How to calculate head for circulator

Steve Garson

Can someone explain how to calculate the head needed for a circulator. The plumber who did my radiant job used a Taco 10 and it makes too much velocity noise.

Here are the details:

3 loops 156 feet
1 loop of 140 feet
3/4-inch copper feed to manifold 60 feet feed and 60 feet return with seven elbows in each run and flowchecks.

I am not sure of the ID of the tubing: it is flexible black rubber. The OD is around 3/4-inch.

He designed the system without taking into account any furniture, throw rugs or bedding. I just ran the heat calcs and it looks like the system needs 157* water with the rugs or 144* with just the wood floors. I am finding that 145* works just fine.

The bathroom, which is the single 140 foot loop shows a design temp of 99* on the software. They both run off the same manifold, with slightly reduced flow to the bathroom.

With this information, what taco circulator should I be using?

Steve Garson

Calculating head and GPM needed for circulator

Can someone explain how to calculate the head needed for a circulator. The plumber who did my radiant job used a Taco 10 and it makes too much velocity noise.

Here are the details:

3 loops 156 feet
1 loop of 140 feet
3/4-inch copper feed to manifold 60 feet feed and 60 feet return with seven elbows in each run and flowchecks.

I am not sure of the ID of the tubing: it is flexible black rubber. The OD is around 3/4-inch.

He designed the system without taking into account any furniture, throw rugs or bedding. I just ran the heat calcs and it looks like the system needs 157* water with the rugs or 144* with just the wood floors. I am finding that 145* works just fine.

The bathroom, which is the single 140 foot loop shows a design temp of 99* on the software. They both run off the same manifold, with slightly reduced flow to the bathroom.

With this information, what taco circulator should I be using?

Jamie_5

calculating head

Siegenthaler uses a resistance model to calculate head loss. Tubing is assigned a resistance per foot and fittings are assigned an equivalence to so many feet of the same tubing. Then you map the system and calculate, using the appropriate fluid density at the appropriate temperature. See his book, Modern Hydronic Heating, for the full method.

That said, I ran the numbers you gave in Siegenthaler's Hydronic Design Studio software. I assumed that your fluid was water at 145 degrees F and your tubing was 1/2 PEX. Using a Taco 010 pump, the highest flow (all circuits open) was 3.36 GPM, which is well under the accepted maximum of 4 GPM for 3/4" copper. So I'm willling to bet it is not velocity noise you're hearing, but something else, perhaps a vibration being transmitted to floor or wall components.

Hope this helps.

Steve Garson

Jamie:

Thanks for the calculation on the GPM. But what amount of head do I need? I am trying to determine the correct Taco circulator to replace the Model 10 which generates a lot of velocity noise.

Thanks again,

Steve

Tom Meyer

Circulator sizing

Steve:

You're missing a lot of information in order to decide how the circulator is going to react.

You would need to know exactly the resistance per lineal foot of the tubing (which varies by diameter and type), the total resistance of all fittings and components the circulator "sees". You need to also understand how the piping is set up. (If it is a primary-secondary, for example.) Also, if this is a single pump system, and a single zone system, things happen one way. If there are manifold valves, then things happen another way under certain circumstances. Also, are there any other circulators in the system? Do they act in parallel or series?

I've attached the technical information for a Taco 0010.

To give you an example of what might happen, if the actual head seen by the circulator is 4 feet, looking at the pump curve, you go up the Head (vertical) axis until you get to 4 feet, then go to the right until you hit the pump curve for the 0010. Follow that directly down to the Flow (horizontal) axis. That tells you what the pump WILL pump at under those conditions. If you don't want 27 gpm, you need to add resistance using something like a circuit setter.

Also, when you look at the circulator performance, you have to anticipate different conditions. (Closed zones, etc.)

There could be a big difference between the flow required and the actual flow. Excessive flow creates excessive velocity which creates not only the sounds you may be hearing, but could cause errosion of the piping fittings, lifting of check valves and other not so pleasant things.

Tom Meyer
Senior Designer/Trainer
Precision Hydronics Corporation

To Learn More About This Professional, Click Here to Visit Their Ad in "Find A Professional"

Michal

not sure but sounds oversized

to be honest, I have never done a radiant floor loop, but a taco 10 sounds way overkill, Looking at the cut sheet in the catalog, that pump puts out 26 gpm at 5 feet head, so yes, the noise you might be hearing might just as well be the high velocities and flow. But like i said, I have never done a radiant floor system, but taco 10 seems too big in gpm. Go back to who ever designed the system.

Steve Garson

I'll get the information needed. This is the only zone fed by the circulator.

