Water flow detection but not for leaks

[Handywill]

I'm just launching on HA, and have a specific problem. I have a hot water circulation pump to minimize my hot water wait time at the taps because my house is old and large, and the bathrooms are scattered. The pump currently runs on a timer which I can transfer to my HA controller software, but I'd like to do something different: detect flow from the supply line to the tank (indicating hot water demand somewhere in the system) to send a signal to my HA controller to trigger the pump to run long enough to charge the circuit with hot water and then shut off. This would reduce my energy usage by both limiting the pump hours and the parasitic hot water losses from the circulation (it's insulated but far from perfectly). This way someone could turn a tap on for a few seconds, turn it off, wait for the pump lag time (30seconds or so) and not have to trigger an event with a switch, remote, etc.. That is, of course, my other option, but I think this idea is a bit more elegant if it can be done cost effectively. My controller computer, by the way, is in the boiler room, as is my network, from a standpoint of considering signalling. I happen to have decided on using zwave with Indigo, as I have an old Mac laptop.

The problem is that the options I found in the forums (sprinkler flow detection, for example) aren't sensitive enough (4gpm seems typical), plus the supply pipe is only 3/4 copper, which is pretty small for this type of solution.

In a perfect world, I wouldn't have to solder in a new fitting, but expect I will have to.

Any ideas?

 

[BraveSirRobbin]

Just did a quick search and found this one.
 
http://www.aliexpress.com/item/G1-4-DIN6-direct-plug-Hall-effect-Water-Flow-Sensor-0-15-1-5L-min-0/537547120.html

 

[RAL]

I agree with BSR, that the Hall flow sensors are the way to go for detecting water flow.   A project I worked on needed a flow sensor and we experimented with several different types of sensors.  The Hall effect sensors worked best.
 
But I must be missing something in your plan.  If you have to turn the water on for a moment to allow the sensor to detect that there is a demand and then turn on the circulation pump, and then wait 30 seconds or so for it to charge the pipe,  why not just turn the water on and leave it on until the hot water arrives and forget the pump completely?  It seems like either way you have to wait about the same amount of time.   What am I missing?

 

[Handywill]

Thanks for the quick response! RAL, the only reason is just to waste less water, as you are correct, there is no particular time benefit. I am also going to create a scene, so if you have a remote handy, you don't even have to use a tap trigger at all, it just seemed elegant to do it this way, and the cost of the sensors you pointed me to is very reasonable. Next I have to figure out wiring...

 

[wuench]

Instead of a flow sensor, I wonder if you could get a motion detector faucet and tap into that for pump control.  It would operate the same way, but might be a simpler install plumbing-wise.

 

[BraveSirRobbin]

Actually, thinking about this even more, I would also add a temp sensor on the copper water line (insulated from room).
 
Here is the reasoning.
 
Say someone only turned on the hot water for a second.  You would recirculate for a dedicated amount of time based on this detection (not needed most likely).  Base this instead on flow AND temperature (circulate only if temperature decreased to xx degrees).  Of course, I guess you could always trigger 'based' on the time flow was present.
 
Also, say no one used any hot water (absence of flow) for a long time.  You would want to circulate possibly?  Only if someone was home?
 
Of course, you could also base your logic on house occupancy (with arm status perhaps)?

 

[RAL]

.. the only reason is just to waste less water, as you are correct, there is no particular time benefit.

 
Of course ... makes perfect sense.  Living in the northeast, where water is plentiful, we aren't in the habit of having to be careful not to waste the small amount of water used by running the faucet for 30 seconds.  So that reason didn't occur to me at first.   14 billion gallons of water in the Hudson river flows by our town of 7500 people each day.    

 

[fcwilt]

But isn't the primary function of such a system to provide "instant" hot water? Yes I understand the savings aspect but that seems a very secondary benefit.

 

[az1324]

Turning a faucet on and then off is more elegant than pushing a button?

