Recirc system add a loop

Plumbing Forums

Help Support Plumbing Forums:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

hardwite

Member
Joined
Mar 9, 2019
Messages
21
Reaction score
0
Location
South Carolina
A recirc system with a timer that works well when the timer is on. But when off, the furthest appliance needs ~5 minutes to receive hot water. It is a long loop. I would like to split into two smaller loops to reduce the time. A shorter time would also allow me to turn the pumps on from the bathroom right before use (on the as-is loop, the user will have to wait ~5 minutes before turning on the shower or hot water faucet).

I am not interested in smart systems that try to time usage.
I do not want to increase the on-time - defeats the purpose and doesn't address use when timer is off.
The loop system is as shown below. It loops hot water from the tank and back to the tank. Here is the loop as-is:


Simplistically, the new loop would look like this:

My questions are: do you advise having two pumps on one tank (pulling from one tank outlet and pushing into one tank inlet)?
What other solutions for this issue?
How do I size the new pump? I have a rough idea of the pex sizes and lengths. I believe 3/4 pex for the loop, and 1/2 for the taps to the appliances. See these pictures of the loop pex and an example tap.


I have the pump information but do not know how to calculate its exisitng head pressure and flow rate, and do not know how to do the same for a two loop system.

Pics of the water heater and connections


Thanks!
Ted
 
Well looking up the specs on this circulator, I see it's capable of 3 speeds and the literature, which I assume you have(?), includes performance curves. I included a copy below.
Before I go any further I'd like to say that I think you should look at what you have now for flow vs. pressure loss(referred to as Head in feet on curve). 5 minutes seems exceptionally long. My first thought when reading your info was why not just increase the flow with a higher capacity circulator.
I have no idea how you select speeds and didn't take the time to look for it. Perhaps you already know.
Before I forget, another thing I noticed that these circulators have a built in Check Valve.
I should mention that I am not a plumber and lack a lot of hands on or detail knowledge of this stuff. But I was involved with engineering and calculating systems many years back. So I can help you figure out your system head loss(based on pipe sizes, lengths and flow.) and then we can compare that to these curves and what we theoretically should be pumping at those pressure loss (head loss). Then hopefully we can see why it's taking 5 minutes for you hot water.
It would take a fairly long distance of smaller piping to cause that pump at minimum flows, but even a flow of about 2 gpm wouldn't take longer than a couple of minutes. We'll see.
So if you're up to it, we need an estimate of total lengths of the loop, types and number of fittings and size(s). I have pressure loss values for PEX tubing and maybe for the fittings depending on the type of connections used.
Grundfos UPS15-55SFC.jpg
 
Looking back quickly, I see you have crimp fittings. Since those have a smaller ID than the PEX tubing we may have to ballpark the equivalent head loss unless I can find a company that has published it.
 
The plumbing was installed during house construction by a contractor, so the workmanship is all his.
I downloaded the documentation: it is a 3 speed with check valve (though the plumber installed a second). The motor is on speed 2. Pictures below.
I spent time under the house and traced the loop. Rough estimate is it is ~295 feet round trip, with ~50 elbows or Ts (clamp/crimp fittings), and all 3/4 inch PEX. Turns can be made with bend supports, but I did not see any of these, only crimp elbows. See picture below. Why the plumber did this - elbows+labor cheaper than bend supports? More opportunity for failures, and defeating a key benefit of PEX.
I don't see any more information on the fittings, but they are crimp type. I am also not sure if this is PEX B - some pics of the markings (some of the 1/2 scap pieces have no markings).
The last room is ~265 feet from the HW supply. A ballpark estimate for the first of the water to arrive is ~5 minutes (I can barely feel a change in temperature at that time, and then around 5:20 a noticeable change. I also measure the water at 129F (in a cup with the faucet submerged), which is a 17 degree drop from my tank setting of 146F.


Grundfos extracts

A better pic of the markings on the 3/4 line. Scraps of the 1/2 had no markings.

Example of routing. It seems sub-optimal, unless support drives this.
 
295 feet! WOW!
I don't think that pump was designed for that type of application.
I don't have the time to work out a system curve(that's what your piping pressure losses are at various flows.) to match it to the pump curve right now but the following will give you an idea where we're headed.
Call it 300 feet of 3/4" PEX = 5.52 gallons of water.
Head loss for a metal 3/4" 90 bend (could only find one mfr that showed it) is equivalent to 9.4 feet of piping. Therefore 9.4'x50 bends = 470' (That's equivalent feet of 3/4" PEX tubing.)
That's 470' + 300' = the equivalent of 770' of piping.
Just as an example, 4 gpm of water flow though 770' of 3/4" PEX = 3.53 psi pressure loss per 100'. Or 3.35 x 7.70 = 25.8 psi loss.
The pump show it's capabilities in head pressure. 28.5 psi x 2.3 ft of water/psi = 65. 55' of head.:confused: as you see the pump is not capable of that. In fact it shows 18' at 0 gpm (referred to as shut off head.)
This may appear a bit confusing but when I show the "system" curve plotted over the pump curve, you'll see that where they intersex is what the actual flow would be.;)

EDIT: So if it takes about 5 minutes to move say one half of the contents of that pipe(2.75 gallons) (assuming the point of use is half way from the source) that would be roughly a 1/2 gpm flow rate. 1/2 gpm x 5.5 min = 2.75 gallons.
Good night!
 
