer values, I’ve checked my log files and they all show values with .00 on the end. So we need to understand why yours have two decimal place values which are non zero.

Stuart

Stuart

Good thinking. However, you are right but with WD/Davis. The decimal points are only in WD/WH3083.

I did wonder after I had posted but was not completely sure whether or not those logs were Davis ones so left the comment to make sure.

Stuart

as I have explained, the solar glitches from the WH3081 solar data will be to WD not handling when the data gets of sync very well
I have for now added to ignore that when its night time, but I need to come up with a way to know if the solar data (which is a different memory location) is out of sync and so ignore it, for the WH3081 data

Thanks, Brian. It would be great if you could do that; I’d be very grateful. However, it is not just on the WH system that it happens, but also the Davis over several years. In fact, as I write (on a different computer; the weather one is dedicated), I’ve just seen a typical glitch on the Davis system as per attachment. This is a short one over 2 minutes, but I’ve seen occasional glitches lasting up to an hour, on both systems. On the attachment on the message of 24 April 2017, 14:13, there is a


Just to be completely clear on this … you are saying that on the 24th April you saw the exact same anomaly for the exact same time frame on both your Davis system AND your WH3081.

If true this would tend to indicate that at least on the WH3081 WD is actually reading the correct value from the sensor and it is highly likely that since both sensors are reading almost exactly the same values there has to be an external effect coming from somewhere to cause this.

This would be extraordinary since I am not sure how any of us could explain the external effect which could cause this at least not by the sun. I suppose it could be some electrical interference but again very difficult to explain.

Stuart

Sorry, Stuart, I express myself badly. I have never seen anomalies on both systems at the same time. They seem to be random on both systems. E.g., this morning’s glitch was only on the Davis system, but one I did not report here yesterday was only on the WH system, as some above. They happen at totally random times, day and night (except for Brian’s suppressing night time ones on WH).

I’ve not done any software development for well over 20 years, so have forgotten more than I ever knew, but my guess would be loss of sync over the 100 bytes transfer. Just one extra or missing bit would do the job. If it happened once at random, that is easily understood: less easy for my feeble brain is the same thing happening over many minutes.