Topic: Interesting observation on solar from WS80

[broadstairs]

I've been gathering data recently on solar readings from my WS80 and even with a calibration factor of 0.75 it is still giving a non-zero w/mē value some 30 minutes before sunrise, so yesterday had a reading of 1/w/mē at 04:48 (about 20+ minutes after dawn) when sunrise was around 05:02 for my location. Later in the day the actual readings are largely consistent with what would be expected for full sun here.

Stuart

[mcrossley]

I have had none zero readings at night, there is probably an element of noise in the DA conversion contributing to it. This is probably why Davis treat anything below 5 as zero. There is no real meaning in very low values.

Looking at the 4 second readings during sunshine, they also bounce around quite markedly when the sun is obviously pretty constant - no clouds.

[broadstairs]

Yes I'm inclined to agree but having been using a Davis for 15 years up to last year it did seem unusual.

Stuart

[Rover1822]

I have had none zero readings at night, there is probably an element of noise in the DA conversion contributing to it. This is probably why Davis treat anything below 5 as zero. There is no real meaning in very low values.

Looking at the 4 second readings during sunshine, they also bounce around quite markedly when the sun is obviously pretty constant - no clouds.

When you say bounce around? what is the range?. For those of us with solar arrays , we know that even on a crystal clear day, inputs will vary. I detect mine from the current output, which of course depends on the size of the array, but directly correlates to the W/M^2 .

[Bashy]

All my nights are zero from my WS80, today at around 4:30 it started at 0.1, sunrise for me today was 5am, sunset yesterday was about 20:45, and the WS80 hit 0 at around 21:20, i do have buildings about 100m to the east, that would slow it down and the same in the west but prob about 150m away, that would speed up the 0. My data times are "around" because of the '5' minute records, so i couldn't be precise, sorry

I don't know if you can see my station but here's the link, not sure how it works, but considering the cost, i am impressed with the setup, comparing it with the budget WMR968 back in the day, it exceeds it. I don't have it as the main station in WD so i cant give you any accurate (timing) data, only the every 5 minute you see in the link, hopefully.
https://www.ecowitt.net/home/index?id=75975

[Gyvate]

30 minutes before sunrise could be due to the so-called refraction effect at dawn (and dusk) time, i.e. depending on air conditions (clouds, water vapour, sand, ...) one can see the sun even before it physically rises above the horizon.
When you can see it, the solar sensor can also detect some light. Of course this happens in steps.
By the way, due to that effect, tomorrow morning, in Western Europe, just before sunrise, you might be able to see the sun in the East with the Moon in full ecclipse and even in "Blood Moon" condition in the West - one (sun) just before rising and one (moon) just after setting - provided the conditions are favourable and you are in a good position.

[mcrossley]

@Rover1822 - not sure from memory, I'd have to check again.

What is also a bit of a laugh is quoting solar rad to a hundredth of a Watt/m^2, when actually it changes in steps of roughly 1.2.

They would be a bit more credible if they rounded their values to something near their quantisation increments of say 1 W/m^2.

[mcrossley]

30 minutes before sunrise could be due to the so-called refraction effect at dawn (and dusk) time, i.e. depending on air conditions (clouds, water vapour, sand, ...) one can see the sun even before it physically rises above the horizon.

But the sun rise and set times (and the solar calculations) take atmospheric refraction into account. OK at a mean value which equates to roughly 0.5 degrees - the actual value depends on atmospheric conditions somewhat; the difference is not hugely significant.

[Rover1822]

You really have to take it all with a grain of salt. I mean it doesn't track, so it will never have the right inclination for a proper reading unless you do live in an area where the sun is directly over head. I mean it tries, but compare it to a hand held solar pyranometer .

I use the values produced for relative use as opposed to some belief in accuracy
.

[Gyvate]

30 minutes before sunrise could be due to the so-called refraction effect at dawn (and dusk) time, i.e. depending on air conditions (clouds, water vapour, sand, ...) one can see the sun even before it physically rises above the horizon.

But the sun rise and set times (and certainly the solar calculations) take atmospheric refraction into account. OK at a mean value which equates to roughly 0.5 degrees - the actual value depends on atmospheric conditions somewhat; the difference is not hugely significant.

This is the theoretical, ideal sunrise calculation, correct, but in order to have a real visibility forecast, this alone won't help (see dependencies). The sensor detects the real visibility which  rarely equals the theoretical one.

But beyond this it also depends on the elevation which is not (and cannot be reasonably) considered in the calculation. The Weather Services do not issue elevation-of-the-observer based values for sunrise and sunset for a given latitude (area) (which would need to be included into the [still ideal] formula). Maybe some consoles do - I don't know. I doubt that the Ecowitt consoles do as the elevation/altitude is not a value to be entered somewhere in the setup. I know that CMX and Meteobridge use elevation, but afaik for correct (or let me say rather better) pressure calculation only. And even then the actual topology of the observer isn't considered either. Simple example - mountain range at the horizon, even hills higher than the observer (~sensor) position in the Eastern (sun) direction.

[Gyvate]

You really have to take it all with a grain of salt. I mean it doesn't track, so it will never have the right inclination for a proper reading unless you do live in an area where the sun is directly over head. I mean it tries, but compare it to a hand held solar pyranometer .

I use the values produced for relative use as opposed to some belief in accuracy
.

Sure.
My remarks were rather meant to be qualitative than quantitative - explaining why some readings can occur rather than claiming they would be accurate. 

[Rover1822]

LOL, Gyvate, wasn't meant directly at you, just a general comment.

As far as when you start getting solar energy , I can start seeing miniscule output from my small hobby array (1.7 KW) well before sunrise, for the reasons you mentioned.

[broadstairs]

I think one of the main reasons I am seeing these early readings is because I am only about 800 yards as the crow flies from the English Channel and of course face east so that will likely generate these early very low readings, I know if I'm up around 4:30ish it does seem quite light to my eyes    Because of the way these programs work they calculate the expected max from sunrise and by the time that happens my actual readings during this time have risen to above the expected max until after about 20 minutes or so at least. Also working on a % of expected max to denote sunshine is highly inaccurate with such low values. I would suggest not allowing sunshine hours to be incremented for a minimum of 15 and probably 30 minutes after sunrise, the similar timing being set at sunset, however only Weather Display allows this, CMX has a threshold value in w/mē as a minimum to consider sun shining and I assume the actual value has to reach this before any increments of hours can begin assuming the % value is high enough.

All of this is very much an approximation unless you have something akin to a Blake-Larson sensor or one of the very expensive tracking sensors. All these home stations have sensors which record global radiation and this leaves them open to effects like cloud edge and other refraction issues.

Stuart

[DelChard]

My weewx records have been pretty consistent.
Detected radiation starts about 10 mins after start civil twilight, and ends 10 mins before the end civil twilight.
My lightning detector had a good test last night, sky was well lit up with no effect on the solar sensor.
I've rigged my radiation chart to display when the sun is 1 degree below the horizon. [ You are not allowed to view attachments ]