Topic: Davis station solar sensor reading very excessive spikes

[broadstairs]

I am just wondering if anyone here with a Davis VP station and solar sensor has ever seen readings spike to ridiculous levels like 1800+ w/mē? There has been a thread running elsewhere (which some here may have seen) where one user has experienced such spikes for a very significant number of years, so much so that they gave up on using the sensor. Now (as some here may know) I have had an interest in solar values from Davis systems for some years so I thought I'd ask here if anyone has seen spikes to this kind of value? I am well aware of how values above the expected maximum can occur for a location and time of day, but these are different in that they significantly exceed the theoretical maximum for the value at the top of the atmosphere which is, theoretically, impossible.

So has anyone here experienced this and if so under what circumstances (if you happen to know) and what software was being used to record these values if they happen?

Thanks   Stuart

[chief-david]

see if other sensors give those weird spikes. Look at graphs too.

Post a few here to see if it is the sensor or the board.

[broadstairs]

We've already spent some time trying to determine whether it is the sensor/software or transmitter board but not been able to definitively determine. The person with the problem did involve Davis themselves originally when it first happened some years ago but even they gave up.

I am just spreading the net wider on the off chance that someone else somewhere in the world has seen such spikes, so far we have one person worldwide reporting this problem. So for now we just need to know if anyone else has seen such an anomaly.

Unfortunately this person has another station type which also exhibits problems with its solar sensor although not so far at the same time as the Davis, however it has been felt that this other station may have an issue with the software which reads the values from it, so for now this is inconclusive.

Stuart

[dalecoy]

Have you asked Weather Underground if there have been such spikes?

How about CWOP? 

[I have never seen such spikes]

[weatherc]

My "9 year alltime record" for solar reported by WD is 1753 w/m2 using VP2...
As is highest gust of 72.9 m/s (163 mph) and highest avg windspeed of 34.5 m/s (73 mph) too...
All have different dates...

[waiukuweather]

certain clouds positioned in the right part of the sky in relation to the sun can reflect the sunlight, giving 2 bits at the same cherry (i.e can cause like a 50% increase), this can be seen when monitoring solar energy panels too
weatherc,what are the dates on those records?
as WD has got better at filtering out spikes with the Davis data over the last couple of years of versions
(you can reset or enter the needed correct values)

[ct]

About 3 months ago my solar sensor reached 1705 w/m².  The UV reading at the same time was 14.2.  The government agency which measures UV (ARPANSA) also had a reading over 14.  The UV forcecast maximum for that day was around 10, but with broken clouds it had a 40% increase in UV and presumably solar radiation

[Bashy]

I have sort of followed the topic on WW and i have seen the sort of spikes that you are on about,
Some of the spikes appear at night time 10, 11pm, i think this is an interference issue....

[broadstairs]

Thanks for the comments everyone. Just a couple of my own.

Firstly I am aware of the reflection issue known as edge of cloud effect and it is generally accepted that this can increase the reading but something up to 20-25%, this coupled with the maximum possible at the top of the atmosphere of about 1370w/mē means we should not see values over 1713w/mē assuming no atmospheric attenuation which is impossible anywhere with the possible exception of somewhere like the summit of Everest!

We have come to the conclusion that this could be an external effect such as interference but when the solar sensor provides a voltage and the sensor is in a domestic situation it is difficult to see how or why this might happen especially when these spikes can appear for several consecutive minutes. We really need to see the spike on the Davis console to know if this is caused by the Davis system and if not we can then review the software being used.

As the software author has recognised that for some non-Davis stations there could be a possible s/w issue and for this persons WH type station he has some spikes at night time. I think for now we can only consider the Davis situation. I have had issues with readings at night time with my own Davis system which were caused by a failing sensor.

Ideally to progress this issue we need either to see the console values at the time or perhaps somehow capture the raw data being sent to the PC from the console prior to any s/w processing. That way can tie down a bit more where these spikes are coming from.

Stuart

[weatherc]

weatherc,what are the dates on those records?