 

[roussell]

Instead of trying to detect flow, I think I'd trigger off of the pressure drop in the line when a hot water faucet was turned on. Just put a pressure switch on the line after the hot water tank that is sensitive enough to trip when the pressure drops or us an analog sensor and measure the actual pressure. Either way it will probably be easier/cheaper than an flow sensor.

Terry

 

[signal15]

How much is this going to save you?  When I was on city water, it was costing me 0.2 cents per gallon.  It's going to take you an awful long time to recoup the cost of what you're putting in, not to mention the time you put into it.  Is it worth it?
 
If I was to waste 25 gallons a day, that's 5 cents.  It would take me 6 years to recoup $100.  And 25 gallons a day is a very high estimate, considering that our family averages about 75 gallons a day according to the meter on my softener.  I'd guess it's more around 10 gallons a day wasted, which is a 15 year payoff on $100.  That $100 would be better spent putting in more efficient bulbs, setting up some lights to be on occupancy sensors, etc.

 

[RAL]

Instead of trying to detect flow, I think I'd trigger off of the pressure drop in the line when a hot water faucet was turned on. Just put a pressure switch on the line after the hot water tank that is sensitive enough to trip when the pressure drops or us an analog sensor and measure the actual pressure. Either way it will probably be easier/cheaper than an flow sensor.

Terry

 
I think using a pressure sensor will be problematic.  The pressure on the output of the tank is created from the pressure on the cold water line feeding into the tank. So when someone turns on the cold water, the pressure will drop on the hot water line, too.   Maybe there would be a small, detectable difference between the pressure drop of cold vs hot water being turned on, and depending on the sensor, you might be able to distinguish the difference if the pressure in the system is otherwise constant.  But if you have a well and a pressure tank, it is common for pressure to fluctuate by 20 psi or more.  That will cause all sorts of false triggers by the cold water if you try and use a pressure sensor.
 
A flow sensor will be much more reliable indicator.

 

[ddennerline]

I am installing a similar hot recirculation line and will use Z-Wave plug, high efficiency pump, occupancy scheduling, and aquastat sensor. I was thinking about putting a button on bathroom/kitchen walls, but instead I will probably just use the Z-Wave bathroom light as additional input. ElkRP rules will drive the whole system.
 
Where I live, water is expensive because the $2bil sewer bondholders require a nice ROI.

 

[Handywill]

Thanks for all the ideas.

Signal15, must admit that the idea of saving water is certainly not enough to justify it financially. It is more the principle of the thing. And because it's cool. Payback on my system certainly won't be fast (if ever).

AZ1324, it isn't only a bit more elegant (no need to have a button that for the guest says "push me for hot water"), but not everyone in my family has a remote or cell phone, or wants to carry one, and adding switches in all the bathrooms and the kitchen would be more than I have planned (initially, at least).

Frederick, while that is the idea of these types of systems in most installations, in my house it ain't instant, it just shortens the hot water interval from well over a minute to 15 seconds or so. The circuit is not what you would do on a new house, it is well short of the tap, but rather an oversized (length) manifold, in effect. Too much destruction to do the full job, but better than nothing. Not a waste of energy in the winter (I have an indirect heater fired by a high efficiency boiler) but it is in the summer.

Sounds like the Hall would work. Need the Hall (about $15), 2 solder to BSP unions, power supply, zwave unit. Can do the plumbing myself. So not TOO crazy.

 

[kurtmccaslin]

You will not get enough flow through that 1/4 inch hall effect sensor so it must be in parallel with your pump with appropriate check valves.
 
I am planning to install a Metlund D'mand system in the house I am building.   It is a circulating pump with integrated temperature sensor.    You push a button (or motion detector) in the bathroom to circulate the water.   When the hot water gets to the end of the line, the temp sensor cuts the circulation off.    This reduces energy consumption-- less hot water down the drain.   Also only circulates the water when needed.     As you pointed out, this also reduces water consumption.
 
I am building a net zero energy house that runs off of rainwater.   Therefore energy and water consumption are important.