Help Selecting Circulator

prettyarmadillo

I am working on a radiant heating system for our home, and I'm largely done with the design, and partially done with the installation.  I've been working from John Siegenthaler's Modern Hydronic Heating to design the system.  I'm in the process of trying to select a pump at this point, and I'm hoping I might find some guidance here.



I am seeking help because I've completed the system curve, and I'm a little concerned about my results (~21 feet @ 3gpm, see below).  I wonder if anyone can reflect on the head figures that I'm getting, and let me know whether it seems higher than it should be for the system? I might have made an error in calculating the system resistance, or maybe I did the calcs right, but made a mistake in the plumbing design?



In addition to feedback on the system curve, I'm also hoping I might get some opinions on the circulator, whatever size it ends up being.  I like the idea of the intelligence that is built in to the Alpha and others like it (Wilo and Taco?).  But maybe I'm better of with a standard pump?  Any help anyone is willing to give is greatly appreciated.



The house is ~70' long, so I have 1" trunk lines feeding three manifolds, with 11 loops across the three manifolds.  My plan is to throttle the loops to get the heat balanced, and then leave it at that.  I'm not wanting to use setbacks, just a single controller.  I haven't fully figured out my control scheme yet, probably a single stat, or maybe an outdoor reset?.  My target for flow is ~3gpm, as I understand noise becomes an issue at higher flow rates.  Since I'm not trying to do setbacks, nor do I have active separate zones, I didn't think that charge time for the conditioned space would be much of an issue; once it's up, it'll just stay up until the heating season is over.



I already own the tankless water heater, I have yet to buy most of the mechanical components (I have the feed valve, but not the exp. tank, trap, backflow preventer, gagues, pump).  The tubing, manifolds and trunks are all laid out, plumbed up and holding air (city required 100psi for inspection).



Here are the specs:

House square footage: ~1800 SF net

Design Load: ~ 21,000 @ 10F

Head Loss Figures (Flow, gpm : Head, ft): 1:3.21, 2:10.79, 3:21.94 (Curve attached)

Heating Tubes: 1/2" tubing on top of wood subfloor

Finish Floor: 1x6 T/G



Heat Source: H-12-2 tankless electric (Specs attached)

Air Trap: Spirovent

Pump: ?

Trunk Lines: ~ 110' (50' supply, 50' return) 1" pex

Loop Lines: ~ 2000' 1/2" pex

Manifolds: Everhot, w/flow meters and throttle valves



Manifold 1:

    Loop 1: 120'

    Loop 2: 200'

    Loop 3: 200'



Manifold 2:

    Loop 4: 180'

    Loop 5: 130'

    Loop 6: 200'

    Loop 7: 280'



Manifold 3:

    Loop 8: 170'

    Loop 9: 200'

    Loop 10: 200'

    Loop 11: 180'



Again, my main concern is whether the system curve looks "right" to the experienced eye, given the general design?  If so, any recommendations on pumps?  If not, where do you think I might have gone wrong?



I've tried to provide all the relevant details, but I'm sure I may have forgotten something.  Please don't hesitate to let me know if there's something missing that would help.



Many thanks in advance for any guidance you wish to give.

Mac_R

Circulators

I really like the Taco 008 Delta T circ.  they also make a 0010.  This pump also has an outdoor reset module built in. 

http://www.taco-hvac.com/uploads/FileLibrary/101-091.pdf

CMadatMe

GPM and Foot of Head

You must be designing around a 15 degree delta-t.

3 x 15 x 500 = 22,500 Btuhr

Your longest loop length is 280' of 1/2" Pex. You have 11 loops so each loops is seeing .27 gpm and your longest loop of 280 has less then 1' of head. We don't total all of the head just the longest. If the pump can overcome that it can over come all.

Your supply length I assume is a total amount? How far is the furtherest manifold to and from the circulator? Even if I do your total amount with the full 3gpm I only have 1.5ft of head. So I'm at 2.5ft

Now to the mistake. Electric tankless..

40,000 btu's divided by temp rise divided by 500 = gpm

Now temp rise it only going to be an estimate. On start up you have about 50 degree incoming water and I'll estimate you need 115 degree water for the radiant. That's a 65 degree rise. Do the math and that unit is only going to give you 1.23gpm until you get the water temp. As the rise decreases your flow rate will increase as return temps get warm but god is that electric meter spinning.



Can't help you out with the pressure drop across that tankless no specs. I'm sitting at 3gpm and 2.5ft of head with the head of the tankless needed.

Mark Eatherton

No fitting allowance?

Even with copper, standard practice is to add 1/2 the worst case length to allow for fittings. With PEX, its pressure drop is even worse. I'd do your longest length time 1.75 for calculation purposes,



Curious of the original poster, did you total up all of the pressure drops of all branches?



ME

CMadatMe

I Used

My pex head loss calculator for the runs. Based is on 100% water at 120 degrees. Point being that he needs to go back and correct his homework.

prettyarmadillo

Losing my Head in the Head Loss Calc

Mac_R, Thanks for the Taco VS idea.  I checked it out, and it seems similar to the other smart pumps.  Sounds like a good idea, I just need to get my head figured out so I can choose an appropriately sized pump.  Also, I have not been able to find Taco data that identifies the pump curves (?).  Probably me not understanding the Taco site.



Chris, Thanks for the thoughtful review and inquiry.  I was thinking toward a 10F delta or so, but perhaps 15 is better?  But yes, in that neighborhood.  I didn't realize that you only consider the longest run, not the aggregate for head loss calculations.  I think I can get my head around that.



Yes, the 100' supply figure refers to the total round trip for the trunk lines.  The longest 1" supply is about 50' of pipe to get from the mechanical room to the most distant manifold, one way (another 50 to get back to the m/r).



Mark, I did run figures for all my loops.  I used the process in the book for calculating parallel circuits; first for all loops on a manifold, then for all the manifolds in parallel, then I tallied that with the resistance that I calculated for the mechanical loop.  Interestingly, total resistance seems to go down when things are in parallel.



I did an accounting of equivalent feet for fittings and components as well; both in the mech. loop as well as in the trunk lines and the manifold valves/meters.  I did each pipe size independently, then tallied their "r" values (equivalent lengths x pipe size coefficient).



I'm confused by the head loss calcs.  I must be doing something wrong somewhere.  The formula I saw was "H_L = r(ƒ)^1.75 = (acL)(ƒ)^1.75".  Breaking it down, I feel pretty good about my alpha number, 0.050764 for the water at high temp, and 0.060073 at low temp, and the lengths are pretty easy, not sure about the "c" value, but I got it right out of the book.



But I just can't get the low head figures that Chris does.  Even just considering the biggest loop, if I have 230 equivalent feet of 1/2" pipe, and I use the "c" of .786 for 1/2" pex from the book, and either alpha value, I get nothing like what Chris does, I get...

r = .050764 x .786 x 230 = 9.18

...and I haven't added the 1" pipe yet!



If I understand the process, you now take 9.18 x 3gpm^1.75 to get 74.27!  It's getting worse and worse!  I know I must be doing something wrong in there, but I can't figure it out.



Chris, I think I understand what you're saying about the output issue with the tankless, please tell me if I'm getting this right: with the initial delta from 65 to 115, I'm running water through that just can't get up to 115 for a while, probably has to make a few trips to get up to temp, and meanwhile it's also throwing heat into the cold floor, so further delays getting that delta down.  Am I following?



I totally agree that the meter is buzzing like a bee.  But isn't it the case that the heat has to get in there one way or another?  So unless I choose another energy source, all that's being wasted there is the pump work to move that tepid water through (not trying to overlook the power consumption of the pump, but it's nothing compared to the elements), whereas if I had a bona-fide electric boiler, it would hit temp on its first pass.  But in the end, don't both the boiler and this unit have to put out the same amount of energy to heat up that water (and space)?



If I'm understanding the situation (which I may well not be!), and let's assume I'm not going to shift to natural gas (not an option in our situation), does this mean that it's a good application for a "smart-pump" running on a delta T or something, so it sees that it needs to run a little more slowly to let more heat into the water, until the floor and the water get charged?



I hope I don't seem cheeky, I'm not trying to be, I'm just trying to understand all the implications, and I value your input highly.  I don't totally understand how the Alpha and its friends do their magic, or where it's best used.



FYI, here are some of the figures I used for the head calc, re-tallied with Chris's input in mind:

200' of 1/2" pex (longest loop at greatest distance from pump)

15' feet for the throttle valve in the manifold*

15' for the flow meter in the manifold*

(*these are made up equivalents; I couldn't find data for the valve and meter on the manifold, but chose the 15' based on other fittings in Siegenthaler's book that I imagined had similar characteristics; I tried to be "conservative")

...which makes 230 equivalent feet of 1/2" pex

times the pipe size coefficient from the book, .786

for an "r" of 9.18.

Now for the trunk line plus a bunch of 90s, the air trap, the tankless*, some Ts for gauges, I had about 165' of 1" pex... I got an "r" of .36 on that.

(Chris, I logged the tankless at 15' equivalent; I called the manufacturer, and couldn't get anything useful... but they said its internal cavity has 1 1/4" entry and exit ports, and the cavity in which the elements sit is larger than the openings, so I'm hoping this is a reasonable guess)



Thanks in advance, again, to all who take their time to help me understand this stuff, I truly appreciate it.

cattledog

calculation using wrong flow rate

Armadillo--



In your head loss calculation where you take 3 gpm to the 1.75 you should be using the flow in each individual loop--i.e. .27 gpm.



