Can I have help finding a Wi-Fi controlled ON/OFF Timer for my water heater's recirculator pump?

Plumbing Forums

Help Support Plumbing Forums:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Jeff, I saw this epic post because "WiFi" in the title and I also want instant hot water, but the electricity free thermosophon is the one for me (and I am an electrical engineer...) You say you want to stick with a pump because it is "awesome" yet it seems like you are describing something rather cranky: It uses electricity, it pumps the water much faster than needed causing turbulence that erodes the copper pipes, the pumps wear out over time, you fuss with scheduling or manual activation with waiting for the water... Wouldn't it be even more awesome if you had the uninsulated return loop to the bottom of the water heater, not deeding a circulation pump, with no bother having to start the pump in advance of using your water?

To confirm for me, Does your setup look like this image View attachment 24036
but with a recirculating pump where this photo shows a check valve? And is your water heater below the level of your sinks like the picture shows? If so, are you able to insulate the hot side and NOT insulate the return pipes so that water cools off and becomes heavier than the hot side? There are a lot of conditions here... if you can't meet them all, then the circulator is probably required.
I took this image from here: Tired of waiting for hot water? Hot water recirculation
Thanks for your family photos! I miss interacting with my local people, so here I am on a plumbing forum!
Jamie
I'm not too inclined to yank the recirculation pump now that I put on a $45 Aqua-Stat and it all works so good. I do realize that having it running causes internal wear on the copper pipe which can cause pinholes (I had 2) but now with the Aqua-Stat it only runs for about 10 minutes every hour & it I wanted to, I could even use a cheap Christmas light timer to limit when it runs.
I'm just wondering, was the pump itself sitting there not running what was keeping my system from thermosiphing? I'd assume that's it. I understand just enough to see how that might slow the water down almost like a check valve would do but just not completely.
 
Uninsulated on the whole loop would cause slightly higher flow, yes, since the heat loss starts earlier in the loop, making water on the return side even cooler, and thus would flow faster. The temperature difference of the water heater's exit water and return loop's temp at the lower entry point of the water heater is what influences the velocity. Bigger temp differential, more flow because the return water is even heavier.

But, faster is not better. Having insulation on the hot side requires more flow needed to keep the water hot enough all the way at the third sink this picture. If the water cooled a lot by the time it reached the third sink, the experience would be "it warm when I open the faucet but gets hotter as I run the water" which is annoying. So, it is best if the hot side is insulated, and the entire return side is NOT insulated so it has a chance to create sufficient temperature differential to overcome the friction losses of every elbow, pipe diameter, and check valve along the loop.

If the entire loop is insulated, there would be less total waste heat for sure, but you might not get any flow because there needs to be enough differential in temperature to force open the check valve and overcome the other friction losses I mentioned. The photo Jeff provided in his first post looks like a circulator pump at the bottom of the water heater. Is it? I wonder if it was first plumbed without a circulator AND insulation on the return line, it didn't work, so a circulator was added?

I added some notes to the internet picture I stole to illustrate what I am saying above:

View attachment 24037
At the end you said "The photo Jeff provided in his first post looks like a circulator pump at the bottom of the water heater. Is it? I wonder if it was first plumbed without a circulator AND insulation on the return line, it didn't work, so a circulator was added?"
This home was built and plumbed in 1993 with everything that's in these photos except for the water heater which I had someone replace.
As far as insulation on the return line, you can see some on the 3' of it that's in the garage feeding the recirculation pump but when they cut into the wall just past that to fix a pinhole leak a few months ago, I didn't see any insulation on that return pipe in the wall.
I don't know if you saw in any of my photos but I bet the guy that installed the 2nd water heater just bent up that long piece of flexible tubing that's between the recirculation pump and the water heater.
Nice work, huh?
 
Last edited:
A general note about the energy cost of the heat loss in these loops:
  • If it is heat loss into a heated home, then it isn't wasted heat because you benefit from it and the boiler has less work to do as it heats your house.
  • If it is heat loss into an air-conditioned home, then the air conditioner has MORE work to do, of course. So this will cost you something. I don't exactly know how to calculate this without researching it more.
  • So, in cooler climates, the heat loss should cost only the difference in energy consumed in heating your home by it's radiators vs hot pipe losing heat to your heated home. If the hot water is an indirect tank or same fuel source, then the cost of this heat source will be possibly the same.
 
I'm not too inclined to yank the recirculation pump now that I put on a $45 Aqua-Stat and it all works so good. I do realize that having it running causes internal wear on the copper pipe which can cause pinholes (I had 2) but now with the Aqua-Stat it only runs for about 10 minutes every hour & it I wanted to, I could even use a cheap Christmas light timer to limit when it runs.
I'm just wondering, was the pump itself sitting there not running what was keeping my system from thermosiphing? I'd assume that's it. I understand just enough to see how that might slow the water down almost like a check valve would do but just not completely.

Adding the Aqua-Stat should solve a lot of problems! You are done, right?

The thermosiphon needs enough temperature difference between hot and return, needs some height in the plumbing from the heater to the sinks, and needs low restriction. So, it needs good design.

A disclaimer: All this I am saying is from a non-plumber who has never installed a thermosiphon loop, but I am skilled in thermodynamics, heat flow, and the design of cooling of electronic systems and LED luminaires. I know heat pretty well, yet imagine there are many topology traps, for example some sinks NOT being above the heater, and spring-type vs swing check valves being used, sticky check valves, feed pipe having more friction loss than the return line AND no check-valve in place (causing hot water to flow backwards from the cold bottom of the water heater), etc. I don't have the benefit of experience.

There are some interesting stories in this guy's blog and the comments that follow: https://www.buellinspections.com/re-circulating-hot-water/ . I see a lot of struggles with heat traps, check valves that require too much force, etc. There clearly needs to be easy water flow, and perhaps 3/4" pipe at least on the feed side. I'll leave a comment in that conversation to see if the host of that article he has anything to add or correct in my points made in the previous post.
 
How many of you commenting here have a thermosiphon system and have lived with it? Or are you just relying on a bunch of various internet articles? I had such a system as I have repeatedly described here for 27 years.

The first expert (jlconline) shows a diagram w/o a check valve at the tank...that’s one of the essential requirements. Now, understand that all check valves are not created equal. There are spring check valves, and these have a "cracking pressure" which is the force required to open the valve. They also have gaskets and seals. These will not work in this application. You just want a swing check valve, and ensure it's mounted properly: that is, horizontal with the flow towards the bottom of the tank. A swing check has no cracking pressure, and generally no wear parts either. Mine was in and functioning for 27 years, no issues.

The "supply pipes" from the water heater to ALL the fixtures in the house were 3/4" copper. Branch lines off the supply lines to the fixtures were 1/2" copper. The return line started at the furthest fixture on the second floor and was solely 1/2" copper.

It’s not “instant” hot water; nothing is. But, it reduces the wait time to seconds.

Regardless of whether you insulate or not, water at the hot water tank exit will always be hotter than water at the end of the return line. With a thermosiphon system as I've repeated, the flow is low and slow. Hot water at your tank tempurature slowing rises to the furthest fixture. It cools off a bit, particularly if there's no use in the supply lines--meaning just the recirculation system flowing. It's still warm as it gets back to the tank, and it takes a very long period of inactivity (meaning no hot water usage) before MY hot water tank turned on again. Remember I had 27 years of experience in my own home (not a customer's home) so I know exactly how it worked.

The return line is just static piping. If you put such a system in and it doesn’t work to your satisfaction you can always add a pump w/o re-piping your system.

My home with this system was in a climate where we heat September-May, so these standby losses were throwing off unmeasurable amounts of heat within the home’s building envelope thus were for all intents and purposes, irrelevant.

This worked so well, and was so simple, that I helped a friend add a kluged system to his house. In his case the plumber did not add it to start. What we discovered was that all the hot and cold water supplies to his second floor were in a wall at the back of a closet on the first floor. At the ceiling point in the closet we opened up a small area of drywall, added a "Tee", and dropped a return line down to the basement. This was not as effective as starting the return line at the furthest fixture, but it made a measurable difference in their wait for hot water upstairs. The work was minimally invasive; just a small drywall patch.
 
