Smart Hot Water Recirculation Solution

I did mine just before Christmas. My old tank was a 14 yr old electric and one of the elements kept burning out. Also I have plans for my basement, and have been working towards moving all of the mechanicals out of there for a couple of years now, and the hot water heater was the final piece.

It’s temperate enough where I’m at that I could do an external tankless mounted on the back of the house, so that was what I chose. The water heater was $899 online, plus I paid a buddy about $1800 to help me install it, including replumbing the whole house for both water and gas. We tore out all the old piping that hung below the basement ceiling- both galvanized water and black iron gas pipe- and replaced with pex and CSST routed in the joist bays and crawl space. Now I can put a new ceiling in the basement without pipes hanging beneath it.

We were probably not supposed to do the gas work ourselves but what the city doesn’t know won’t hurt them. I read all the codes and I’m confidant everything is compliant. I’ll probably go ahead and get the job permitted and inspected this summer. I didn’t do it at the time because I didn’t want to get red tagged over something and have no heat for christmas. I pressure tested the gas lines, so I know there are no leaks, and everything is bonded to ground as per code.

Heh. I guess I’m still kind of proud of this job so I want to tell everybody about it. It still makes me happy every time I take a shower!

@NoWon @wgmcg @rossetyler

I understand now. Thanks for clarifying. In my case I’m going to run a return line, particularly now that I understand about the implications of the crossover valve and the end of the run.

what DH are you using for the qbino? Ah… nevermind.

Here’s the pinout for the qbino:


This coupled with the brewpi pinout for a standard pots jack:
Black - unused
Red - 5v
Green - GND
Yellow - Data

qbino black -> pots green
yellow -> yellow
red -> red

Seems to be working great!

1 Like

thanks I hadn’t sat down and figured out the qbino ts pinout yet still waiting on my brewpi sensor to arrive.

I actually just used this device handler found at The smartest house website where I bought the qbino. But that one is likely the same.

interesting. That’s a different DH entirely. A few months newer than what I’m using. The erocm123 DH above reports in fahrenheit, which I thought I remembered you saying your qbino didn’t do.

Maycock seems to be using the ST “composite device” DH model, which he explains at this link:

Dunno which one is better. I just realized however that I can use this device as both a temp probe and a switch. I have a little 2Gal electric hotwater heater as a buffer in my hotwater line that’s right at the location of the temp sensor. I’m using a ge-zwave switch to control it right now but i think I can do both jobs with the single qbino device and use the GE switch elsewhere. So, the qbino has paid for itself already!

yes mine is reporting in Celsius I played with erocm123 DH but I am running my setup on my Hubitat and I was not able to convert his DH to work I think due to his using child devices but I have not spent much time with it.

Is the hubitat ST DH cross compatible? Is it faster to respond that ST?

you have to make a minor name change mod in the code to most DH to make them work with Hubitat.
it does not like DH that uses child devices but they are working on fixing that.
As all custom code is run locally it is faster.
In Hubitat the only things that use the cloud is stuff like Alexa, Ring doorbells, Sonos or networking to the SmartThings hub
Hubitat when first pairing a device does not see devices as fast as SmartThings (it might take an extra 30 sec - minute) but they do pair.
SmartThings is more polished but with running all the custom code and my presence sensors on the cloud vs locally made me add the Hubitat.
SmartThings was also limited to 32 zigbee devices and I was already past my limit where as Hubitat does not appear to be limited to 32 zigbee devices.

Is there a rules engine for it? How do you program it? I’m pretty deep into webcore is there anything like that?

it has an app called rule machine which was on SmartThings a year or 2 ago but the developer pulled it and is part of the Hubitat group now.
Rule Machine is more basic than webcore not as many options but basically the same
Others have converted Webcore for Hubitat but I have not tired it i think there are still a few bugs in it.
I am just moving everything I can run locally over like my locks, garage doors, alarms, lights and presence sensors and then run cloud based devices like my Sonos, doorbells and nest off SmartThings.
One thing Hubitat still does not have is a remote app you have to be connected (wifi or hard wired) locally to the same router the Hubitat hub is connected to. So you have to view/control it from home, although you should be able to remote in to your home computer to view/control the hub and it’s devices if you home computer is on. I just have not gotten around to setting that up.
You can see most devices in the Smartthings app when you network Hubitat with SmartThings they show up as virtual devices in SmartThings. That requies a community created app
Hubitat is working on an official remote android/ios app and a direct connection app to SmartThings but neither has been released yet.

Just bought one. Thanks for turning me on to it.

Funny last major ST outage a few months ago I railed about the cloud failure leaving my circulator pump on (I caught it in 20m or so, but still) and called for a community supported local processing engine with a port of webcore.

Seems these guys were already way ahead of me.

off topic but to save you a little time
to make SmartThings DH work in Hubitat

  • copy the DH (ctrl A and copy) into notepad
  • select edit and replace physicalgraph with hubitat and replace all
  • copy the modded DH into the Hubitat interface
    (it uses web browser but goes right to the hub just like if you logged in your router)
  • not all DH will work it may take a little trail and error or a search on the forum. Support will usually get you a DH in a few days.

This next bit might be over kill and really applies to SmartThings as well (most people will not bother doing it)
Like SmartThings, Hubitat when you fist set it up scans your zigbee channels and chooses the channel with the least noise on it. This is normally fine. But I did have a GE switch that would not pair to the channel it choose. but paired to another channel fine. But then you have remove and pair all your previous zigbee devices to that channel (you could try leaving them for 24 hours and they may find the hub by themselves depends on the device).
others say zigbee devices will find the new channel automatically but I have found not all devices have that feature.

