[danner]

Nope it is still doing it. The other 2 stations near me with pyranometers reached 0 w/m2 at 7:10pm and 7:24pm. The 7:24pm station is directly north of me so my sunset should be similar.

I'm going to set the clipping at 15 for now.

[danner]

What is the best method to measure the output of the SP-110, between the positive and negative screw terminals? Or is there a test point somewhere on the board to sample from?

[werk_ag]

The best method to measure the output of the SP-110 is disconnect if from the board and measure directly from the wires. Keep the black and clear wire together. Use a millivolt scale

As a last recourse, you can also try reduce the gain of the OP by slightly change the value of R2.
Take the Solar / UV interface schematic, measure its actual value between GND and pin 2 of the OP and carefully increase its value 100 Ohm at a time.
Do this with the Solar / UV interface unpowered.

[danner]

The best method to measure the output of the SP-110 is disconnect if from the board and measure directly from the wires. Keep the black and clear wire together. Use a millivolt scale

OK, I'll have to find my alligator clip leads.

[danner]

Found my clip leads and measured the SP-110 with a black cloth on top of it (there is a full moon out and my neighbors outside light was on) and measured 0 volts. Removed the cloth and still 0 volts. I shined my LED flashlight directly on it and kept it there for 15-20 seconds then removed it and the meter went back to 0 volts. I repeated shining the flashlight and removing it to see if it would output any spurious readings and each time it returned to 0 volts.



When you get time what is the next step, reducing the gain?

Thanks for the help.

[werk_ag]

Yes, try to reduce the OP gain by setting R2 to around 4900 Ohm. After that, do an additional test: Leave the pyranometer disconnected, and connect the input of the interface (+) to ground. Leave it running for 24 hours or so. Then check on the graph if any spike was recorded during that period.

[danner]

Thanks for responding. I assumed that was the next step so I reduced the OP gain this afternoon. I measured 4.70k before adjusting and did my best to increase it 100 ohms, but the closest I could get it was 4.81k ohms. It is extremely twitchy, but I figured it was going to be from trying to adjust the pot for the FO wind instruments.

Today was very overcast because we have heavy rain approaching so the maximum reading was well down from the theoretical max. The graph does show the output taper off to 0 w/m2  after the sun went down instead of the reading floating at 12 w/m2 for hours. I am getting the occasional 1 w/m2, that you previously explained why it could happen, but so far no spikes above 1 w/m2 at night.

[werk_ag]

That's good. Keep monitoring without applying any clipping yet.
If the occasional spikes never go above 1W/m2, you can then set the clip to 1 or 2 (you know where you have to do it).

As we have reduced the OP gain, the readings now will be somewhat lower than before, but this can be compensate on the software. Wait for a clear sky day, preferabily at full noon (as you know) and increase the 1600 value, to 1650, compare the readings with the nearby stations, and if necessary fine tune the value until you get a reading close to your reference station.

[danner]

Here's the Cumulus MX solar graph from last night. The further into the evening it went the more the solar sensor increased the number of 1 w/m2 readings. Nothing was going on in the house any different at 8pm than at 8am so I cannot think of any additional electrical device being turned on to cause interference.

Frustrated that the readings continued to climb ever higher I swapped out the U1 battery and trickle charger for a 12VDC regulated power supply at 5:00am. This had no effect on the readings and they continued to increase until sunrise.

Official sunrise was at 7:48am.

[werk_ag]

It seems clear that you may need to reduce the sensibility of the system even a little more.
I think you can't expect that the readings will be coincident to theoretical graph or that they start exactly at sunrise time, some time before that, already there are enough radiation (IR + visible light) on the sky for the sensor to start to read.

I'm somewhat surprised by the frequent use of the expression "Frustrated"...
Sorry, I wonder if you have read what is said on the topic where the Solar /UV interface is announced:

"Please bear in mind that to a successful build of this sensor, you may need to spend a considerable amount of time and effort.
With the Solar / UV interface you can build a reasonable solar radiation sensor, but don't expect that this task will be just a matter of connecting the parts and that's all. You may need to spend a considerable amount of time and effort to calibrate it.
For installing it in a suitable case, you will also need a lot of ability and imagination.
If you are not sure about what you're doing, PLEASE, don't order this device. "