next topic »

[on:]

I want to sense water level in a river, just inside a water mill actually.

I don't need to be very accurate +/- 5mm would be just fine +/- 10mm at a push.

To avoid debris and potentially constrain a float, to improve stability if necessary, I am considering mounting the sensor in a 110mm plastic pipe.

I was wondering if anyone had tried this. I don't know much about ultrasonics, even less about this particular sensor chip.
I expect that enclosing the module in a smooth walled container will exclude any extraneous signals but I also think I may get echo, like you might in a cave.

Any thoughts, better yet any experience?

[Reply #1 on:]

...the ranging accuracy can reach to 3mm.

You may want to play with the sensor in the bathtub.  Even under ideal conditions you may not get 5mm accuracy trying to measure water level.

[Reply #2 on:]

Ultrasonics can be a tough nut to crack sometimes.
I have found you need a nice clean water surface, no froth or muck, as it will not reflect the signal back.
Also, I would tend to avoid placing the unit in a pipe of any sorts as the signal will bounce of the inner walls and you will most likely not get the reflection you are wanting.
That also means keeping the sender unit away from walls. or other structures.
Hey, I thought you people in the UK used imperial measurements, glad to see you speak metric, has there been a change?

How about using an optical sensor targeting a floating disk or object on the water?
Some of the industrial optical sensors I have used are quite good, albeit expensive. They are typically adjustable for sensitivity and range.

Is the water still or is it flowing?
What is the total rise and fall you expect?

You mention you are not wanting much in the way of accuracy, I would say that +/- 5mm is pretty tight for this application.
Consider also not only resolution, but also repeatability.

rockwallaby . . .

[Reply #3 on:]

I have found you need a nice clean water surface, no froth or muck, as it will not reflect the signal back.

You could put a hard, loose-fitting float in the tube to avoid this, maybe.
You'd probably have to put a filter on the tube to avoid debris entering and jamming the float.
Limescale could also be a problem.

Maybe a U-bend to isolate the sensor?

[Reply #4 on:]

Yes you could, but again, avoid the tube like the plague, it will give you problems with reflections.
With the idea of a floating plate, you need to make sure it will always be horizontal so it will reflect back to the sensor and not somewhere else.
Clean still water, itself will always be horizontal by nature.

rockwallaby . .

[Reply #5 on:]

I only need to read a change of about 1m but I will need to do it from a at least 1m above that or the thing will drown when the river is in flood.

The water is flowing, disturbed and far too dirty, floating debris of all sorts and sizes, to offer a flat surface.
That is why I was asking about the tube ...
A float seems like a good plan, polystyrene perhaps?


However if ultrasonics don't work in a tube then I probably wont be using ultrasonics

I will try it anyway I think whilst expecting to be disappointed, I may learn something from the attempt that points to another solution.
Thanks for the input
Al

[Reply #6 on:]

Maybe a laser range finder like this http://diydrones.com/profiles/blogs/11-gram-arduino-powered-laser-rangefinder

[Reply #7 on:]

Very interesting, thanks.

I have also been wondering about capacitive sensing, although that would require me to build a sensor which is basically a tube with foil on either side that is insulated from the water but arranged so that water will fill the space between the plates as it rises.

More fiddling I think

[Reply #8 on:]

The water is flowing, disturbed and far too dirty, floating debris of all sorts and sizes, to offer a flat surface.
That is why I was asking about the tube ...
A float seems like a good plan, polystyrene perhaps?

If the dirty floaty stuff is small and moving it should work.
How big is the biggest debris, twigs will be fine, logs will be a problem, probable also for other reasons too 

Anything small that floats by will not really cause a problem, but all this depends on how good the ultrasonic receiver is.
The electronics in most units usually do quite a bit a processing and you end up with a smoothed output signal.
Oops, I have just seen that the HC-SR04 is a simple device and has no real post processing 

How about putting a trash rack just ahead of where you want to read, so all the debris is shunted around that area if you think the debris will be a problem.

I would only use a floating disk of sorts if you were to use an optical measuring device, not for really or ultrasonics as the float.
I don't know the specs of the HC-SR04, though I did use such a thing many years ago for a different application not involving water.
The units I have used are industrial types and are fully IP rated and send a 4-20mA signal out.
If you use this unit you may need to think about protecting it against moisture that will have not so good an effect on the sensor over time.

Now here is an idea that is also used, how about putting a pressure sensor at the bottom of your weir or tank.
You can either drop the sensor in or if you can mount it in from the side wall so the cable is not in the water.
Again, I have done this for industrial installations.
With the correct sensor you can get very good constant readings.

rockwallaby . .

[Reply #9 on:]

You are absolutely correct RE industrial options, we use them all over the place, at their industrial price of course 

I was just exploring options really, call it proof of concept, playing whatever ...
I like to know what will and will not work, it helps me to conceptualise when I am doing real work.

Thanks
Al

[Reply #10 on:]

Further to the idea of using a pressure sensor I just came across this article which you might find interesting.
Read through the comments.

http://kayno.net/2010/02/11/wireless-picaxe-based-water-tank-level-sensor/

Paul

[Reply #11 on:]

Maybe you could use a float with quadrature gradations (or turn a rotary encoder) and count pulses.