Cloud Sensor

Peltier Method

This cloud sensor is based on the H&R TM92HVC2178 Peltier cooler module run as a generator. I used an aluminum plate designed as a hand plate for a door push painted black. Plastic screws were used to clamp the junction because the instructions said to not short the top and bottom plates. The surface seems to be ceramic so I don't know why they said that.
The output with the Sun shining on it is -100 millivolts and on a clear night about +12 millivolts.

Building a Thermoelectric Sensor to Measure IR Sky Transparency for Cloud, Cirrus and Dust Detection, at Night (Lou Boyd, of Fairborn Observatory)

IR Thermometer Method

28 Oct 2004 A better cloud sensor may be a common IR Thermometer. The ZyTemp TN105 from Metric Instruments ($20) has the following readings:

Clouds	Temp F
total overcast & rain	43
daytime clear blue sky	-11
patchy clouds & clear sky	depends on where pointed

This sensor works around 10 nm wavelength and so detects heat. If you stand up from a chair and point the sensor at the chair it reads 80 deg F but a table top reads 72 deg F. The box says the range is -33 to +220 C (-27 to +428 F) so the above sky temperatures are within it's range. I believe that these units are individually trimmed at the factory and that the on board micro controller compensate for manufacturing variations allowing this non cooled sensor to perform much better than raw sensors.

The TN9 IR thermometer could be combined with a small PIC microcontroller and packaged in a weatherproof way and could send the sky temperature back to your PC using RS-232. A plot of sky temp vs. time would give a good picture of the cloud cover.

The view angle of the TN105 & TN9 IR sensors is what they call 1 to 1, meaning that at 1 foot from the sensor a 1 foot diameter circle is being sensed. This is much too wide to see clouds moving. By aiming the sensor down onto a parabolic mirror that angle could be reduced and as clouds drifted by the variation in temperature would give a feel for what was going on. The trick is finding a parabolic mirror good ar near IR (a slightly lesser requirement than for optical use, but not by much) and has a useable f/D ratio.

Edmund Optics does have Off-Axis Parabolic Metal Mirrors that I think would work but they cost hundreds of dollars.

Fixed Mount versions of the TN105 that are already interfaced for RS-232 (ZF-5283) are in development.

SparkFun IR Thermometer Evaluation Board - MLX90614

2016: Steve built a cloud sensor using a Melexis sensor, a picaxe 20X2, and a Python program to handle the data. The sensor is maintained at 40F or above to minimize condensation/frost. The data is on the web.

Back to Brooke's Products for Sale, Weather, home page

This is the [an error occurred while processing this directive] time this page has been accessed since 28 Oct 2004