EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: ledtester on September 30, 2021, 02:57:09 am
-
I seem to recall that there was a relatively recent thread on different ways of detecting when a drop of water hits a sensor. I've done a search but I can't find it.
Does anyone recall the thread or have ideas?
My scenario is that a ceiling tile is getting hit with water drops from an upper floor and I'd like to know when it happens so I can correlate it with human activity.
-
I don't know the topic, but any recording device you may have might be enough
- a light bowl sitting in the fake ceiling, record the audio noise made by the drops
- a LED shining on a cling film (or on Al wrapping foil) that will catch the drops, record the changes in the reflected light with a webcam
- two electrodes stuck in the fake ceiling tile, measure the resistance with a logging capable DMM (or make an audio oscillator and record the resulting frequency with a laptop)
- remove the wet tile entirely and film the pipes instead
-
One could detect the drop hitting the surface directly, e.g. with some piezo under a hard (e.g. glass or metal) target surface. This may need some hight of fall to get enough sound to get above the normal environment sound. There is a chance for false alarms from loud noise if too sensitive.
The other point would be detecting the presence of water, like via conductivity in some paper of the like. It may take quite some time for the system to work again once it is wet.
-
Replace the tile with a polycarbonate sheet and point a camera at it.
-
Commercial devices are made for this. They have a tape (looks like medical adhesive tape) with wires in it, and some salt substance that becomes conductive when wet. I suspect you could make up something with common household items. Then, rig a transistor to amplify current, and a Sonalet beeper and a 9V battery. Put it all on top of the ceiling tile and wait for the beeping.
Jon
-
Commercial devices are made for this. They have a tape (looks like medical adhesive tape) with wires in it, and some salt substance that becomes conductive when wet.
...
I went searching and found this:
https://www.dwyer-inst.com/Product/Level/WaterLeakDetectors/SeriesWD (https://www.dwyer-inst.com/Product/Level/WaterLeakDetectors/SeriesWD)
On another sales brochure (https://www.dwyer-inst.com/PDF_files/WD.d.pdf) it says:
The small and discreet Model WD Water Detector is designed for dependable
detection of low levels of conductive liquids. The module features a sturdy and reliable
aluminum enclosure and is powered by 24 VAC or 24 to 30 VDC. Water sensing tape
attaches to module and if any liquid comes in contact with the tape the resistance is
changed and the alarm will be triggered. The tape is hydrophobic so it does not absorb any
of the liquid it is detecting which makes for a faster drying time and faster return to service
after a water leak.
The sensing tape is 1˝ wide and can be bought in lengths of 5´, 10´, 15´ and 25´. Multiple
tapes can be connected together to extend the coverage area which makes it ideal for
domestic as well as commercial applications. Features include power and alarm LED’s,
alarm test switch, continuous tape integrity self check and extendable tape sensor.
-
Commercial devices are made for this. They have a tape (looks like medical adhesive tape) with wires in it, and some salt substance that becomes conductive when wet.
...
I went searching and found this:
https://www.dwyer-inst.com/Product/Level/WaterLeakDetectors/SeriesWD (https://www.dwyer-inst.com/Product/Level/WaterLeakDetectors/SeriesWD)
Hmm, the detector is kind of expensive, and the tape is a bit pricey, too!
Maybe there's a second source? eBay, alibaba?
Jon
-
Here's something more affordable:
https://www.networktechinc.com/tape-style-leak-detector.html (https://www.networktechinc.com/tape-style-leak-detector.html)
-
An older Nissan I had used a disk sensor about 1" diameter in front of the windshield. Not sure if it was piezoelectric or a dynamic mic, but it used the sound of the rain hitting the disk to control the wiper speed. Worked quite well. Your problem is not knowing where the drops are going to hit, so you need a bigger target like a sheet of plastic film.
-
"My scenario is that a ceiling tile is getting hit with water drops from an upper floor and I'd like to know when it happens so I can correlate it with human activity."
OK. Quick and dirty.
This is what I use in my basement. Very cheap at the local hardware store. Battery last over one or two years. Loud alarm.
Wire the two electrodes to a protoboard with a parallel wire array pattern to increase the detection area and even use a cheap disposable party dish platter with a hole cut in the middle to bring the water to the sensor array if uncertain of where it comes from.
The whole thing can be placed on top of the ceiling tile as it will be quite light.
You can connect the beeper to a relay or arduino or whatever to keep track.
Good luck with it.
PS: make sure there is not a water pipe or air conditioning pipe at that location the could drip condentation... as another source of water.
-
Thanks, everyone, for the ideas and suggestions.
I don't reside in the building so in this case I think a passive data-logger would be the best way to proceed.
