Solar anomalies in WD

For years, I’ve had anomalies on the solar curve on WD, shooting up to the top of the scale, way beyond any expectations. I blamed Davis for a faulty sensor, which question was denied by the makers. This happens only with solar and it has happened on at least 3 computers with three different versions of Windows and goodness knows how many versions of WD.

Over the last 24 hours, I’ve had the same behaviour with the new equipment (different sensor, different connection, same software WD10.37544 and earlier).

The first attachment shows WD for the last 24 hours with the anomaly, on the WH3083
The second shows WD for the last 24 hours without the anomaly, on the Davis
The third shows Cumulus display for the last 12 hours without the anomaly, using the same WH3083 data as in the first.

I’m flummoxed!


By chance, I’ve happened to find an example of the same problem on the Davis system, last December, albeit for less time.


What shows in the logfile?

Thanks for your reminder! I haven’t done this recently (I’m having problems with my memory :frowning: ) but have just opened the 42017 vantagelog.txt around 5 o’clock when there was no sun. It is showing regularly batches of 1870.32 W/m


Strange, if it was only the Davis I would say it’s a cable problem, but if the new station is showing it too I wonder if you have some other data source for solar set up the metar or similar area :? Did you clone the Davis install for the new WD or set it up completely fresh?

In many years of Davis+Solar+WD I haven’t seen huge values like that. I have seen cloud reflection spikes but that’s not what’s causing this…

Is there any possibility that something is reflecting on to your sensor? Does indeed seem very strange.

Stuart

50% brighter than the noon day sun :?

Just wondering what a sensor might read if you used a mirror to reflect the sun on to it as well as it being in full sun… would it read 150% or not?

One other thought… would be good to see it on the actual console as my VP1 console does appear to have 5 digits in that area. Might be a big ask if it is totally random.

Stuart

Clear sky, yes, I think it could. However, looking at the first screenshot this was happening in low sun, and during the night.

One other thought... would be good to see it on the actual console as my VP1 console does appear to have 5 digits in that area. Might be a big ask if it is totally random.

Stuart


It’s a good question, I rarely touch the console, does it keep its own max solar?

If he has raw data logging turned on then he might see something there.

I think the absolute max that the Davis sensor can output would be 3 volts (= supply voltage). Calibration is 1.6 millivolts per W/m2 so the resulting solar value would be = 3000/1.6 = 1875 Watts/m2. That’s why I suggested a cable problem, intermittent supply short to output, but the new station doing it too pretty much shoots that idea down in flames :?

I’ve been over this again and see that Cumulus does not appear to see the extreme values. Some how we do need to see the raw value but if the only station is the new one I don’t know how to do this as I only have experience of the Davis.

One question I can think of is if both Cumulus and WD are running how is the data shared?

Stuart

Thanks, all, for trying to help me, much appreciated.

Had one anomaly this morning at 09:20. (Att. 1) on the WH3083 system. Note since the reboot yesterday evening, the intervals on the log file are correctly 1 minute, even for the anomalous reading. This latter does not show on the Cumulus software, only WD.

I’ve investigated the Cumulus Apr17log.txt file. The data is recorded only at 30 min intervals and it has no header text, so I had to deduce the meaning of the active columns. There was no anomaly on the presumed light column. I also found a file today.ini which has a lot of data that is not otherwise easily seen. The [Solar] shows a high solar radiation of 962 W/m


“I have only the Davis and WH3083 running simultaneously on WD for most of the time, and they share.”

How are they sharing? Both WD’s accessing the same serial port?

No! Davis uses COM1 and WH3083 uses a USB port!

OK, no station sharing, oops :oops:

All very mysterious…

Sorry if I’m being a tad daft here but I am still puzzled. You are saying that Cumulus does not have a problem at the exact same time as WD does have the anomaly. How is that possible with no port sharing or data sharing? Or have I completely misunderstood the issue?

Stuart

Stuart

If you look at the first data in post #3, you can see that the anomalies last for seconds on end. It shows perhaps more clearly in post #2 where the anomaly lasts continuously for perhaps 5 minutes or more.

No, the phrase you question is perhaps my poor ability of expression. All I was trying to say was that it appears that when there was an anomaly in WD, it didn’t show in WH3083. In fact, I’ve never seen one in it, only in WD, many times over years, which I blamed, incorrectly, on the sensor.

Yse I can see but you are saying that the #3 is using the same data but only 12 hours worth. I think this is where I am puzzled. If you are saying the #3 is using the exact same data to display how was this data collected and passed to both WD and Cumulus? That is my first problem to understand.

Stuart

Stuart

I feel our differences are more semantic than technical. Both systems sample the different types of digital data sequentially, so the actual moment that one variable’s data is recorded is quasi-unpredictable. However, here, I’m generally not talking of a single byte of data going haywire but of data bytes that are haywire, every time that single variable is interrogated over a period of time. In short, the timing of a measurement is almost unpredictable in real life. Think of waiting at an Oxford Street bus stop: the bus you are waiting for arrives after 5 others have passed but is full and you have to wait while another 5 buses pass before one comes that you can board. Each packet of data waits its turn to be interrogated for a nanosecond or so. If it ‘misses the bus’ for any reason, you will have a byte of wrong data at the end.

I’m no expert in what I’m saying but I think that the time of any event is ordered but appears random. In most cases of ‘my’ problem, tens or hundreds of busses may pass without stopping in each sequence of errors. Of course, I have no idea of the sampling rate (the frequency of the buses).

Whether a display shows 12 or 24 hours of data is not material because what you see is past history.

Since 01:07 this morning (I’m rarely a night owl but there are occasions!), I have reset both systems on WD only (no Cumulus nor my dirty fingers touching mouse or keyboard). So far, no problems on either WD or WH3083, but random problems can happen in a few minutes or after several days!

How to log the incoming raw data.

I’m thinking more of some local (electrical?) glitch causing event rather than bad data from the station(s), but I can’t imagine how that could target only the solar data :dontknow: The Davis at least has a conductor on each bus (data packet) checking all passengers have tickets (CRC checksum) so it’s unlikely one parameter could be corrupted in transmission.

Niko

That is interesting, thanks.
On the Davis WD, I got a list of 99 items 'final_data(nn):=nnn; which, of course, were meaningless to me.
On the Program Event log, I got 12 lines 'error with saving nexstormdatahour to registry Failed to set data for " ’
followed by two lines ‘Resetting warmest day values 06.00:31 17/April/2017’ and ‘…06.01:31 17/April/2017’ - whatever all that means.

On the WH3083 WD, both files remain blank; guess its looking for data from COM1 and not USB.

Of course, when I want it to glitch, it is now performing OK; just being patient for one, other or both to glitch!

BTW, the computer is powered from a 12 V source from a PSU fed from a filtered UPS, so glitches from that source are unlikely. In any case, the same fault has occurred over several years on at least 3 computers.

The 2 systems are still tracking each other remarkably well (between glitches!), other than temperature when the sun shines. I’m surprised how well! :slight_smile: