Thermosiphon vs. Recirculating Pump
My HP water heater is being placed immediately below the master bath and an adjoining bath in my new home. However, the kitchen. laundry and powder room at on the other end of the house, so I was considering an on demand recirculating pump for that end. In my google search for dedicated return loops I came across the idea of a thermosiphon loop as described here: http://www.buellinspections.com/re-circulating-hot-water/. My plumbing setup (water heater in basement and fixtures a floor or more above) seems perfect for this.
I think I could set this up with a timed valve, so it did not “run” while sleeping (thus not wasting energy to reheat the water).
Here are pros and cons of thermosiphoning as I see them so far:
Pros:
– no pump to buy ($500)
– no pump to maintain
– no need to ever replace the pump
– no cost of running a pump (some sources say it’s about the same as the cost of heating the water for a standard WH. A HPWH should be a lot less, of course).
Cons:
– when the valve is open, hot water will be siphoned from the tank into the loop, which will increase the heating the water due to reheating.
– when the valve is open it will be using electricity, though not much,
– timer cost and possible replacement cost
Anyone have any experience with these systems?
Norman
CZ 3A
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Norman,
A thermosyphon loop is an energy nose bleed -- constantly drawing hot water from your tank, even when no one is using hot water.
For more information, see Hot water circulation loops.
I guess the question is: Is a periodic nosebleed (running 4-16 hours a day) more costly than an on demand pump (running maybe <2 unit?
Note that in Winter, all the additional energy used goes to help heat the house (in your case, at resistance electric rates).
There is a cost analysis below, although note that insulating the pipe would greatly decrease the heat loss (the vast majority of the cost). And that different assumptions can make a big difference (eg, pump running 12 hours/day, my NG/elec rates, 50' pipe, no insulation and not accounting for Winter benefit/Summer loss, I get $80/year). Maybe $30/year with insulation.
google:
site:www.treehugger.com /clean-technology/ask-pablo-will-a-water-saving-hot-water-recirculation-pump-really-save-me-money.html
Norman,
I don't know why you assume that a thermosyphon will operate for 4 to 16 hours per day. My assumption is that it will run for 24 hours per day.
Martin,
The article I linked had a valve with a timer that turned off the loop at night. I thought I could create a more precise schedule based on my real use which might get it down to as little as 4 hours.
Norman,
You're correct that such a thermosyphon works, at a lower cost than a pump, with some additional energy penalty. Of course, the energy penalty may seem reasonable to you.
I'm not a fan of solenoid valves, because many brands have a reputation for failing regularly. (A failed solenoid valve is a maintenance headache.) Note also that if you install a solenoid valve to turn off the thermosyphon for 8 to 20 hours per day, that means that no hot water will be available at the tap for 8 to 20 hours per day -- so if you change your daily schedule, all you have is cold water.
My last house had a long run between the hot water heater and the various use points on the far side of the house. To resolve this issue, I had a plumber install a recirculating line with a Grundfos pump. The pump had a permanent magnet motor and was very quiet and durable. Grundfos now makes a "Comfort" pump that has a auto adapt function that controls the operation based on use patterns. It might be work considering.
When I redid my current house I considered doing something, as there are about 40 feet of pipe between the hot water tank and the master shower.
It turns out it doesn't really bug me, and since I have my own well, I do not stress out over the social cost of heavily processed water going down the drain. Also, all the other fixtures are on the way to the master bath, so if I am not first to use hot water, it isn't much of a wait.
I cannot see wasting heating energy constantly recirculating hot water, but if it is a problem, how about this idea:
Set up a recirculating pump, only it is switched by the shower light. If you add a DOM [delay on make] relay into the circuit with the pump wired through the NC contacts, you could have a system that runs for say 20 seconds when you switch on the shower light. IOW,when you switch, the pump turns on, after 'x' seconds[the relays are adjustable] the DOM relay kicks, and since the pump is wired through the normally closed contacts, it turns the pump off, not on. When you turn the light off, the relay turns off and the process is reset.
No wasted water, no wasted heating fuel[other than what is always wasted by water cooling in the pipes.
If the relay dies, the pump will run whenever the light is on and the relays are cheap.
Keith,
A thermosensor that turns the pump off when hot water shows up would work similarly.
But would it not run all the time whether there was demand or not?
The control on my recirculating pump loop with electric tank less heater works this way
Anytime there is a draw of hot water that will turn on the water heater. My control senses the heater current. If the current loop temperature is below the set point of 110° the pump will run until the loop reaches the set point. The pump will time out after 3 minutes and stop regardless of loop temp.
It takes a little less than 2 kWh and 3 minutes to heat my 175 foot loop.
So I splash on the hot water for a second or two and then off and from any faucet then get undressed and put the laundry in the hamper by then the shower is hot.
The down side is if you turn on hot water and change your mind the loop will still heat and I wasted 22 cents of electricity to no benefit.
Walta
So am I correct in thinking that whenever the water heater runs, and the loop is cold, the pump runs?
Yes if the loop is cold and there is a call for hot water the control will heat the loop. Since the heater is tank less it only runs when there is a call for water or the pump is moving water.
I wanted a button switchs to start the control but the wife was not having any part of that.
Walta
Sounds like a good trade off between energy use, ease of use and time to get hot water - without the false triggers of a motion detector.
"I wanted a button switchs to start the control but the wife was not having any part of that."
That was my first thought and came to the same conclusion, SO's would not tolerate/use a switch, but being it one's own house, remembering to turn the light on seems a small price
https://www.mcmaster.com/#7268k21/=1asmp7b
55 bucks plus the wire is not an unrealistic investment
> It takes a little less than 2 kWh and 3 minutes to heat my 175 foot loop.
