Head_Gnome
Member Since: August 25, 2011
Country: United States
-
Epic failure. We epoxied a fitting to the top of the sensor. This worked great for about 10 minutes. When the fixture got repositioned on the bench, the pressure on the fitting warped the top ever so slightly. Turns out the sensor is extremely sensitive to the slightest structural deformation. So either go with iklln6's suggestion of small enclosure or epoxy a fitting to the board area around the sensor as opposed to the sensor itself.
-
When the rain gauge switches, the switch contact mechanically bounces. A fast IRQ routine counts some of those bounces thereby giving a false reading. There are two solutions:
Hardware: check out the MC14490. It's a great chip for solving this type of problem.
Software: add a 100ms interlock to the IRQ routine. Get the current millisecond time and compare to last time IRQ was called. If less than 100ms, do nothing. If more than 100ms, update time of last IRQ and increment count.
No public wish lists :(