Cloud Sensor

Something that would be useful for me to have is a cloud sensor that I could have somehow wake me up if it got cloudy so I could close the observatory roof.

Commercial ones are expensive, rainsensors.com/ being the only one under €100 and it's a rain sensor, so I'll need to go the DIY route.

The theory is that the temperature of the sky is low when it is clear but warms up when it's cloudy. So by using an IR temperature sensor (something like www.melexis.com/Infrared-Thermometer-Sen...rs/MLX90614-615.aspx ) to measure the sky temp, a TMP36 or similar to get the ambient temperature, and a arduino I should be able to determine temperature differences that correspond to clear, overcast, and somewhat cloudy. And then get it to SMS me or something so I can close the roof before it rains :)

Has anyone else tried this?

Hi Albert

Having done some work with arduino in the past I looked into this myself in the past and stumbld onto a design by a stargazerslounge member:

stargazerslounge.com/topic/67790-arduino...ol-and-cloud-sensor/

Looks solid and he has documented the results.

Thanks, I'll have a look through that thread. Seems pretty similar to what I was planning on. I had a sensor in the attic from a previous project so I re-used that with an arfuino tonight the test the idea. Seems to work well.



This was from a clear sky to lightly overcast.

Looks like just a matter of putting it in a box and writing a bit of code around it to wake me up when it gets cloudy

I've messed a little with the focuser I have from JMI and sending voltages over Arduino PWM pins to it and a little GUI to control it. Though I'll need to get a chip to control the motor better so it can go in reverse as well

Is use of precision focus tools necessary?
Are we not after average sky temperature - does it matter that sky image is blurred?

I do see that the more light is focused on the sensor the higher the photon count and the better the measurement - but to detect significant difference between sky temperature and ambient do we need to be so precise?

Can't we just have an upward pointing sensor - even without any lensing (or very rough focus to avoid IR from nearby ground objects)?

I could be completely missing something fundamental but the above are questions I currently have.

Mark C.

The focusing isn't for the cloud sensor, it's purely an attempt to get as much out of the arduino as possible. No point paying £20+ for the board and not using it to the fullest extent of its capabilities. The sgl_observatory_automation group ( groups.yahoo.com/neo/groups/sgl_observatory_automation/info ) are using one as a cloud sensor, focuser controller and dew-heater controller.

You could conceivably use the arduino to do *much* more.

Alternatively, it would be cost effective to buy a bunch of the ATmega 168/328 chips and the minimum required peripheral components and have seperate circuits.

Yea the focussing is a separate project but will be using the same arduino board for economy. I have a JMI motorised focuser already and its controlled from a hand controller. I'd like to get that working via the laptop and possibly eventually write an ASCOM driver for it so software can do the focussing.