Via the Adafruit Forums, user Govner shares a build of a high precision rain gauge:
I have never been happy with the resolution or operation of “rocker-type” rain gauges. Moving parts and constricted water entry openings are inevitable sources of problems. Why not have a no-moving-parts, water-proof detector. “Waterproof” ? How can that be? Enter stage left, a piezo crystal detection unit.
This unit is capable of counting rain drops at a rate (about 2 usec between drops, est.) that far exceeds the world’s record rainfall rate (Google: South America 1.5-inches/minute). In any case, my goal was not worrying about the fact that research assigns 1 uL/drop or some such thing. I just wanted to get the big picture so I could inhibit irrigation pumps when enough rain warranted that. And with a simple evaporation algorithm, re-enable the pumps. All control had to be wireless so I just added another NRF24L01+ 2.4Ghz transceiver ($2 USD) to my existing network. Piece of cake. I used a glass mason jar which is used for preserving foodstuffs and is completely waterproof. I used a glass one (as you can see) but plastic would have been easier for drilling the small hole in the bottom for DC power. I used a diamond drill bit and it took all of about 30 seconds. No biggie.
Happy to share anything/everything done on this project. I used Pro Mini, 4 piezo elements, differential amplifier, Buck power supply (I like the buffering between my solar panel system), BME280, NRF24L01, DS3231 RTC.
I just wire-wrapped the assembly on a PCB that I designed for general purposes like this. I recommend using electrical contact grease around the lip of the glass jar just to get a great, lasting seal. (That’s another reason why I chose not to use plastic. Glass has a far superior long-term reliability in this regard).
The unit works far better than I expected and has other applications. The processing speed is so nice that, when I did the “shower stall test”, I could increase the droplet rate and see the drop-to-drop LED response. At the maximum rate, the LED just can’t keep up with the rate and appears to be lit steady.
See the forum post for more and pictures below. Nice work!