Cheap flow detectors

I don't understand your response.... sorry. The "watts video" link I supplied is for sinks, refrigerators, toilets, dishwashers, washing machines, etc... Basically, anything fed with a twist-on type water supply hose. (So, this probably doesn't include water heaters, but I guess one could be built if it doesn't). The video shows the basic mechanism, which is a pair of valves held apart by an internal spring.

Think of the design like a latch, and it makes more sense how it works. I think they're pretty cool. Not sure what they cost, however.

Only when the flow is fast enough to be turbulent - a Reynolds number in excess of about 2300 in a lot of situations

And sound propagation. And if the flow is turbulent, you can detect vortex shedding behind a bluff body in the channel.

This article doesn't mention vortex shedding flow meters, but they are commercially available.

-- Bill Sloman, Nijmegen

It's called a turbine meter, and you've been able to buy them since the 1970's - probably earlier.

That's a magnetic flow meter, and they've been around since the 1960's at least.

-- Bill Sloman, Nijmegen

is

p,

?SYNTAX ERROR

Put a "balloon" in series with the pipe to "detect" water pressure; a "Y" for flow.

...another way is to use a ball in a vertical pipe,,size of ball a little smaller than pipe ID; increase of flow rate will raise ball more.

On Thu, 15 Apr 2010 16:43:07 -0700, whit3rd wibbled:

Wow - I like that - very elegant.

What about a thermistor in self-heating mode? A pair (one in self- heat mode, one not) should get dowt to a reasonable flow sensitivity.

The shuttle (piston) type runs around $20 apiece one-off. Cheap enough?

Best regards, Spehro Pefhany

I didn't even know Slowman.

Who would even want to?

The feeling is perfectly complementary.

Mike Terrell and krw are perfect examples of the fact that evolution - as an essentially random process - throws up a lot more maladaption than useful variations.

-- Bill Sloman, Nijmegen

Yeah, but that would be an "invasive" approach; "cut the pipes". I was hoping for something that could sit "outside" them. E.g., John's sound and temperature ideas are on the right track. I don't know how to even approach the sound technique as I don't know how I can "filter" out the effects of neighbors, etc. Maybe I could put something *in* the supply line to accomplish that (?) -- but, its a big unknown.

OTOH, temperature might work *if* I do everything relatively; with a slow learning algorithm (so, as soil temperatures change and interior temperatures change, I change my expectations of what constitutes a 'difference' in the pipe)

Hmmm... but, would that be "sensitive enough" (whatever that means)? Or, would it just detect "big flows"?

speed of the flow.

As I said in response to Charlie's suggestion: how sensitive would this be? I.e., I would assume it reacts to the *speed* of the water/fluid (and not the mass)?

Yes, I've used some "fan sensors" like this.

Ouch! I guess a lot would depend on the thermal mass to which the sensor (and heater) is attached? I.e., using this on a 1" galvanized pipe outdoors might prove "expensive"?

How would you sense a *leak*? (presumably, that is a much finer resolution)

Welcome to engineering. You can either work to optimize the galvanized pipe outdoors (insulate well, use a pulsed strategy, budget for a separate heater so you can more reliably detect small changes with the thermistor, or even the two thermistors someone else suggested) or you can change a section of pipe or the location of sensing - it depends on the constraints. If the location must be outdoors and not cutting the pipe is a prime constraint, then you have to work with those constraints.

At some point cutting the pipe (or calling a plumber to cut the pipe) and putting a water meter in-line becomes the more cost effective approach.

Normally you put two contacts near the floor, and when there's a puddle, you have a connection and sound an alarm. They have issues with contamination causing false alarms when things like salt films building up across the sensor body and connecting the contacts without liquid water. Floor paint will reduce problems as compared to bare concrete floors. If you can contain the leak (ie, use a pan under the water heater or clothes washer), a float switch is less prone to false alarms. Won't do a thing for your sprinklers, of course.

Or, find another sensing technology that doesn't have these liabilities. :<

I would, ideally, like something that I can use regardless of location (within reason) and type of pipe. E.g., the thermal approach won't work well on PVC, I suspect (?).

Yup. But, if I can find a noninvasive technique that does what I want, then I'm good for gold! I need way too many of these to be paying much for them *or* the labor to fit them. (I had hoped eliminating the requirement for actually *measuring* the flow might make things easy enough)

Ah, OK. Yes, I am aware of that approach. I was thinking you were discussing some way of measuring a very low flow (exit) from a "sealed/stopped pipe" -- which would essentially give me what I was looking for, indirectly.

Exactly. I'm more interested in the "detect *some* flow" as I can deploy it for different purposes (i.e., detecting flow where there *shouldn't* be any as well as detecting absence of flow when there *should* be)

I was thinking of looking at temperature vs ambient. I.e., water fed from the main would be at ground temperature while water "elsewhere" (in the house, in the irrigation lines, etc.) would be at slightly different temperature -- at

I.e., in the winter time, ground water is considerably colder than water that has been sitting in interior pipes. In the summer, the reverse would tend to be true (or, at least the difference would be *considerably* less... our "cold" water is more like "lukewarm" in summer)

This wouldn't work in the hot water outlet from the hot water heater. :< Nor in *buried* portions of irrigation lines.

So, at lower flow rates, it would take *less* power? Or more?? (hmm... this is counterintuitive :< )

What about *heated* water? Does the signal fall into the noise?

I've used mass flow meters (coriolis) in industrial applications, in the past. Considerable overkill. :>

I was hopping to avoid cutting pipes as that gets messy and expensive really quick.

Crap. This may be one of those "unsolvable" ones.

But don't I have to insert that *into* the line monitored?

It's starting to look that there's no clever solution -- especially one that would scale well *and* be inexpensive! :<