Chuckles_3

An "experimental" way to do it is to put in a Grundfos 15-58. This is a three-speed pump where even the top speed doesn't move as much water as a Taco 0010. It doesn't cost any more than a Taco 0010. Don't use the highest speed that causes no noise. Rather, find the lowest speed that will heat the zone.

Joe Mattiello

Taco 0010performance curve

What is the load? The load will dictate the flow rate and the required flow through the ID of the piping used will dictate the head. Flow is additive, but head isn’t, so all that’s needed is the length of the longest loop 156 lineal feet. Assuming you have a 20,000 btu load, at 2 gpm, through 156 lineal ft x 1.5 for the valves, and elbow will be 234 ft through ¾” pex tubing at .02142 frictional loss per ft = 5 ft. Your design is the collective flow amount @ 5 feet. For your reference I have attached the submittal data sheet for the 0010, which has the performance field of curves for the 00 series to successfully size the right pump for your system design.

Taco, Inc.
Joe Mattiello
Technical Service Technician
joemat@taco-hvac.com
401-942-8000 X 484
www.taco-hvac.com

Steve Garson

I checks the specs on the radiant tubing: It is Watts Radiant 3/8-inch Onix Heatway EPDM Tubing. Capacity is 6.25 gallons/1000 feet. So the four loops hold 3.8 gallons of water, plus the feed and returns. Thermal conductivity of 0.17 BTU/hr-ft Degrees F.

Joe Mattiello

easy fix

Incidently, if you have velocity issues you can throttle back on the flow, reducing the flow with no detriment to the pump, and at no cost to you. A lot cheaper then changing out the pump. Try it…. what do you have to lose.

Taco, Inc.
Joe Mattiello
Technical Service Technician
joemat@taco-hvac.com
401-942-8000 X 484
www.taco-hvac.com

Steve Garson

Velocity Noise / GPM Flow / Head

I need some guidance. With the Taco 10, there is a ball valve at the output, so I can throttle down the flow, but when I close it as much as 80*, where 90* is closed, I still hear the noise. It only stops when the valve is closed. How can I correlate this to an alternate circulator?

hr

If the circ is oversized

closing down a ball valve at the discharge just moves the velocity noise to another spot in the system

Anytime you restrict flow and cause it to go through a smaller opening you get a velocity increase, and often the noise associated with it.

IF in fact you have noise caused by an oversized circ, your best bet is to install the correct circ. The 3 speed Grundfos, mentioned above, would be a wise choice. A quieter pump to begin with. IMO, and the speed adjustability will fine tune your flows.

hot rod



To Learn More About This Professional, Click Here to Visit Their Ad in "Find A Professional"

Jamie_5

calculating head

I am not a professional like the other posters, just a homeowner, so I may be missing something important. That said, if my assumptions, which I list below, are correct, I believe the total head of your system will be between about 11-1/2 and 12 feet, depending on which zones are open.

Assumptions: there is one pump which pumps through 120 feet of 3/4" type M copper piping containing 14 90* elbows and two flow check valves, which is the supply and return piping from the single radiant manifold with zone valves of some sort which feed four zones of 3/8" PEX-AL-PEX tubing (I believe Onix is actually different, but I assume the resistance to flow of the two tubing types is similar), three loops of which are 156 feet long and one of which is 140 feet long. I assume no pressure differential bypass valve is installed and the fluid is assumed to be water at 145* F.

Using a Taco 0010 pump on the setup described above, flow will range from a low of .62 GPM with just one of the longer loops open to 2.1 GPM with all four zones open and each zone running at just over .5 GPM. So I don't think you are suffering from velocity noise.

The people posting here who suspect that your pump is wrongly sized, however, may well be right. Looking at the pump curves posted by one person, it looks like a number of pumps could produce flow rates close to that of the 0010 at the same head. What you need to know now is what the designed flows for each of the zones is supposed to be, which will depend on the amount of heat you are trying to carry to each zone and the difference in supply and return temperatures you are shooting for. If the flow per zone was designed to be at about .5 GPM, the pump curve chart that was posted suggests that even a Taco 007 would work, and that appears to be confirmed by Sigenthaler's software. If you need slighly higher flow, it looks like an 008 or 009 would be a better choice, since the head of your system falls closer to middle of their pump curves. The 0010 and 007 are at the low flow end of their curves at 10 to 12 feet of head.

Somebody else mentioned that if the piping arrangement were primary-secondary, the results would be different. I don't understand why, but I am probably missing something.

NB: The above results depend on the accuracy and completeness of my assumptions. I interpreted what you said in your posts to arrive at those assumptions, but I could have misunderstood or missed something. For instance, if there is some component in your circuit which has high resistance to flow and which is not included in my model, like a Munchkin boiler or some types of heat exchangers, the model's results are meaningless. GIGO, and all that.

Hope this helps.