They are from 2013-2016. I have not checked if they are in WD-data only or if they are in WL-data too tought [offtopic](would need a functioning re-import of WL-data to WD for that what i actually mentonied in the bugreport-thread on WW some times ago). [/offtopic]

[archae86]

Firstly I am aware of the reflection issue known as edge of cloud effect and it is generally accepted that this can increase the reading but something up to 20-25%

Source for that limit?  I've been seeing cloud reflection high excursions well above the zero-cloud "maximum" ever since I got my Davis solar sensor years ago.  As they show up with clouds, and don't otherwise, I don't think it is some kind of mysterious RFI or other oddball effect.  Sadly Cumulus does not seem to be maintaining records for my station regarding solar (maybe I'm missing a configuration item), so I can't supply my personal maximum.  However in brief browsing of my old data I come across a June 2015 case in which I logged slightly over 1350 Watts/square meter in a period in which my no-cloud value would have been just under 1000.  So that is over 35%, which rather exceeds your "up to" limit.  I doubt Albuquerque cloud cover is ideally configured to maximize this effect.

[broadstairs]

The 20-25% is only an estimate, however it is based on the recommendations when estimating the capabilities of solar panels for generating electricity where they recommend you increase capability of the circuits involved. I suspect it might just go a little higher but not much. If you Google the 'edge of cloud effect' you will find a lot of references. The 1370w/mē values is the maximum  value at the top of the atmosphere, so not only no clouds but no atmospheric attenuation at all - zero. I very much doubt values of over 1800w/mē are possible. Also when you consider this user has seen values like that order of magnitude when the before and after values were perfectly reasonable and around 700-800 or so. One other point about all of this is that you must also consider error % of the sensor itself which I think about 5%.

Stuart

[R_o_B]

Stuart, I have been following the posts in the other forum...

The situation is not with the high readings but, as you indicated in your first post, it is with high level spikes that can last for a few seconds to many minutes.

I cannot remember if the spikes also happened at night time (knowing that the specific weather software used has recently been modified to ignore pulses/spikes during the night).

Unless I have really misunderstood your first message, you are looking for other weather station operators that might have noticed 'pulses/spikes' in the reading of the solar values when the solar reading values should be fairly uniform...

[broadstairs]

R_o_b yes that is what I am looking for primarily. We are discounting the night time errors for now as at least some are happening on the 'other' station. It is a real puzzle and has been plaguing this person for some years. We are all baffled and I was just hoping widening the search might just turn up something or magically someone who had this and fixed it 

Stuart

[waiukuweather]

@weatherc, the older build 81 might work better for you for wlink import
@broadstairs, reflection of clouds can mean a value higher than the maximum that comes into the top of the atmosphere
why would it not?. Its like having a mirror in the sky

[LFWX]

@weatherc, the older build 81 might work better for you for wlink import
@broadstairs, reflection of clouds can mean a value higher than the maximum that comes into the top of the atmosphere
why would it not?. Its like having a mirror in the sky

...or multiple mirrors reflecting from multiple locations...in addition to the radiation coming directly from the sun!

[broadstairs]

I think that most if not all of the possible sources of terrestrial reflection have been discounted, this is a quite normal domestic situation and the chances of reflections from anything other than edge of cloud effect are negligible and it is also unlikely that any such reflections might provide the time span where these rogue readings can happen (from a single spike to several minutes).

I'm not sure that values that have sometimes been seen could be accounted for by edge of cloud effect, yes it is possible for readings to exceed the 1370w/mē but from my research on a different topic it seems very very unlikely that the values sometimes seen could be valid as the conditions either side of the spikes do not indicate gin clear conditions.

Stuart

[dendrite]

Are you seeing any high spikes under perfectly clear conditions? I'm up around 43.5N and my highest is about 1500wm^-2 with towering cumulus. Generally my spikes above the solar constant last a couple of minutes until the cloud repositions itself. If you're getting the spikes with clear skies there's definitely a problem.

[broadstairs]

I think we have decided on a hardware issue with the Davis station having eliminated most of the possibilities. The person with the problem has an old VP1 which has exhibited the issue for some years and acused him to turn off the solar because of it until recently when his WH station had similar issues. Howvere the WH station problems are likely a complete red herring in terms of the Davis issue. The Davis system will be soon retired anyway now.

Thanks for all the suggestions and advice.

Stuart