Last edited:
Thanks Frodo but could you answer a few questions?
1. Were you aware that this pump has been installed with this same plumbing since 1993 and the pump & it's plumbing is that old too?
2. What makes you say that I'll be replacing the pump within 3 years? I assume something about the elbows being too close to the pump?
3. Can you tell dumb me what having an elbow too near the pump can do to hurt it & what the damage is? I'd assume that the elbow directs water at a higher pressure towards the outer side of the pipe after the bend, causing swirls which work like sandpaper on the tubing?
Thanks for your help. You are right though when you said that I probably won't have it replumbed and that's since it has worked just fine for almost 30 years now with this same pump & plumbing.
Typically the concern for elbows(or other devices) being too close to the suction side of a centrifugal pump is the potential for turbulent flow(vs laminar flow) effects on a centrifugal pump. Stuff like cavitation. However, the distance does vary but I'd say 3 pipe diameter distance would be considered okay. And your arrangement satisfies that.
 

Attachments

  • index.jpg
    index.jpg
    40.4 KB · Views: 1
Last edited:
I wonder why the thermosiphon works at all if you use 3/4" Cu for the hot riser and 1/2" for the cold faller. A back-of-the-envelope calculation tells me that for a height difference of 20', there will be about 4.2lb of water in the hot pipe and 2.0lb in the cold pipe. Maybe I'm missing something, but I don't see how 2lb pulling down can raise 4lb. (I assumed constant temperatures in both sides for starters.)
 

Attachments

  • Thermosiphon.jpg
    Thermosiphon.jpg
    31.8 KB · Views: 9
One thing is you're assuming a specific point where a change takes place between hot and cold. Assigning a 20 ft length for both and determine a weight.
The density is different along the entire length of the piping.
I don't know how it is figured but we know there are working systems with 1/2" return lines.
 
Last edited:
The density is different along the entire length of the piping.
I don't know how it is figured but we know there are working systems with 1/2" return lines.
Yeah, I just assumed an easy case to figure, but clearly the cold-ish return line doesn't get cold at the top. I know it does work, just wondering why. Maybe somebody who stayed awake in calculus and thermodynamics could tell us.
 
It is not the total weight of the water in the tank vs the return pipe because we aren't putting both on a scale to compare, but instead the PSI difference at the bottom port of the tank that matters. The denser cooler water is heavier per unit area, and it will thus be at a higher pressure at the port of the tank. The higher pressure water will enter the tank and displace the lighter, warmer water at the bottom of the tank.

I found a video that shows this:
1586296403880.png
and the it describes why the PSI of each different shaped vessel is the same (given the same density of fluid is in all three)

This video explains it better than my attempt was:

1586296576565.png
The Internet is awesome!
Writing here is more fun that moving plants in the yard!
 
Adding the Aqua-Stat should solve a lot of problems! You are done, right?

The thermosiphon needs enough temperature difference between hot and return, needs some height in the plumbing from the heater to the sinks, and needs low restriction. So, it needs good design.

A disclaimer: All this I am saying is from a non-plumber who has never installed a thermosiphon loop, but I am skilled in thermodynamics, heat flow, and the design of cooling of electronic systems and LED luminaires. I know heat pretty well, yet imagine there are many topology traps, for example some sinks NOT being above the heater, and spring-type vs swing check valves being used, sticky check valves, feed pipe having more friction loss than the return line AND no check-valve in place (causing hot water to flow backwards from the cold bottom of the water heater), etc. I don't have the benefit of experience.

There are some interesting stories in this guy's blog and the comments that follow: https://www.buellinspections.com/re-circulating-hot-water/ . I see a lot of struggles with heat traps, check valves that require too much force, etc. There clearly needs to be easy water flow, and perhaps 3/4" pipe at least on the feed side. I'll leave a comment in that conversation to see if the host of that article he has anything to add or correct in my points made in the previous post.
Not that it matters since I'm sticking with the recirculation pump & Aqua-Stat and it's working really good, but as far as using thermosiphon, from what you said about needing height difference between the heater & sinks, I don't have any. In fact in this 2-story house built on a slab, everything downstairs is almost the same height as the bottom of the water heater, maybe only 1'-2' higher.
So from what I think I've learned here, that's not enough differences in height for it to work. Right?
 