Is this a large pipe? I get more like .6 kWh for 3/4".
Jon
When ran the numbers on paper the answer was much like yours. On paper the pump should cycle the loop in 45 seconds.
But in the real world it take 180 seconds for the warm water to make the trip and when I read the power meter it seems to go up by 2 kWh.
The plumber used more elbows than I would have preferred and I think the elbows are mixing the hot and cold water in the pipes, on paper we assume they are separated and one displaces the other. On paper I did not estimate how many BTU to warm up the tubing.
Walta
Walter: Let's analyze this:
"But in the real world it take 180 seconds for the warm water to make the trip and when I read the power meter it seems to go up by 2 kWh."
180 seconds is 0.05 hours. An energy rate of 2kwh in only 0.05 hours is 2kwh/0.05= 40kw, or 40,000 watts. At 240VAC that's 167 Amperes Within error that implies a ~35kw tankless(?) on a 240V/150A circuit?
2kwh= 6824 BTU. If the water started out at a cool 60F basement temp and was raised to 110F, that's a 50F rise, which implies there is 6824BTU/50F= 136 lbs of water + water-equivalent thermal mass in the loop. At 8.34lbs/gallon, that comes out to 16 gallons.
The volume of 3/4" PEX is 1.84 gallons per 100', and you have 1.75x that much, or 3.2 gallons.
The volume of 3/4" L type copper has 2.5 gallons/100' which would come to 4.4 gallons in 175'.
The thermal mass of the pipe is negligible.
Unless you have 1.25" - 1.5" plumbing on that loop, the numbers aren't anywhere near what you think they are, but I'd believe 1/4 of that on the first draw, when the water temperature really is that cool.
I didn't follow the role of elbows in causing mixing. This is a single hot water line running to a point where it is connected to some type of return line? A lot of energy is going somewhere....
Yes Dana you are correct in calculating my heaters wattage.
It is in fact a 36 kWh heater running on 4 240 volt 40 amp circuits.
The 2kWh reading is from the power company’s meter. Being a 400 amp service the meter does not display increments less than 1 kWh. So the display changed by 2 kWh when heating the loop so the actual usage is more than 1.1 and less than 2.9 with some small background loads running.
I am ashamed to say the heater was set to 128° so that gives us a 68° delta.
So 6824/68=100 pounds of water
100/8.34=12 gallons
The 175 foot length number was a guess I was surprised to find the real number is 230 feet for the loop length of PEX B seems to hold 1.84 gallons per 100 feet. 2.3x1.84=4.2
Now I am thinking electrical load is more like 1.1 kWh for the first run in the morning.
1.1 kWh=3753 BTU
3752 BTU/68°F=55 lbs of water
55 lbs/8.34= 6.6 gallons
I am thinking the hot and cold water is mixing in the pipe so I need to circulate 6.6 gallons in the 4.2 gallon loop to flush out the cold water.
My original post said less than 2kWh
Walta
Jon
My loop has a dedicated return line so it is more like this system.
How the water in the pipe mixes depends on the velocity of flow.
This is a great article on hot water systems.
http://www.garykleinassociates.com/PDFs/15%20-%20Efficient%20Hot-Water%20Piping-JLC.pdf
Interesting point, thanks. So you are moving water at about 2.2 gpm and if you had a bigger pump, you could move 4 gpm and achieve a ~35% more efficient "plug flow". And get close to a 1 min wait time.
I was not my intent to hijack this thread.
I am on my second pump the first was rated for 48 liters per minute at 95 watts. Its performance was unacceptably slow.
https://www.amazon.com/gp/product/B00O1N69VQ/ref=oh_aui_detailpage_o05_s00?ie=UTF8&psc=1
So I replaced it with 60 liters per minute at 260 watt
https://www.amazon.com/gp/product/B01NCUHO7K/ref=oh_aui_detailpage_o04_s00?ie=UTF8&psc=1
Clearly my loop is not flowing at anywhere near the speed the pumps spec sheet claim my loop only has about 6 feet of elevation change I suspect the flow restriction is PEX B fitting that fit inside the tubing and have a smaller inside diameter. The other possible restriction is the ¾ inch brass check valve.
Walta
Elevation change doesn't matter, but the pipe itself is a big factor. My guess would be two Taco 009 pumps in series.
I have low water flow. I was tired of waiting for hot water so installed a thermosiphon system. I cut 3" wide strips 1 1/2" XPS then routed a groove with a 7/8" core box bit the length of each strip for the 3/4" copper supply line and 5/8" for the 1/2" return. I clam shelled the entire loop with two strips glued and sealed with PL300 foam adhesive held in place with zip ties. I used closed cell foam pipe insulation for the feeds to each appliance. This is part of a gut remodel so I was able to optimize pipe runs and detailing. With airtight R9 insulation I'm confidant that the system is economical and responsible. It works great.
Thrifttrust
How much change in elevation do you have in your system?
What is the temperature differential from supply to return?
Does your heater have a tank?
I like your insulation idea.
Walta
Thanks Walter. My heater is a tank type. (I think this is necessary.) The shower valve (highest point) is 66" above the tank outlet. The heater is a 68K BTU/hr high performance condensing gas unit. It has side inlet and outlets designed for space heating. The return goes to the lower port, 36" below the top outlet. The entire loop is about 60'. The water leaves the tank at 140F and returns at 116F. I don't know the resting flow rate but I have never heard the heater run unless there had been hot water demand. I did not put any valves in the loop for fear a malfunctioning unit could allow water to stagnate.