Rain Detector

Pulsed microwave radar between 1 and 10cm is quite good for detecting rain as it falls.

Peter

Please define more precisely what do you mean by "rain measurement". Do you want just to (a) detect it is raining or (b) measure the amount of precipitation? There are many simple solutions for (a), a simple switch activated by the weight of water collected in a small vessel, detecting the change in conductivity of "something getting wet", etc. (b) will require collecting rain water and measuring it somehow. Weight, height of water in a container of known dimension, pressure generated by the collected water, changes in light transmission through the water, etc.

You should focus first on the physics of the problem. Once you know what to measure and how to measure it, the interface to the parallel port will probably be the least important part of the project.

1) PC sends constant stream of data from Parallel port 2) PC is left out side. 3) Rain kills PC 4) Stream of data stops. Rain detected.

Seriously though, I'd be inclined to use the serial port because it is simpler and needs fewer wires. Use USB and get power thrown in for free! However, the PC end is the least of your troubles (but don't leave writing the driver until the day before the project is due because it can take a while)

As for the detection, the water level detection in the Industrial Boiler Control I'm working on uses capacitance probes and is pretty sensitive. They work by using an oscillator whose frequency varies with the capacitance of a probe dipped in the water. Measure the frequency - know the depth. The concept will probably work with a large flat probe to be spattered by rain instead of a long thin dipper. Or you could have a bucket with a dipper in it.

Of course I have seen simple rain detectors which are just two wires beside each other. When a drip bridges the two then the circuit is completed.

Look at web sites for commercial irrigation systems. Most of them have a rain detector nowdays to prevent watering when its raining or has rained recently. Get some ideas from them and then design a better one. Conductivity through collected water and periodic cup dumping are common.

(b-1) Measure the total rainfall over a period of time (e.g. total inches the past 24 hours).

(b-2) Measure the instantaneous rate of rainfall (inches/hour).

--
Grant Edwards                   grante             Yow!  People humiliating
                                  at               a salami!
                               visi.com

In a cold climate, you should define how to handle snow and sleet and how to prevent the collected water from freezing at night.

In a hot climate you have to consider about evaporation especially if the air becomes dry and windy after a shower.

Paul

An interesting technique is to count the impacts from raindrops in a given area, using a suitable transducer. It can be quite accurate, apparently.

Leon

Do also have to measure the impacts to get an estimate of drop size?

--
Grant Edwards                   grante             Yow!  .. Should I get
                                  at               locked in the PRINCIPAL'S
                               visi.com            OFFICE today -- or have
                                                   a VASECTOMY??

I can't remember. It might be accurate enough if an average size could be calculated.

Leon

You would have to apply the Marshall-Palmer-law and the Laws-Parsons-distribution to get a better relation between how much is raining and the drop size distribution.

Paul

The 10 cm (3 GHz) wavelength might be usable for detecting huge hails, but even a heavy (150 mm/h) rain will attenuate the 3 GHz signal with only 0.1 dB/km, so I very much doubt that you would get any usable reflections from the raindrops.

Frequencies between 10 GHz (3 cm) and 100 GHz (3 mm) would be more suitable, although it might be hard to separate between drizzle and fog at the higher frequencies.

Paul

Picking up on this point - I'm interested in making an accurate aquarium level detector at the moment (saltwater) , and wonder how good this method could be. In an ideal world something like 1 mm resolution would be good. Could this method achieve this. Any suggestions on the general type of oscillator and ballpark frequency that might work best?

Steve

scilent-project scrobe on the papyrus:

No, it seems from your post that you want the good folk here to design it for you, so that you can take the credit. Go away and design something, then come back here when it doesn't work.

--
John B

Oscar scrobe on the papyrus:

.

Ah, the Chinese water clock.

--
John B

Steve - what's the depth of the tank? ie 1mm in what total?

For a possible start, a double sided pc board cut to 10mm by ?mm (radio shack?) with a capacitor, inductor & transistor soldered on and some paint should give you first info.

Keep in mind that the dielectric constant of water is around 70 & epoxy board material is about 4 (check both).

Hul

Steve wrote: > > >As for the detection, the water level detection in the Industrial Boiler

biological control unit, definitely.

use a pic with usart, program that and hook it up to the pc serial port via a long serial cable that goes outside on the front porch. use level converts so you get a good distance on the cable. attach the pic to a switch. when the switch is activated that sends a signal to the pc that does whatever you want it to do. set a big glass of tea on the porch to attract control unit (i.e. your mom) out on the porch. when she sees rain she pushes the button.

alternatively, for more fidelity, you could have a pot and use a/d conversion to get a range of values. label the pot something like "ain't raining", "ain't much rain", "little bit a drizzle", "it's rain'n perty good", "raining a lot", "cats and dogs", and "i think i just saw toto". then instruct control unit about the finer points of polling vs. interupt driven i/o and instruct control unit to periodically poll (i.e. look) and see how much rain there is and to use the pot to select the proper value.

Thanks, it's a pretty shallow tank, about 30cm. I'd need to make a non-copper electrode design for this as there are inverts in there. I was thinking about a stainless probe or pair of to do the sensing.

Steve

Steve - what are inverts?

Hul

Steve wrote:

Ah, a bit OT I guess, but invertebrates, as in non-vertebrate organisms, include corals and other non-fish marine things - hence the need for precise level (== water specific gravity) control to keep them. Copper in greater than trace concentrations is highly toxic to marine invertebrates, hence you need to avoid same in any electronics in contact with aquarium saltwater.

Steve