Now I get it; thanks. Pressure differential is pretty small; shows why it takes so long. I'll stick with the pump :).
 
I agree with everything! See some comments inside your quoted stuff.

How many of you commenting here have a thermosiphon system and have lived with it? Or are you just relying on a bunch of various internet articles? I had such a system as I have repeatedly described here for 27 years.

Jamie: Yep, that's me! Your experience is actual data, my words are my analysis from various, sometimes conflicting, Internet articles that mostly don't exactly say how a plumber should do it, or why. I am curious and a future DIY installer for my house, so I want to know what really makes it work. I do try to back up what I say with math or references. I love finding out I am wrong on something, because then I am smarter for it.

The first expert (jlconline) shows a diagram w/o a check valve at the tank...that’s one of the essential requirements. Now, understand that all check valves are not created equal. There are spring check valves, and these have a "cracking pressure" which is the force required to open the valve. They also have gaskets and seals. These will not work in this application. You just want a swing check valve, and ensure it's mounted properly: that is, horizontal with the flow towards the bottom of the tank. A swing check has no cracking pressure, and generally no wear parts either. Mine was in and functioning for 27 years, no issues.

Jamie: I saw an example read here where some installations without check valves kind of work, but not at the last fixture, as it has so much backward flow in the return line it never gets hot enough.

The "supply pipes" from the water heater to ALL the fixtures in the house were 3/4" copper. Branch lines off the supply lines to the fixtures were 1/2" copper. The return line started at the furthest fixture on the second floor and was solely 1/2" copper.

It’s not “instant” hot water; nothing is. But, it reduces the wait time to seconds.

Are you saying the water starts warm and then rises a bit as it runs? I think that means heat loss in the feed side is higher than the velocity of flow and you would benefit if the water either flowed a bit faster or the feed side was insulated better.

Regardless of whether you insulate or not, water at the hot water tank exit will always be hotter than water at the end of the return line. With a thermosiphon system as I've repeated, the flow is low and slow. Hot water at your tank tempurature slowing rises to the furthest fixture. It cools off a bit, particularly if there's no use in the supply lines--meaning just the recirculation system flowing. It's still warm as it gets back to the tank, and it takes a very long period of inactivity (meaning no hot water usage) before MY hot water tank turned on again. Remember I had 27 years of experience in my own home (not a customer's home) so I know exactly how it worked.

The return line is just static piping. If you put such a system in and it doesn’t work to your satisfaction you can always add a pump w/o re-piping your system.
Jamie: And, if a circulator is used, it would be the most energy efficient if it were variable speed circulators with a delta-t control having a temperature sensor on the water tank's output and on the return line perhaps 2 feet from the tank. I think this Taco variable speed Delta-T circulator set to move the water just fast enough to keep the return water within 5 degrees of the feed water would draw the least possible energy and would deliver instant hot water at every fixture in the loop. With that, insulating the whole loop would help in climates where heat loss is not beneficial.

To make your static system flow a bit faster, you could swap out a section of your return line with baseboard radiator (as high above the water heater as possible) so the return water cools down more and at a higher elevation. That return line filled with denser water at a longer height will elevate the PSI at the point it enters the bottom of the heater. That higher the PSI differential at the bottom port of the heater, the faster the flow.


My home with this system was in a climate where we heat September-May, so these standby losses were throwing off unmeasurable amounts of heat within the home’s building envelope thus were for all intents and purposes, irrelevant.

Jamie: Yes! Heat loss is not "heat wasted" if that heat is beneficial.

This worked so well, and was so simple, that I helped a friend add a kluged system to his house. In his case the plumber did not add it to start. What we discovered was that all the hot and cold water supplies to his second floor were in a wall at the back of a closet on the first floor. At the ceiling point in the closet we opened up a small area of drywall, added a "Tee", and dropped a return line down to the basement. This was not as effective as starting the return line at the furthest fixture, but it made a measurable difference in their wait for hot water upstairs. The work was minimally invasive; just a small drywall patch.
 
