Solar anomalies in WD

I think that’s what is flummoxing all of us, how do you get such wrong but sometimes consistent errors over a few minutes, which is why we are clutching at straws. I suspect Brian is correct over the WH system. The Davis system is a different kettle of Surstr

maybe you could try covering the solar sensor on the Davis
and see if the spike still occurs
(also note that WD does use the CRC error check in the Davis LOOP data …i.e they needs to pass before WD will accept the data)

…and the console does a CRC check on the data packet from the ISS. It’s virtually impossible for couple of rogue bits to get into the data :?

@Devil: Have you always had the same Davis hardware, or has some of it been swapped out over the years?

The WH and the Davis are “chalk and cheese” as the Brits say. I think it’s confusing the situation to be looking for similarities…

I changed the Davis system for the same type at the same time as I bought the solar sensor: before that I didn’t have solar. It is a wired type with a fairish length run of Davis’ black cable (guess ~25 m) which is unscreened (unshielded for the US). Of course, the other system is wireless.

Regarding the Davis:

There are two separate CRC checks on the data before it hits WD making data corruption in transit extremely unlikely.

Solar is only present in 1/24 of the total data packets and yet solar is the only parameter affected. Statistically unlikely in a data corruption situation.

The spikes are very close to the full scale reading from the solar sensor that would result from the sensor output being equal to the supply voltage.

The sensor/ISS hardware has not been changed.

My conclusion is that there is a hardware problem either with the sensor, the sensor cable/connector, or in the solar measurement circuitry on the ISS.

I think Niko is correct
because in checking my code in WD for handling the solar data, I put in there that if the reading is over 2000 for 2 data points in a row, then WD set the reading to 0
which explains the zero readings after a spike

As there is so much in common in the software on the two systems, e.g., in the registry but also some functional parts of the two systems, I’ve decided, for clarity, to separate them entirely. I’ve uninstalled completely the WH system and Cumulus from my weather computer. I have therefore ONLY the WD/Davis system on my weather computer. On another computer, I have ONLY the WD/WH system, no Cumulus, with a totally clean installation. The only thing in common between them, now, is a stand-alone UPS.

Since the divorce and separate households, there has been no glitch on either.

Didn’t the Davis have spikes for a long time before you had the WH?

Yes, indeed, but I had a feeling (unsubstantiated) that it had become more unstable when both were on the same computer, sharing the same registry and some other items. Maybe a kind of symbiosis! :slight_smile:

Fair enough.

Have had anomalies on both systems since they were physically separated.

143143 is the Davis system and shows a single negative glitch
143610 is the WH3081 system and shows a 29 minute inclusive positive glitch

If you look at post #33, the UV index reading shows a 29 minute inclusive negative glitch (the solar irradiance reading, with a positive glitch, at 14:34 could easily be correct, within normal tolerances), also on the WH system.
Question: is it coincidence that we have had two separate glitches with exactly 30 error values???


You should remove all common denominators, including the UPS (for your testing purposes) and as you are getting interference from somewhere on both, you need to look ate what is in common and as youve said, they are both on the UPS still, so for this, i would remove the UPS totally, it could be the source.

they are not occuring at the same time
so not related,between the 2 systems, and as I have already pointed out, WD could be improved some what for how it handles the solar data from the WH3081 station

as for the 0 reading from the davis station that you got in the post above, I have already pointed out that if a raw reading goes over 2000 then WD sees that as a hardware glitch at sets the reading to 0 (if that raw reading is over 2000 for 2 readings in a row)

Definitely not the UPS, as it is brand new; the problems started years ago.

One problem with the new UPS has come to light. If the load on it is less than 15 W (my main weather microcomputer alone), it cuts out with a power cut. If I have a load >20W, it works fine. However, this is irrelevant to the thread, as I have a desktop on it as well as the microcomputer.

A 9 minute positive glitch on the Davis system (no decimals ID)


The more I consider what is happening here with the Davis system the more I believe that this is a bug most probably with either the solar sensor or the circuit board in the ISS.

If it were an issue with WD then I’m sure that someone else somewhere in the world would be seeing this problem and so far no one has reported this. I did start a thread on the WWXforum.net asking if anyone there had seen anything like this but no one has seen this.

The chances of interference just affecting the solar value in the communication between the console and WD is extremely low and the CRC checking would tend to eliminate these anyway. I suppose it could be interference in the cabling (it is a cabled ISS isn’t it) but again this would almost by definition affect more than just solar. So we come back to the sensor itself or the ISS board. If this is as I believe a VP1 then that rules out changing the ISS board unless you get very lucky finding one, I know how rare they are as I was looking for one for my VP1 a while ago.

As Brian has said on more than one occasion WD could be involved in the WH station errors and I believe connecting the two is wrong especially as the error times do not coincide.

Stuart

Thanks, Stuart.

Since my earlier post, the Davis system has gone berserk with both +ve and -ve glitches for no apparent reason (graph attached). I’m not too worried about it, as it will be going to the recyclers as soon as I have sorted out the temperature screening on the WH. Even if I have to replace the WH twice or thrice as often as the Davis, it is a quarter of the price. I’m satisfied that it is no less accurate than the Davis, as the superimposed curves track really well within cited tolerances (wind excepted as the Davis wind is kaput).

If, as you and Brian suggest, the glitches in WH are of a different origin to those in Davis, then I’m at a loss to understand. For me, it is too big a stretch of the imagination that I should get similar-appearing glitches on both systems only in solar and nothing in temp., hum., wind, rain etc., although I’m willing to understand the coincidence. Even more difficult to understand that it is only here that it occurs! I’m flummoxed!


2017-05-03_130342.png

the glitches with the WH3081 are likely software related, as I have suggested
(if you use the cumulus software instead then that might prove that)

Long shot: How close to RAF Akrotiri are you? Their over-the-horizon radar causes havoc on the amateur bands. Installed about 2003(?), you said the problem started years ago.

But as others have said, why only solar. . .?

My guess >60 km with a range of 800 m hills in between. I’m in a fairly EM ‘quiet’ area, although neighbours have detectable WiFi signals (plus my own, of course). As the problems can happen at any time 24/24, at apparently random times, I think external human EM activity is unlikely. Good thinking, though, as I’ve heard that UK bases have caused problems.