Last edited:
Well I'm back.
Based on the input info I used to figure the theoretical flow rate of the existing system, it appears to have come on the high side. About 1.6 gpm @ 13.6 feet of head.
Of course the friction loss values used may have been low for the fittings(?), etc. but I was hoping to see a closer number.
Then we have the branches off the recirc loop but that shouldn't add much to the time since it's actually adding to the rate of flow.

Have you ever considered incorporating an aquastat to sense and control pump on/off based on return temperature of circulated water. This could be used in conjunction with a timer used for general time frames when you want the recirc system to operate.

Here's what the system curve looks like (in red). The actual flow rate of a pump is based on its head pressure capabilities vs the piping systems resistance to flow and is reflected by where the two curves intersect each other.
CIRC DWG UPS 15-55 SFC CURVES.jpg As you may know PEX has a smaller inside diameter than it's equivalent nominal size in copper, but it's the distance that's a killer.
You mention temperature drop of 17 degree, which is substantial. I wonder how long the payback would be for energy savings if pipe insulation was used?
Way back when I was working industrial design, heat tracing with insulation, was actually mentioned as a way maintaining delivered hot water temperatures to the point of use. Never hear about anymore. I assume it's not very economical.
 
Last edited:
circ_loop.png zne_valve_3.png

A 3 way zone valve , if wired like the 2 --3 way switches in your kitchen,
will allow you to turn on each loop as you need them
with a wall switch in your desired area.
the zone valve will tell the pump to turn on, as it opens or closes the port to zone A or zone B
this will require a 24 v transformer

It can also, be on the end of the circ line keeping the loop constant circulating
then when called for port B opens and the flow goes to both zones
This cuts down on wait time,keeps the loop hot

4x3_teee.png
 
I understand your #7 post where you use a zone valve to select one of two loops for different set of appliances.
However, I have no clue as to what your thinking is in post #8, with the a/c position selection recirculating a smaller loop with no water available to the appliances.
Or the second one where you add remote control isolation valves for each appliance. Give us a hint as to philosophy of operation. Or purpose?
 
#8
the op said his fixtures are a long way, he emphasized the word LONG. that he had a 5 minute wait for hot water
what #8 does is make the loop shorter/ closer to the fixtures
recircs water 24/7 no wait for hw

#9 deleted...to expensive to install
 
Last edited:
@hardwite
I thought this might interest you. It's a heat loss calculator that I used to come out with your 17 degrees temp loss. I didn't know what the ambient temperate was so I used 60 degrees. I found I had to reduce the flow rate down to about 1.4 gpm to come up with 17 degree drop.
https://www.wbdg.org/guides-specifi...s/temperature-drop-calculator-hydronic-piping

I couldn't determine what was being used as pipe material, as I'm sure it's a bit different than PEX.
But playing around with it you can see, for example, that 1" of fiberglass insulation reduces the temp drop substantially.(To about 4.2 degrees)
Temperature Drop.jpg
 
Last edited:
I installed a Watts (essentially the same as the Grundfos, I'm told) circulator to drive two loops ( each about 30' around) in 3/4" copper. Insulating the pipes using big-box foam sections made an enormous difference in perceived performance with very little temperature drop following pump shutoff. If I had the OP's 295' loop problem, I would replumb with Type A PEX, with as few fittings as possible, insulate the heck out of it, and run it continously.
 
I installed a Watts (essentially the same as the Grundfos, I'm told) circulator to drive two loops ( each about 30' around) in 3/4" copper. Insulating the pipes using big-box foam sections made an enormous difference in perceived performance with very little temperature drop following pump shutoff. If I had the OP's 295' loop problem, I would replumb with Type A PEX, with as few fittings as possible, insulate the heck out of it, and run it continously.
Yeah...but that's a lot of work. I think he could get by with a circulator with a higher Head/Capacity.
Those fittings sure do add a lot of head loss. Most companies don't even list head loss value for them. But I would suspect all the metal ftgs would be close to the same and all the plastic ftgs would be close to each other, as well.(Plastic have almost twice the pressure drop as the metal.)
Insulation could be considered but at higher velocities the heat loss goes down.
The other thing I would recommend is an aquastat to have the circulator run just enough to get the return temperature to desired value.
1. Bigger pump
2. Aquastat control(in addition to a timer to shut off when not typically required.)
3. Insulation to reduce temp drop and subsequent pump on/off times.

Edit: Both Watts and Grundfos make circulators in a wide variety of flow and head capacities.
 
Last edited:
Here's a rough table of PEX fitting loss, given in equivalent lengths:

View attachment 20186

His estimate of ~50 elbows in 3/4" PEX looks like ~500+ feet equivalent. I was surprised to see that Type A and B (Brass crimp) were so close.

