Topic: Calibration for solar irradiance with WS90

[luk5566]

Hi,

I've recently got the GW2001 kit from ecowitt and I've set up WS90 in a temporary location. I'm also using ecowitt2mqtt open source software to send the data to home assistant.

But I'm also using bird spikes on ws90 (I have lots of tiny birds around here). So I wonder, has anyone had much success using a static multiplier on the solar irradiance with this sensor package and bird spikes?

I worry it is much more complex than simple "multiply by 2". The spikes obscure, but they also reflect. Maybe painting them black would be better?

I hope others have already found a good solution so I'm posting this. I've searched the forum, but I found nothing.

[Gyvate]

my solar readings with gain factor 1.0 stay regularly very nicely within the theoretical maximum for solar radiation for the longitude/latitude combination of my location.
(the lower readings yesterday come from overcast sky  )
 [ You are not allowed to view attachments ]

[luk5566]

my solar readings with gain factor 1.0 stay regularly very nicely within the theoretical maximum for solar radiation for the longitude/latitude combination of my location.
(the lower readings yesterday come from overcast sky  )
 [ You are not allowed to view attachments ]

Do you have the bird spikes installed?

I'm not sure because there wasn't any really clear sky since I installed my ws90.

Perhaps it is fine, but how can this obstruction not affect it?

Here it is compared with my PV output that does peak at 2280W in absolutely best conditions (panels point 145 degrees, my latitude is 53.5N, angle is 40 deg, panels are nominal 2400W and they reach this on the inverter input, but the chart is AC output so it suffers from 5% losses, therefore the max is 2280W at best blue sky in summer).

 [ You are not allowed to view attachments ]
 [ You are not allowed to view attachments ]

At 12.04 it was producing 1767W (so 77.5% of max). For my latitude the max clear sky solar irradiance on a flat plane is 766W. 77.5% of that is 593W. The actual reading was 507W.

So perhaps I need to multiply it by 1.171?

[Gyvate]

careful - you cannot simply compare your PV panel with the solar (light) sensor of the WS90 1:1.
Many factors to be considered before you can make a "direct" comparison.

Already the angle by which the direct portion of the irradiance hits the panel and the light sensor of the WS90 are different.

What do you want to adapt ? To the maximum solar radiation ? This value is for ideal conditions, no clouds, no or very little water steam in the air (which you can neither see nor tell).

Also, the irradiance determination of the sensor is (as it measures light) gained by photometric analogy for the human eye maximun. So it cannot be easily compared to your solar panel.

YOu want to deep dive into the matter, you need to study a bit more  .
We've also written something in our WiKi about it. (see signature)

I would just leave it as it is. But of course you can do as you like. Whatever makes you happy ...

[havtrail]

I believe the question is whether he needs to compensate for the shadows cast by the bird spikes...

[Gyvate]

I believe the question is whether he needs to compensate for the shadows cast by the bird spikes...

1. they don't significantly impact the light reading
2. even if, one would need an interval compensation depending on sun position - which is not supported.
You just use a fixed factor - 1.0 = nothing or > 1.0 or < 1.0
3. the sudden drop in the graph is definitely not caused by a bird spike ...
Rather by clouds or a physical obstacle

but adjusting the gain such that the maximum solar radiation is always reached as seemingly suggested by the OP is not a good idea.
Who says that the light readings are always the theoretical maximum....

[luk5566]

I didn't consider different spectral response of the sensor vs solar panels. I thought it is measuring solar irradiance with a tiny solar panel. But as it isn't it makes sense that the results would be different.

I suppose water vapour will scatter/attenuate visible light similarly across the entire spectrum (because it has no visible color when clean), but add some pollen or dust to the mix and it is a different story.

I'll leave it as is. The dips are almost certainly caused by the shadow of the bird spikes as they seem to repeat daily (when sun is shining) at almost the same time. There are 3 during the day and the last one was at 14:45 the day before yesterday , next day it was at 14:52, today is too early at the moment.

So it is definitely not ideal. I'll have to figure something out for it as it is not the same amount taken every time. If there is heavy cloud cover the dip will be minimal. In full sun the shadow will be sharp.

If anyone has got a solution other than removing bird spikes I'd love to. Hear about it.

[SamiS]

Unfortunately I do not have a suggestion how to totally fix it except removing the bird spikes.

But for me it makes perfect sense, that the shadow from the spike most likely will show up in the readings, since the actual sensor is physically very small. The effect most likely is different depending on the geographical location, season and time of day, that all affect the angle of sunshine.

There is still something that you can do. You could rotate the spikes a bit so the shadow does not hit the sensor at the most problematic time. You also might consider removing one spike from the problematic side. I think that if you can climb to the sensor during the problematic time you can visually see the shadow and react accordingly.