-
If someone send me an order to develop a "water DROP detector" I would assume "drop" is a keywork, so a microphone or similar device attached to a stretched plastic membrane probably will be my first version.
If WATER is the key then I would start with 2 sheets of plywood separated by a sheet of plastic. Measuring the difference in resistance between top and bottom layer connected into a Wheatstone bridge will give me a device that can sense if something is one of plates wet while keeping it relatively insensitive to natural humidity variations.
I neither WATER nor DROPS are key - meaning that any liquid should be detected then perhaps same 2 layer sensor like described above but with ultrasound speaker/receiver on edges would work? Fluid will change the speed of sound in plywood or any material that absorbs it so it should work. Water will evaporate so this kind of sensor will self-balance itself after time when leaks stop but for oils it will not happen because one of plates will just saturate over time.
You can start blowing air into a pipe with nichrome wire(s)/strip stretched inside (air intake airlow sensor from any car should work). Keep the wire relatively hot. Drops of liquid such as water or others will hit the wire and start evaporating. By sensing change in resistance you can find out if any liquid present and even separate water from oil or diesel fuel for example. It is insensitive to a dust, but you will need to get another stream of air taken just outside of sensing area to measure the difference. if you arrange the pipe in for of Y, you can reverse the flow of air sometimes and auto-recalibrate periodically.
Commercial rain detectors use interruptions in light to count drops. Dust and junk will not be undistinguishable from water drops so probably it is now what you want.
-
If there's any height involved, I like the little microphone way, just have it shielded from getting splashed on, then wire up some level triggering, or program it, and if you're in to it, add bluetooth.
-
I asked this as a general question to not prematurely rule out any ideas, but now I'll give a few more specifics of my situation...
I want to know when water is dripping onto the ceiling tile. I want the data recorded because I don't live in that building and can't rely on the occupants to be around to inform me when an audible alarm goes off.
The water is dripping from a height of between one to two feet and is hitting a specific spot (as suggested by the pattern on the ceiling tile). I know this because above the tile there's a piece of plastic with a small hole in it which is acting like a funnel.
The detector needs to run on batteries and I'd like it to run for at least a month (and hopefully longer) without having to replace them. Visiting the detector means bothering the occupant of the apartment and I'd like to avoid that as much as possible.
Possible sources of the drip include:
- leaky plumbing
- occasional spills of water by the upstairs occupant
- water ingress from the outside due to rain or melting snow
So it could be a long time before any drops are detected.
I'm currently leaning towards using both acoustic and resistive methods for detecting the drops. Ideally the detector can remain in a very low power state and be woken up by drop of water hitting, say, a piezo sensor. It then can use resistive methods to verify that it was in fact a drop of water that hit the sensor.
I'm thinking of placing the piezo sensor under a ramp so the water doesn't hit the piezo directly and the water is shed away from the detector.
Problems to solves include:
- the low power piezo sensor part -- one that can run for weeks on batteries
- the water detector on the ramp -- currently thinking of two interlocking finger traces which run across the ramp (not with the grade but perpendicular to it) which could be used for either a resistance or capacitance measurement
- wireless communications mechanism that has a low stand-by power requirement -- I just want to interrogate it once every week or so to see if it has new data
-
I came across this discussion:
https://forum.arduino.cc/t/drop-counter-rain-gauge/308122 (https://forum.arduino.cc/t/drop-counter-rain-gauge/308122)
The objective here is accurately measure the amount of water collected -- and this would also work for my problem.
In medical and laboratory settings there is a need to measure exact drops of a fluid and the procedure is:
1) make the fluid slowly drip in calibrated size drops, and
2) count the drops usually using an optical sensor (i.e. a LED and a phototransistor).
There is a rain gauge which works on the same principle except that it uses a four wire sensor which counts drops as as they impact the wires. They claim much higher precision over conventional tipping-bucket rain gauges.
"Ultimeter Pro Rain Gauge"
- https://www.peetbros.com/shop/item.aspx?itemid=14 (https://www.peetbros.com/shop/item.aspx?itemid=14)
You can also get just the 4-wire sensor:
- https://www.peetbros.com/shop/item.aspx?itemid=77 (https://www.peetbros.com/shop/item.aspx?itemid=77)
I think the 4-wire sensor assembly includes the drop calibrator -- not sure though.
The optical method is used a lot in the "strobe droplet" / "levitating water" displays.
I also learned from this discussion that you should get a double pulse from an optical sensor when a drop passes through:
https://forum.arduino.cc/t/drop-counter/5083 (https://forum.arduino.cc/t/drop-counter/5083)
(https://farm4.static.flickr.com/3589/3341699447_7047edb530.jpg)