As describe in the website 'Bricolsec,' we use the diode's thermal feature to establish a really accurate sensor. We kept and reproduced the same approach with two small copper squares black placed on a hot diode balanced with a reference diode hidden and protected from sun light.
The most important thing is always to define correctly the casing to be compatible with the outdoor conditions. So we define a design close to a professional sensor and be focused on the following points:
- find a polycarbonate dome to avoid opacity deviation
- improve reference inertia => create a concrete space to hold the reference diode
- still provide so enough place to handle silica gel packet and reduce condensation*
*Condensation is a real issue, not for the electrical system really simple but to the optical filtration apply to the light. A professional sensor use a double skin to avoid this...
A first correlation's been done with the first design and was really good. The deviation in lower insulation level can be assigned to a bad inertia over the reference diode. It's precisely why we decided in the final design to implement a plug with a concrete cylinder to handle this component. (that mean, we don't have new correlation)
This final configuration (second design) is running for one year now and the sun power analysis is really interesting. In the following you could appreciate the sun diagram.
The sketch is really simple, it's just an analog conversion (with a ponderation from past consistent test) and an average measurement recorded each 10mn but refresh each second.
The main difficulty concern the electrical shield with voltage and current regulator, follower, converter / subtractor... I used a full Mega proto shield to realize this as describe in the following:
Comments
Please log in or sign up to comment.