  • turn off any 2.4 wifi your router has that you do not need (for this test) after the test if you need the 2.4 WiFi change it’s channel away from the zigbee range.
    zigbee and 2.4 uses the same band and 5g is faster anyways.
  • power off SmartThings and remove the batteries (just for this test)
  • Then go through each channel one at a time and try to pair a zigbee device 3-4 times. If it does not pair after
    3-4 attempts move to the next channel. If it does pair you will have to remove it or use another device for the next channel.
  • After you have completed all channels from 11-26 you will notice only a handful paired in the first few attempts.
  • go back and repeat the test on those few channels and narrow it down to one.
  • If it is the same channel as SmartThings is using (found in the hub info on ide) then you will have to contact SmartThings support and have them change the SmartThings zigbee channel so it does not cause any interference.
    (SmartThings does not allow us to change the channel ourselves)
    Some people have found the lower channels to work better.

Over kill but may save you a lot of headaches down the road.
This should probably be done for SmartThings as well if we had control of changing the zigbee cahnnel.

I have not had any issues with zwave devices.

Thanks for the info. My hubitat should be arriving this weekend. I plan on integrating it very very slowly, given that my ST install is working pretty well at this point.

In other news, I finally went to install my brew pi thermostat, only to discover that those stainless barbs don’t fit 1/2" pex. I know you had trouble ordering them, and frankly you’re better off for it. They fit 1/2" ID hose, which is not the ID of either 1/2" or 3/4" pex. I thought they looked a bit funny but didn’t give it much thought beyond that. I’m glad I didn’t do this project late last night. I’m off to the big box to buy some 1/2 barb to 3/4 male npt fittings.

This is a great thread. I have been looking for a better pump and noticed that some people were using the BACOENG pump. This does not seem like a good option is it is not designed for drinking water and is not NSF rated as far as I can tell.

I did find another pump which is made by Taco and sold by Rheem( Homedepot) and Richmond (Menards) with part number RH17920 or RP17920. This pump is designed to run for 2 minutes every 30 minutes which makes me think it has much higher flow compared to the Watts pump. It could easily be wired for on demand. It is designed for tankless systems

This pump would need to be used with some sort of valve between the hot and cold lines if no dedicated return line is available.

Also noticed that Taco has a hot to cold water control valve. It is a different design than the Grundfos valve. The Taco valve is part of the Taco HotLink system. Seems like a beefier design.

1 Like

It looks like there is a stainless and a cast version. They seem to recommend the stainless for this sort of use.

It looks like the Rheem/Richmond pump is based on a Taco 0015. The Taco 0015 seems to have slightly better specifications although it is much more expensive compared to the BACOENG.

Taco 0015 - 18 MaxGPM/17 FT Max Head 17 FT
BACOENG is 13 MaxGPM/20 FT Max Head

So, I’ve installed the Watts pump and have it on an Iris Smart Plug that also measures energy usage. From what I’ve gathered, the pump uses a little less than 27 Watts when on. This is true if it is pumping while the valve is open or not. I used the below information to help me with my tendency to throw the “Law of Diminishing Returns” right out of the window when it comes to “SmartThings”.

27 Watts X 24 Hours = 648 Watt Hours or .648 kWh (Daily Use)
.648 kWh X 365 Days = 236.52 kWh (Yearly Use)
236.52 kWh X $0.12 (Energy Cost per kWh) = $28.38 (Yearly Cost)

Yep, that’s less than $30 a year if the pump was always on. Now, something else to consider is recirculating water when there is no demand! The cost to reheat that water, although is not calculated, has to be more than $30 a year…

So, how important is it to know if the water is hot or not versus if the pump should be running or not? From my understanding, it takes about 1-2 minutes for the water to recirculate through the pipes, of course longer or shorter times depending on how much linear foot of piping you have. If there is a motion event triggered in the bathroom, then more than likely, someone is going to be in there at least 1 or 2 minutes before they wash there hand or get into the shower…

I also considered just letting the pump run 20 minutes each hour while the house is in Home Mode. (Simple Button)

So, I can already cut 6 hours off of the run time of the pump just by having the pump off during Night Mode. That leaves 18 Hours of just 1/3 of that time on to equal 6 Hours per day.

Pump being on 6 Hours per day = Less Than $8/Year to operate! Spending $60 on a temperature sensor will take 7 Years to recoup the cost. By that time, the pump and the valve will probably went bad. :rofl:

So…As I hit the order button on Amazon, I just can’t help myself even though I know that I am just wasting money.

Side-Note, I hope my Wife doesn’t see this post!!!

Is there any reason why a regular contact sensor with builtin temperature couldn’t be taped or stuck to the raw copper pipe coming out of the wall? Even if it doesn’t read the actual temperature of the water, wouldn’t it serve as a benchmark as to how how the pipe is “relatively speaking”?

Yes, doing this basically upgrades it to an under-sink system with a smart timer ( It’s not much more expensive that the basic Watts. You save time and headaches with a system that does what you want out of the box.

1 Like

Yep, I can see this as being a more streamlined solution. Would still need a Smart Plug/Switch to make it “Smart” but at least it’s “Smart” enough to know when to turn on motor and when not to. Also having the added benefit of one less part to fail.

I have the Watts installed so it is what it is. I’ve created the Piston below to see if I would need to install a separate temp sensor. In essence, it would limit the amount the pump runs by at least 1/3 of the daily operations. This should lead to less energy both in electric and gas…Maybe…

We will see.