Pump sizing help

[fin5281]

I just completed a 40 x 40 x 16 pole barn and have installed half-inch pex in the floor on six parallel loops with the manifold connected to a siesco SH 22 tank less water heater. The longest loop is 272 ft and the other 5 are close in length. I am trying to size the correct circulator pump.


Sent from my iPad using Tapatalk

 

[Dagny]

Taco book is in my truck but guess a 007 would work you have very little head and 007 is available in a zoning model.

 

[fin5281]

Taco book is in my truck but guess a 007 would work you have very little head and 007 is available in a zoning model.

I must be doing the math wrong. Can you help me out with it


Sent from my iPad using Tapatalk

 

[Boilerhouse]

Here is a pump selector guide from Taco. It may be helpful.

https://www.taco-hvac.com/uploads/FileLibrary/SelectingCirculators.pdf

 

[fin5281]

Here is a pump selector guide from Taco. It may be helpful.

https://www.taco-hvac.com/uploads/FileLibrary/SelectingCirculators.pdf

I just used the same document and it calculated out my head pressure at something crazy.

It's a 72000 btu system so the gpm comes out to 8.3

Using their formula for head and 1/2 pex in a 272 ft loop is like 32.3

That seems really high fora simple single zone system


Sent from my iPad using Tapatalk

 

[Siteman25]

Based on what you provided, for your flow rate, the pipe is too small, that is why the head is coming out very large. For water systems we try to keep water velocity under 7 ft per second, per the Taco chart the hydronic heat range is 2 to 4 feet per second. For 1/2" pipe the flow rate for that range of fluid velocity is 1.25 gpm to 2.5 gpm. To use a flow rate of 8.3 you need at least a 1 inch pipe, for a velocity of 3.4 feet per second. For your loop length you will need to be on the low end of the flow rate to get one of these Taco pumps to work properly due to the head loss in the pipe. You are probably looking at an 009, 0011 or 0013. In order to help you with those calcs we would need to also know what fluid you are using, the average fluid temperature, and what your expected temperature drop is.

 

[fin5281]

Based on what you provided, for your flow rate, the pipe is too small, that is why the head is coming out very large. For water systems we try to keep water velocity under 7 ft per second, per the Taco chart the hydronic heat range is 2 to 4 feet per second. For 1/2" pipe the flow rate for that range of fluid velocity is 1.25 gpm to 2.5 gpm. To use a flow rate of 8.3 you need at least a 1 inch pipe, for a velocity of 3.4 feet per second. For your loop length you will need to be on the low end of the flow rate to get one of these Taco pumps to work properly due to the head loss in the pipe. You are probably looking at an 009, 0011 or 0013. In order to help you with those calcs we would need to also know what fluid you are using, the average fluid temperature, and what your expected temperature drop is.

Planning on using 50/50 mixture of antifreeze at 100degrees. Shooting for a temperature drop of 20 degrees. I already tried a bell and gossett 100 that I had laying around.... needless to say it did not work. I was only getting .4 gpm. I have the manifold, seisco sh 22, expansion tank , spirovent, and 1" valves plumbed in with 1" copper. Manifold spans off to 6 parallel loops of 1/2" oxygen barrier pex.


Sent from my iPhone using Tapatalk

 

[Siteman25]

Single zone?

 

[fin5281]

Single zone?

Yes single zone. It's a 40x40x16 polebarn


Sent from my iPhone using Tapatalk

 

[Siteman25]

With your 1/2" Pex, 50/50 mix, and loop at 272 feet (use about 281 feet, to account for 4 90 degree bends and a branch tee), calc is as follows:
H=0.0374*1.582*281*4.970 = 82.6 feet of head (using flow rate in 1/2" pex of 2.5 gpm, velocity slightly higher than 4 feet per second, none of the Taco pumps can do this.

H=0.0374*1.582*281*3.364 = 55.9 feet (flow rate of 2 gpm)

H=0.0374*1.582*281*2.033 = 33.8 feet (flow rate of 1.5 gpm, slightly higher than 2 feet per second velocity)

The Taco 009 circulator can do this, but when you back up and look at your system, the change in temperature is:
f=Q/(450*T)
1.5=72000/(450*T)
T=106.7 degrees

So realistically what would happen is
1.5=Q/(450*20)
Q=13,500 BTU delivered

1.5=Q/(450*30)
Q=20,250 BTU delivered

In other words due to your tube size, you can't circulate enough water to deliver all the heat you need. The flow rate should be around 8 gpm with a pex tube size of 1 1/4" for a 20 degree drop. Based on those numbers:

f=Q/(450*T)
8=72000/(450*T)
T=20 degrees

1.25 inch Pex
H=0.000794*1.582*293*38.055 = 14 feet head

The 0012 model matches that pretty good

1 inch Pex
H=0.00223*1.582*288*38.055 = 38 feet head - No Taco pump works, head too high

if you did a 30 degree drop

f=Q/(450*T)
f=72000/(450*30)
f=5.33 gpm, use 5.5 gpm

1 inch Pex
H=0.00223*1.582*288*19.753 = 20 feet head, the 009 and the 0014 match up pretty well.

