Hi,
I am using the usual arduino rain sensor which relies on the conductivity of water to detect the presence of rain. However only after a few moths of exposure to the elements of nature, corrosion renders the sensor useless. Have a look at this picture bellow :
Can barely see see the cooper tracks on the pcb any more.
Anyone got any experience with other types of sensors that can last longer? What do cars use for rain detection?
The usual connection to those sensors has them powered up all the time, which leads to what you observe.
To minimize corrosion try powering the sensor from an output pin, and turn it on as rarely as possible, for a very short time.
Rain can be detected without electrical interaction by detecting the change in oblique reflection from a glass surface, but that requires sophisticated optics and electronics.
jremington:
The usual connection to those sensors has them powered up all the time, which leads to what you observe.To minimize corrosion try powering the sensor from an output pin, and turn it on as rarely as possible, for a very short time.
Yes I thought of that too. Switching on the current when needed, makes the sensor lasts longer but not that long.
I think it also has to do with acid rain.
Rain can be detected without electrical interaction by detecting the change in oblique reflection from a glass surface, but that requires sophisticated optics and electronics.
Is that what car windshield sensors use?
How about a capacities sensor something like the soil humidity sensor but on a bigger scale?
Some interesting possibilities turn up if you google "non conductive rain sensor" or "optical rain sensor".
Another possibility is to use a piezoelectric element. This person used one to turn the kinetic energy of raindrops into LED light:
Maybe a less extreme implementation would be to attach a piezo to a square of roofing tin and use it + an amplifier as a microphone... sample the volume/spectrum on the mic regularly to detect raindrops.
This would have the advantage that you could quantify the intensity of rain, but you might have to do some filtering to reduce ambient noise. It also won't detect moisture as such; damp fog, mist or drizzle might not register, for example.
I’ve just bought a used optical rain sensor off a car - used to turn the wipers on/off.
Worth a look ?
hammy:
I’ve just bought a used optical rain sensor off a car - used to turn the wipers on/off.
Worth a look ?
Probably yes!
I ve see some on ebay.
But woulnd know how to connect them. There is no diagram.
Which one did you buy? Did you use it?
Aaah there is the rub
Most are connected via the LIN bus to a control module - I’m interested in finding out how to do that . If that doesn’t work I’ll look at taking it apart and just using the sensors/optics
LIN bus? That makes things more complicated . Need to get into the protocol etc.
If you manage to tap on the sensor itself inside, please let me know. I d be very much interested in this as no ither viable solution seems available. ..
I d be very much interested in this as no ither viable solution seems available.
Why are sensibly priced, professional quality optical rain sensors, like this one, not "viable"?
Thats looks like a really professional solution.
Unfortunately the total cost including shipping to EU, where i live, and taxes is over 100 euros which is more than am willing to pay...
You could try making your own , there are some optical rain sensors on you tube .
Eg
That might show how to use the sensor part in the car one ( my plan B)
LIN bus shouldn’t be hard ... it’s only serial stuff, and similar to CAN .. there is info out there and shields for it too , just not had time to look in detail /find someone who has done it !!
As an alternative , if you made a grid from stainless steel wire and arranged your circuit to :
A) . Only power up and take a quick measurement , say every 5 seconds.
B). After each measurement reverse the power supply to the grid for the same duration to reduce electrolytic action . Have a look at H bridge for the reversing .
Or ..
A “tipping bucket” sensor makes a simple device for detecting rain. There are probably spare parts for the maplin one on eBay or make one ..
I am considering this now :
More info here :
Do you need to measure volume of rain, rain rate or just rain/no rain?
It would be nice to be able to measure the intensity of rainfall but it looks like i will have to settle with just rain/no rain..
You could try two temperature probes, one exposed and rain and one not. When one gets wet it’s temp should drop slightly. Might need some thought around ensuring even sun and wind exposure.
warrenboyle:
You could try two temperature probes, one exposed and rain and one not. When one gets wet it’s temp should drop slightly. Might need some thought around ensuring even sun and wind exposure.
Good thinking !
Although it sound viable as a principle, I d rather use something proven that works instead of doing R&D on my own! 
Humidity rises to 100% when it rains so there will be little difference between the two probes.
If you have heated surfaces , you would see a difference in the power required to keep them at an elevated temperature .
