Greetings -
An acquaintance of mine made this sensor about 10 years ago as a side hobby and had passed away and I’m trying to re-create the sensor. I’ve been tasked with taking over sunshine records and need to replace the pre-historic sensor the local National Weather Service office is using before that one fails. The NWS no longer is interested in recording sunshine minutes but the media and other news outlets like to use the data.
So, I’m trying to determine parts. Here is what I know from the manual he had informally produced:
- Dual hermetic sealed planar silicon photovoltaic sensors with translucent diffusion lenses.
- Spectral response: 400 to 1100 nanometers
- Operating range: 0 to 1500 watts per square meter
- Operating threshold: 50 to 400 watts per square meter (via adjustable control knob)
- Output signal: +9 VDC (battery voltage) to Ground (during sunshine)
- The signal went to a Veeder-Root totalizer to record sunshine by seconds, minutes, etc.
- The sensor was supplied with an internal 9 volt lithium battery which needed replacement every 30 days (if operating from only that power source).
- During sunshine conditions, the signal is at ground (0 volts). During non-sunshine conditions, this signal is pulled up to approximately 9 VDC (battery supply voltage) through a 1 megaohm resistor.
Sensor description:
When correctly installed and aligned with the solar ecliptic, the signal output generated by the amount of sunlight falling on the lower exposed sunshine sensor is compared with the signal output of the sensor positioned under the shadow band. During conditions of sunshine, there is a differential signal between the two sensors, and the sunshine display will count and totalize elapsed time in seconds.
During the night, or when the sky is obscured, or when clouds cover the sun, the differential signal is greatly reduced, and the sunshine display stops counting seconds.
Thank you for any help you can provide - I’d be most grateful!! : )
Andy
Greetings,
The basic idea from what I gather, is that the device has two weatherized photosensors, one provided with a shade and one not. Shiny conditions cause a distinct shadow to be cast, resulting in a larger difference in the signal produced by the two sensors than in darkness or diffuse/overcast conditions, which state is indicated by a logic signal.
Though not rocket surgery by any means, a few difficulties do present in terms of trying to recreate such an instrument in a manner that produces results that are consistent with those yielded by the current art. It would be nice for example, to get similar sensors or at least info on those in current use; there are optical considerations (viewing angle, spectral response, etc.) which could potentially influence results, and I don’t recall seeing such things on the shelf in an environmentally-sealed format. Perhaps they do exist, but I’m not immediately aware of where to look. An approach that comes to mind as an alternative might be to build such a thing using a translucent switch boot such as the 335-1013-ND as a cover, in conjunction with a machined mount for a photodiode such as the 1080-1143-ND .
Regarding an application circuit, something on the order of the below is something that comes to mind as a starting point. It’s very simplified and not particularly well thought through, but it’s an idea. Basic concept would be to run the two diodes in photovoltaic mode across a few resistors, pick up and amplify the difference with a low-power instrumentation amplifier, and feed the result into a low-power linear comparator with a built-in reference. The ICs should consume maybe 100uA, so a 9V could potentially last a few months, or a set of AAs maybe a few years depending on how much current the totalizer input consumes. A person could hang a solar cell on the box to extend things as well, potentially.
At risk of taking things into the weeds, this notion is useful for me in understanding the importance of maintaining the data source. Should it fail, I’d not be surprised in the least to see headlines announcing that the sun went out… 