 

[fin5281]

With your 1/2" Pex, 50/50 mix, and loop at 272 feet (use about 281 feet, to account for 4 90 degree bends and a branch tee), calc is as follows:
H=0.0374*1.582*281*4.970 = 82.6 feet of head (using flow rate in 1/2" pex of 2.5 gpm, velocity slightly higher than 4 feet per second, none of the Taco pumps can do this.

H=0.0374*1.582*281*3.364 = 55.9 feet (flow rate of 2 gpm)

H=0.0374*1.582*281*2.033 = 33.8 feet (flow rate of 1.5 gpm, slightly higher than 2 feet per second velocity)

The Taco 009 circulator can do this, but when you back up and look at your system, the change in temperature is:
f=Q/(450*T)
1.5=72000/(450*T)
T=106.7 degrees

So realistically what would happen is
1.5=Q/(450*20)
Q=13,500 BTU delivered

1.5=Q/(450*30)
Q=20,250 BTU delivered

In other words due to your tube size, you can't circulate enough water to deliver all the heat you need. The flow rate should be around 8 gpm with a pex tube size of 1 1/4" for a 20 degree drop. Based on those numbers:

f=Q/(450*T)
8=72000/(450*T)
T=20 degrees

1.25 inch Pex
H=0.000794*1.582*293*38.055 = 14 feet head

The 0012 model matches that pretty good

1 inch Pex
H=0.00223*1.582*288*38.055 = 38 feet head - No Taco pump works, head too high

if you did a 30 degree drop

f=Q/(450*T)
f=72000/(450*30)
f=5.33 gpm, use 5.5 gpm

1 inch Pex
H=0.00223*1.582*288*19.753 = 20 feet head, the 009 and the 0014 match up pretty well.

This situation seems very close to what I am trying to do. What makes mine different ?

http://www.pexuniverse.com/content/how-size-circulator-pump


Sent from my iPad using Tapatalk

 

[fin5281]

With your 1/2" Pex, 50/50 mix, and loop at 272 feet (use about 281 feet, to account for 4 90 degree bends and a branch tee), calc is as follows:
H=0.0374*1.582*281*4.970 = 82.6 feet of head (using flow rate in 1/2" pex of 2.5 gpm, velocity slightly higher than 4 feet per second, none of the Taco pumps can do this.

H=0.0374*1.582*281*3.364 = 55.9 feet (flow rate of 2 gpm)

H=0.0374*1.582*281*2.033 = 33.8 feet (flow rate of 1.5 gpm, slightly higher than 2 feet per second velocity)

The Taco 009 circulator can do this, but when you back up and look at your system, the change in temperature is:
f=Q/(450*T)
1.5=72000/(450*T)
T=106.7 degrees

So realistically what would happen is
1.5=Q/(450*20)
Q=13,500 BTU delivered

1.5=Q/(450*30)
Q=20,250 BTU delivered

In other words due to your tube size, you can't circulate enough water to deliver all the heat you need. The flow rate should be around 8 gpm with a pex tube size of 1 1/4" for a 20 degree drop. Based on those numbers:

f=Q/(450*T)
8=72000/(450*T)
T=20 degrees

1.25 inch Pex
H=0.000794*1.582*293*38.055 = 14 feet head

The 0012 model matches that pretty good

1 inch Pex
H=0.00223*1.582*288*38.055 = 38 feet head - No Taco pump works, head too high

if you did a 30 degree drop

f=Q/(450*T)
f=72000/(450*30)
f=5.33 gpm, use 5.5 gpm

1 inch Pex
H=0.00223*1.582*288*19.753 = 20 feet head, the 009 and the 0014 match up pretty well.

So.... here is the real problem. The concrete has already been poured so I am stuck with the underground set up that I have in place. Obviously I am not going to jackhammer the brand new floor to put new pipes in lol. So that leaves me with a problem of how do I make it work the best it can with what I have in hand. What would you do?


Sent from my iPhone using Tapatalk

 

[pstnbly]

The tubing is in the slab now so the numbers to maximize efficiency based on the heat source may not work. Also if he can get the slab up to the steady state temp the delta will narrow. The point here is to provide a sufficient circulator to supply the loops at around 4'p/second. In my shop I'm pushing 8 300' loops on a 1" manifold with a 15 - 58 Grundfos on speed 3 and if the floor is cold I get about a 20 degree delta. A 15 - 58 on speed 2 is about equivalent to a Taco 007. A Taco 11 is a multi speed pump. Either of those should work.