Not that it matters since I'm sticking with the recirculation pump & Aqua-Stat and it's working really good, but as far as using thermosiphon, from what you said about needing height difference between the heater & sinks, I don't have any. In fact in this 2-story house built on a slab, everything downstairs is almost the same height as the bottom of the water heater, maybe only 1'-2' higher.
So from what I think I've learned here, that's not enough differences in height for it to work. Right?

RIGHT!
 
Not that it matters since I'm sticking with the recirculation pump & Aqua-Stat and it's working really good, but as far as using thermosiphon, from what you said about needing height difference between the heater & sinks, I don't have any. In fact in this 2-story house built on a slab, everything downstairs is almost the same height as the bottom of the water heater, maybe only 1'-2' higher.
So from what I think I've learned here, that's not enough differences in height for it to work. Right?
It would work to circulate the water based on the second floor height and if the first floor was part of the same piping system feeding the second floor, that would circulate as well.

It wouldn't work for the first floor sinks, only if they were on a separate loop or branched off so far from that 2nd floor loop that the branches wouldn't get hot.
 
It would work to circulate the water based on the second floor height and if the first floor was part of the same piping system feeding the second floor, that would circulate as well.
It wouldn't work for the first floor sinks, only if they were on a separate loop or branched off so far from that 2nd floor loop that the branches wouldn't get hot.
As a non-plumber, I'd have guessed that the downstairs loop would recirculate in such a shorter loop than the upstairs would that warm water in the downstairs loop would heat up the AquaStat so soon that it wouldn't yet have gotten nice hot water upstairs before it stopped the recirculation pump.
I was SO jazzed the first time I tried it after I wired it up that since it took 10 minutes to finally heat it up & shut off, that was long enough for good hot water to have reached my bathrooms upstairs.
Sure, it probably hasn't heated the upstairs delivery pipes up fully so they'll cool the water upstairs down a little bit too soon but still, it's hot enough to take a shower. If I need to adjust the faucet a little as I shower, no big deal. 👍
That's way better than before I had the AquaStat & had to waist 5-10 minutes of water before I could shower.
 
Last edited:
As a non-plumber, I'd have guessed that the downstairs loop would recirculate in such a shorter loop than the upstairs would that warm water in the downstairs loop would heat up the AquaStat so soon that it wouldn't yet have gotten nice hot water upstairs before it stopped the recirculation pump.
I was SO jazzed the first time I tried it after I wired it up that since it took 10 minutes to finally heat it up & shut off, that was long enough for good hot water to have reached my bathrooms upstairs.
Sure, it probably hasn't heated the upstairs delivery pipes up fully so they'll cool the water upstairs down a little bit too soon but still, it's hot enough to take a shower. If I need to adjust the faucet a little as I shower, no big deal. 👍
That's way better than before I had the AquaStat & had to waist 5-10 minutes of water before I could shower.
It all depends on how the piping is arranged. You make it sound like it's 2 independent loops.
 
It all depends on how the piping is arranged. You make it sound like it's 2 independent loops.
I don't have a clue how that recirculation system is plumbed. It's hidden in the walls & the builder, Standard Pacific, or their plumbers didn't give me a blueprint or even describe it in anyway when I bought the house new in 1993.
So, my not being a plumber, I can only guess what's inside the walls. 🤔
 
It all depends on how the piping is arranged. You make it sound like it's 2 independent loops.
Now that I think about it more after what you said, I guess that I had assumed that the different areas of the house have their own loops that joined each other somewhere & that would allow hot water that's in the shortest section to get back to the AquaStat the quickest. That's got to be basically how this works, right?
You aren't saying that there's just 1 big recirculation loop that goes to everything in the whole house, are you?
 
NO! Basically just saying that you can't speculate on how the system may function when, like you say you have no clue as to how it's piped.
But you don't have to worry about any of that. You have a functioning system and you are happy about how it's doing.
 

Latest posts

Back
Top