Here's a little project I've been messing around with that I'd love to have some input and/or suggestions on as I've been somewhat perplexed about the best way to handle it...

Here's a quick synopsis. Water is flow is turned on or off via a solenoid valve. There are two possible "signals" that could turn the solenoid on. On the output side of the solenoid is what essentially amounts to a "sprinkler" system. (It's actually "foggers")

The problem is, occasionally debris will get stuck in the solenoid and prevent it from closing all the way, resulting in the foggers staying on, though generally in a diminished capacity.

-------------------------

My idea is to somehow monitor this setup and determine if the water is still flowing when it shouldn't be...

-------------------------

My first idea was pretty basic...

I found a "flow switch" on a surplus site... It could detect down to .1LPM of flow. Whenever water was flowing through it would close a contact.

I monitored the two signals that could call for flow and OR'd them together. I then passed that to an XOR along with the input from the flow switch. The output of the XOR was buffered slightly with an RC circuit (to allow a small delay for solenoid opening/closing time) and then used to trigger an alarm...

The result was that any time the solenoid was on, but water was not flowing, or any time the solenoid was off and water was flowing, it would trigger an alarm output...

The following are the problems with that version of the project...
1) The switches were expensive even on surplus (about $90 per if I recall) and availibility seemed very limited.
2) The switches themselves could possibly fall victim to the same issue. Since it worked on water pressure being able to push open a "valve", theroretically debris could get stuck in the switch and cause a false alarm. While that's better than the other way around, it's still not great...

-----------------------------------

Another thought I had but have not attempted to test...

If a filter was installed before the solenoid, to keep debris out, then I could monitor the pressure differential between the two sides of the filter while the solenoid was on. If the differential exceeded a certain value I could trigger an alarm signalling that the filter needed to be cleaned.

The biggest problem with this that I see is that it potentially adds more maintenance to the system... I'm sure there is a lot of debris that goes through unnoticed, whereas the filter will now catch it all and probably clog frequently.

----------------------------------

I've also considered using a single pressure sensor on the inside of the solenoid. I think I'd have to do some testing though to see what the normal pressure is with water off (i'm sure the foggers plus height of the water columns getting up to the foggers add a bit of pressure) and see if the pressure generated by a partially stuck open solenoid is greater... If so, then this could be a fairly simple solution... If the solenoid wasn't energized and pressure was higher than the normal "off" position, then I could trigger the alarm...

----------------------------------

Last but not least I've also considered trying to find some sort of rotary with a magnet and use a hall effect sensor... Only issue I have with that is that thus far I haven't had much luck finding many distributors (i.e. mouser, etc) that stock such things, and when I have found them they tend to be very expensive... Not to mention an exact flow rate like this would give isn't really what I'm going for here, so it's really a bit of overkill...

----------------------------------

Anyone have any advice / suggestions? Basically I need to make sure it's relatively inexpensive, parts are easy to get ahold of, etc...

Thanks in advance,
Greg

 

How about temperature, with or without an upstream heater? In fact, the heater and temperature monitor can be one and the same device (e.g., resistance of a heated wire).

John

 

Here's an idea, based on the clear dome things they use to use on old gas pumps when you fill up your car.

A little glass tube/container, inlet at the bottom. Coloured glass/plastic bead in the bottom and a IR sensor (reflective type) shining at the bead. If there is any flow the bead will "dance" causing a repeating signal at the IR sensor. Provided the bead is light enough and the inlet underneath it is an orifice (acts like a little "jet" nozzle) the bead will dance even with very low flowrates.

And buy a filter anyway man!

 

Here's an idea, based on the clear dome things they use to use on old gas pumps when you fill up your car.

A little glass tube/container, inlet at the bottom. Coloured glass/plastic bead in the bottom and a IR sensor (reflective type) shining at the bead. If there is any flow the bead will "dance" causing a repeating signal at the IR sensor. Provided the bead is light enough and the inlet underneath it is an orifice (acts like a little "jet" nozzle) the bead will dance even with very low flowrates.

And buy a filter anyway man!

Hrmm.... that's a good thought... actually, a little neo magnet in the bead along with a hall effect sensor would probably make it even more reliable...

Thanks much!!!

 

Why not fix the root cause of the problem and filter the water before it reaches the solenoid valve?

 

I agree...

Filter = good.

But even then there can always be a ruptured solenoid diaphragm or leaky pipe etc. A flow sensor will indicate ANY fault condition where water is lost (after the flow sensor).

 

Or you could sense water coming out of one of the sprinklers.

 

Have you looked at turbine style flow sensors/transducers?

These give pulses at a rate proportional to flow and there are many types from quite cheap to stupidly expensive.

A couple of examples of moderately priced parts:
http://uk.rs-online.com/web/search/searchBrowseAction.html?method=getProduct&R=0257149

http://uk.rs-online.com/web/search/searchBrowseAction.html?method=getProduct&R=5114772