As Mark and Chris have indicated, the head loss in the loops and other piping is modest. The big unknown is the head loss at different flow rates in the water heater. What is the manufacturer's data for head loss with flow through the heater? It is going to be the biggest factor in your system, and you can't get to a pump selection until you figure out the resistance of the tankless. It is surprisingly large.

Gordy

Phantom information

On pressure drop for this unit. They make an RH-12-2 for radiant heating only difference being an added pressure switch. I don't think you will get much out of the unit if you go higher than 2.1 gpm anyway 40* rise at that rate with 125* max temp setting.

prettyarmadillo

Wrong Flow & Phantoms

Thanks Cattledog, for the scoop on the flow rates.  So you just divide desired flow by number of loops?  I think I get it now.



I hear what you're saying re; pressure drop across the unit, it's clearly a critical number to obtain.  There are no published specs that include the pressure drop or any sort of equivalent pipe length figures.  I talked to the factory, and they had no idea.



I hate to go too far off into farmer John land, but since I am already in possession of the unit, is there a way to clinically determine the pressure drop?  I guess I'd need a way of pushing a known rate of flow through it, with gauges on each side?  Seems a little out of reach.



Gordy, do you know if there are published numbers for the RH?



Or, is it possible for me to back into the pressure drop, knowing some recovery rate information?  I've attached a chart from the mfg's web site, in hopes that maybe there's something there.

CMadatMe

The Manufacture

If a stand up company should have someone in their technical or engineering dept who has the anwser. Guessing is just cause you more headaches.

Gordy

Same unit

As yours. Only difference in the engineering is recommendation of 500sqft of radiant. Pretty vague. not even a gpm rating in that chart.



I can tell you this much. I have an ariston 4 gal electric hot water heater hooked up to about 300sf of radiant. 4 loops 160' each with a laing 40 watt pump as an alternate heat source. Its a 1.5 kw unit If I use it never shuts off. EXPENSIVE to operate in my grid. But It works.



Gordy

prettyarmadillo

Alternative Heat Sources?

Thanks Gents, for the further thoughts.



I'll try the factory again to see if they can come up with a pressure drop figure of some sort.  Not sure how to integrate if they produce pressure drop rather than equivalent feet though.



Even if I'm able to get a good number out of the factory, I'm hearing pretty loud and clear that there are strong feelings that this is a poor choice of heating unit, mainly centered around the fact that it'll likely be running pretty constantly to try to maintain water temp.



I'm open to changing, don't want to seem like I've got my heels dug in, and I certainly don't want to saddle the system or our house with the long term punishment of a poorly chosen heat source.



May I ask for suggestions as to alternates? I can't afford an air/water or ground-source heat pump right now, so, without going to gas or oil, I'm not sure what my alternatives are? (I have the quiet hope that perhaps one day I might retrofit a heat pump of some sort into the system, if I can afford it).  Love to hear what your thoughts are?



Thanks.

Gordy

Mod/Con NG

Dunkirk, Knight, Peerless, Triangle tube. Make models close to your heating load. 95-98% EFF.

If your Grid rates are cheap. Plenty of good Electric boilers that modulate to the load, and have ODR.

Gordy

Wondering

 If the rh-12-2 is saying it can only heat 500sf. If 2.1gpm is all this unit can handle. That leaves an average flow rate of .19 gpm per loop.

prettyarmadillo

Sticking w/Electric...

Thanks, Gordy, I guess I'll try to start digging into the electric boilers, can't do the NG.  Not sure what ODR stands for?



Still don't totally get the boiler vs. the little guy though, simply from the question of total heat input.  What's the difference whether the boiler gets the heat into the system at high output right away, vs. a slow climb with the smaller unit... if the floor needs 21kbtu/hr, does it matter if it gets it all in the first 20 minutes, or over 40 (other than a longer pump run-time)?

Gordy

The difference

Is modulation ability with some, and outdoor reset control (odr). If you get a 4 element then those elements will stage to the load one at a time not always 12 kw all the time. Plus the ability to get higher flow rates across the HX

prettyarmadillo

Electric Boilers Elaborated

Thanks for the info.  Outdoor Reset certainly seems like a good idea, and higher flow is sure to be good.

prettyarmadillo

Hmmm...

Gordy, I'm not sure I follow your train of thought on this?

Gordy

Not capable

Of the flow rates you may need

Gordy

Not capable

Of the flow rates you may need

Gordy

Not capable

Of the flow rates you may need

Gordy

Not capable

Of the flow rates you may need

CMadatMe

Electric Boiler

Uponor, a leading mfg of radiant has a electric boiler they promote as radiant ready.



http://www.uponorpro.com/en/Modal-Pages/Products-and-Tools-Modal-Page.aspx?id={58BC4834-4676-4A2D-B493-1EED6079AF80}