At that point if you can't meet your set point on the radiant loops alone you will need an additional zone such as a unit heater to fully utilize the output of your water heater.

As an aside, I really think members here should discourage people from using water heaters for space heating unless those water heaters are specifically rated for space heating. Use the proper appliance for the job. A modcon boiler is surely more expensive upfront but will pay dividends by having heating curves specifically for high mass (garage floor) radiant applications, and outdoor reset controls. Modcons thrive on high delta systems and the selectable heating curves mean in many applications no mixing valves, boiler protection loops or hydraulic separation is needed.

 

[Radix2]

You guys seem to be missing the fact that the OP states he has 6 parallel 1/2in 270foot pex runs?

So if he needs 8gpm at 20deg to carry the load, that is only 1.33gpm per loop. Which is well within the capability of common pumps.

 

[fin5281]

The tubing is in the slab now so the numbers to maximize efficiency based on the heat source may not work. Also if he can get the slab up to the steady state temp the delta will narrow. The point here is to provide a sufficient circulator to supply the loops at around 4'p/second. In my shop I'm pushing 8 300' loops on a 1" manifold with a 15 - 58 Grundfos on speed 3 and if the floor is cold I get about a 20 degree delta. A 15 - 58 on speed 2 is about equivalent to a Taco 007. A Taco 11 is a multi speed pump. Either of those should work.



At that point if you can't meet your set point on the radiant loops alone you will need an additional zone such as a unit heater to fully utilize the output of your water heater.



As an aside, I really think members here should discourage people from using water heaters for space heating unless those water heaters are specifically rated for space heating. Use the proper appliance for the job. A modcon boiler is surely more expensive upfront but will pay dividends by having heating curves specifically for high mass (garage floor) radiant applications, and outdoor reset controls. Modcons thrive on high delta systems and the selectable heating curves mean in many applications no mixing valves, boiler protection loops or hydraulic separation is needed.

Thanks for the help. I am using a seisco sh22 which is made specially for radiant heating. The theory behind it is I want the circulation pump to kick on when it hits what the thermostat is calling for . At that point the boiler also kicks on. In theory the boiler should on,y run when the pump calls for heat. I suppose it would be better to have the pump run less times with more volume than a slow and steady flow. What do you mean by a unit heater. Unfamiliar with that term. Would I be better to divide the 6 loops into 2 or three zones with separate pumps?


Sent from my iPad using Tapatalk

 

[fin5281]

You guys seem to be missing the fact that the OP states he has 6 parallel 1/2in 270foot pex runs?

So if he needs 8gpm at 20deg to carry the load, that is only 1.33gpm per loop. Which is well within the capability of common pumps.

I like where this is going. Does anyone else have this opinion? So each parallel loop divides the total pressure and head pressure by 6 . Hmm maybe I did not make as big of mistake as I thought. Hopefully there are others that think this is correct.


Sent from my iPhone using Tapatalk

 

[Radix2]

I like where this is going. Does anyone else have this opinion? So each parallel loop divides the total pressure and head pressure by 6 . Hmm maybe I did not make as big of mistake as I thought. Hopefully there are others that think this is correct.


Sent from my iPhone using Tapatalk

there isn't any question that you need to calculate the head of your floor loops based on 1/6 the target flow per loop.

your 1in feed to the manifiold is correct for 8gpm

your 1/2in loops are correct for 1.3gpm per loop.

now plug in and get your head for each section.

the head across the manifold is not additive if the loops are balanced - the same pressure will be seen across each loop - so any single loop can be used to calculate it, - i.e. 1.33gpm through 270' plus fittings.

then add the head for your 1" manifold feed.

putting all the flow through a single loop is a fundamental error, I think siteman just misread the layout.

 

[pstnbly]

I suppose it would be better to have the pump run less times with more volume than a slow and steady flow. What do you mean by a unit heater. Unfamiliar with that term.


Sent from my iPad using Tapatalk

Actually the theory goes that having the boiler output match the heat loss of the building with constant circulation would provide the most efficiency. It gets into the weeds technically so I wont go there.

Secondly, just manifold the loops together (in your case I would use manifolds with flow meters and setters). and run a 007 Taco or a 15 - 58 Grundfos. The 15 - 58 is just slightly more money and has multi speeds. Also the pump should pump away from the boiler, the air scrubber/scoop, and expansion tank. This is called the PONPC (Point Of No Pressure Change).

A unit heater is one of those hanging heaters you see in garages, in your case you would need a hydronic model (air - water).