Good general design guide at http://www.texasinspector.com/files/pex_designguide.pdf.
Thanks for that link and info. I just spent an hour or so going back and forth to different sources. Nothing seems to agree exactly but reasonably close.
But that table that you posted, showing Uponor system with higher values for the 3/4" over the metal insert fittings, doesn't make sense. But on the other hand,it doesn't say what thoses numbers represent. Where did you get that chart?
:D
 
Last edited:
Thanks for the great comments. Some corrections to my information:
  • Outside temperature was ~68F when I measured the 17 degree drop. + a few feet of 1/2 PEX from the recirc loop to the faucet.
  • The 50 fittings are 90 degree elbows and T-offs for the 1/2 PEX run to the appliances. ~34 elbows and 16 Ts. When analyzing the circ-loop, the T's would be a straight-line flow of water not a turn; therefore less resistance?
  • All of the fittings are metal.
  • The pipe is PEX, type B or C. Can anyone tell from the markings?
I don't understand most of the calculations, other than the long wait time (w/pump off) and temperature drop is due to the excessive length and resistance - even with error, all of this fits. If the pump is not suitable, does this mean I risk an early pump failure?

These can be resolved with using two loops or a higher flow rate pump. My concern with the higher flow pump is stress on the fittings and leaks down the road. How probable is that risk?

Re-plumbing into two loops can be via 1 pump per loop, or one pump with a zone valve for 2 loops. Same cost and complexity between these two options? -add a pump or add a valve?

Is another option to have a balancing valve, or does that again overload the pump?

Note: whichever choice, I will use fewer elbows and insulate.

Is this the one-line for a 2-loop with one pump and a 3-way valve?
 
Last edited:
Let me go back for a moment regarding the statement, "when off, the furthest appliance needs ~5 minutes to receive hot water." I lost track of that statement. In other words, it takes 5 min. when the tap is open and running. Based on the approx. 265 feet of 3/4" tubing that would calculate out to an approx. flow rate of only about 1 gpm.(4.75 gallons/5 min.)

Is your primary concern to reduce the time it takes the recirc loop to provide hot water to the furthest point? Say when you hit a switch?
And/Or not to have the 77 Watt recirc pump running constantly while a timer dictates?

With regard to concern for higher flow rates, let me remind you that what you're doing by reducing the friction losses (with shorter loops, etc.) is in fact increasing the flows. That's partly how the time gets decreased. Except of course when a loop/line is added shortening the distance.
Also with higher flow rate for recirc system, we're only talking rates of flow that it would have been if it was a much shorter run of piping and/or what normal use flow rates would be.

I do like your proposed layout with the added line directly to the furthest fixtures. Not sure how you house/bldg is laid out but I would assume that would reduce the length of run to those furthest fixtures as well as possibly reduce some losses on the first loop. That alone should result in a higher flow rate.

EDIT: You could always make the proposed piping improvements and leave the pump as is with the idea that you could increase size after the fact, if need be.
 
Last edited:
Sorry for the confusion - it takes 5 minutes for the pump to send the water to the furthest appliance. I was confusing my goal, which is to leave the pump off (for efficiency) except
  1. during typical use times (e.g. morning) and
  2. when a user indicates they need hot water (via a button, motion, or other sensor to run pump).
A 5 minute travel time makes #2 not convenient or useful. A 2 minute travel time would work.

After your calculations and the above comments, I now worry that this poor design will lead to early leaks or a failed pump. Similarly, if that a larger pump would put the pressure or flow out of 'good design practices'.

The plumber laid out the loop as follows:
  • It runs in the crawl space (accessible).
  • It runs from the water heater to bathroom#1, then to the kitchen, then turns and goes upstairs to bathroom#2, then back down to the crawl space to bathroom#3, then up to another second floor bathroom, then back down, etc. etc.
  • It runs from the water heater in this way, to the furthest 'geographically distant' bathroom, then back towards the water heater picking up other bathrooms. This results in the 'furthest appliance on the loop' being only 5 feet away from the water heater. And, he put the bathroom#1 shower on the outbound loop, and the sink on the inbound loop, so the shower gets water quickly, but you die of old age waiting at the sink (the plumber used to be a writer for the Twilight Zone TV show in his previous life).
Because my kids will be out of the house and the upstairs unused, I was thinking that a good loop split would be one loop for upstairs and one for downstairs. This should cut down the distance nicely.
 
Yes, my first thought was the 5 min was for the recirc system but when I reread you original comments I had to ask, to be sure.

What you describe sounds quite reasonable. Particularly since it would appear to reduce friction loss substantially. I don't think there will be a concern for too great of an operating pressure being required, to carry a reasonable flow, with a reduction in pressure losses. We'll take a look at the new loops.

Is that upstairs loop represented on your diagram as those 2 points on the dashed line?

What's a your estimate of the resulting length and fitting count of each of those 2 loops?

Yes in my first look, I had purposely just used the ells, based on the purpose of the estimate.

Are your branch runs all 1/2" and what's the longest(approximately only).

upload_2019-3-14_13-16-50.png
 
Last edited:

Latest posts

Back
Top