I made this sensor a while back as a prototype and put it outside earlier in the year and it still seems to be working OK. It has been a dry spring and summer here in Seattle though and it really hasn't had much of a workout. It was my first project with the MSP430G2.
It is pretty simple. The funnel catches rain from a known area where it falls into two "buckets" that tip back and forth. Adjustment screws are used to calibrate the bucket volume. A hall sensor detects each tip of known volume and sends a signal to the microcontroller which timestamps and stores/transmits the data. Collected rain falls out the bottom through weep holes once it is measured.
It was cheap to make, here is a bill of materials:
- Hall sensor - less than $2
- Funnel: 1$ for three at the dollar store
- Magnets: I think I paid a couple of dollars for a tube of them
- Empty plastic nut container
- bottom of a coffee can to make the tip bucket mechanism
- Scrap wood
- Miscellaneous wire, nut, bolts, and nail I had around
I stole shamelessly from this guy and he has a good write-up so I won't repeat that here: http://www.instructables.com/id/Arduino-Weather-Station-Part3-Rain/
These device have been around a long time. When I was an undergraduate engineering student I worked one summer in a lab associated with the university where they needed to digitize rainfall data over a long period from several locales. This was in the days of mainframes, well before PCs, microcontrollers, and spreadsheets. The rainfall data was recorded on 24 hour charts that were attached to a clock driven drum and changed out daily. Each time the bucket tipped on the rainfall gauge it would make a tick on the chart. My job was to go through years of data (it had been stored on microfiche by the time it got to me), write it down and then later punch cards that were read into the mainframe. This project was more fun, but I appreciated the money at the time 