How to calculate pump size for multi zone conversion

jeremypbeasley

Hi, folks. 

I’m converting a single zone hydronic system into a 3 zone system (one zone for each of 3 floors). We had to remote the low hanging pipes in the basement ceiling due to the lack of head clearance and asbestos wrap. All radiators now have 1/2” hePEX home runs and converge to 3 Uponor 7-loop 1” manifolds. 

My piping between the boiler and manifolds mirrors what is called for in the boilers manual below.

I’m sold on a 3 pump system vs a 3 valve system, but I’m hung up on choosing my pump size for each zone.  

The system before, on the same boiler, but with larger diameter steep pipe had a 1/25HP taco cast iron pump (007-F5) circulator for the entire house. And it seemed to work just fine. 

When I calculate all the new plumbing GPM, I’m led to believe I need MUCH more power from 3 pumps. 

Am I calculating this correctly? 

For example, let’s use the the top floor (Zone 3). 6 radiators and one stubbed to be added later. Based on available info online for these American Radiator Co. Roccoco radiators, the total BTU is 17,800. We’ll call it 21k to allow for a needed addition 7th radiator. 

Let's assume the temp drop is 20F. I understand the formula for GPM to be 0.002*BTU/(Temperature Drop, F) So, that's 2.1 GPM.

Each radiator has a vertical run of 1/2” 11ft hePEX That's 77 feet of vertical tubing. Plus an average of 30 ft per radiator of horizontal run in the basement ceiling. 30x7= 210 + 77 = 287 ft. Let’s round up to 300 ft. 

Uponor hePEX sized at 1/2" with 180F water has a pressure drop of 12.2 head per 100ft of tubing. So with 300ft of tubing. That’s 36.6 drop of head. 

That means we need a pump that can deliver 2.1 GPM flow and a total head of at least 36. 

This Taco 009 1/8HP model fits that description.

Taco 009 Stainless Steel Circulator Pump, 1/8 HP

https://s3.amazonaws.com/s3.supplyhouse.com/manuals/1347286901614/79539_PROD_FILE.pdf

Is this the correct pump size? 

hot_rod

The pressure drop is for the longest loop off the manifold, not all the home runs combined. So would that be more like 40’ up and back to the most remote radiator? So the head is for 40’ of Pex, not 300’

jeremypbeasley

hot_rod said:

The pressure drop is for the longest loop off the manifold, not all the home runs combined. So would that be more like 40’ up and back to the most remote radiator? So the head is for 40’ of Pex, not 300’

Well that’s one heck of a detail to get wrong! Thank you. 

So if 1/2” hePEX at 180F water has a pressure drop of 12.2 head per 100ft and my longest run off each manifold is under 100ft, my pump requirement then is 2.1GPM and 12.2 head. 

Is it also true that the hePEX pressure drop can be calculated in shorter spans? For example, if it’s 12.2 per 100ft, is it also true that every 10ft, the drop is 1.22? If so, assuming my longest run off each manifold is was 40ft, I could get a pump that’s 2.1GPM but only 4.88 head (1.22 x 4). Is that sound science? 

hot_rod

You add up the gpm requirements of all the radiators. So the upper floor has 6 radiators and you calculate 1.7 gpm each, so 10.2 gpm , at the head of the longest run from that manifold being served by the pump you are selecting.

Possibly not all radiators need 1.7 gpm, it depends on the load of the room they are located in, not the size of the radiator.

You may have some additional piping to the manifold. Add a foot or so, call it 10 gpm at 7’. Well within the range of a small 1/12 Hp circ, or ECM style

Teemok

hot_rod beat me to it I'll post it anyway. You are designing for boiler flow and zone flow. When calculating the head needed for each manifold we look for the highest head loop on that manifold at the desired flow.
Loops in parallel don't add the head they do add flows. So each radiator needs 2.1 gpm and they are each on their own 1/2 loop. Find the most restrictive loop and design for that head while the pump moves 7x2.1gpm=14.7gpm I'll guess: loop head is under 4'. Now you add the head of the 1" distribution piping, near boiler piping and the boiler. Wild guess again: + 3'. So for this single zone the pump with 7 loops must move 14.7 gpm @ 7ft of head. Hypothetical.
That is just one zone/manifold. Let's say the other two floors are the same. With all three pumps running you'd have 44gpm trying to go through the boiler. It wouldn't because the flow dynamics with all pumps might have a much higher head in the piping and boiler so flows would be lower.
The boiler has to compatible with one zone calling and all three. Desired 20F delta T flow is around 1gpm per 10,000 btu typical. A 140kbtu boiler would be happy with one zone but all three is a lot of flow.
This where primary/ secondary piping shines. Give the boiler it's own pump sized for its needs alone and size the distribution piping and pumps or single ecm pump for the piping and zones head and flow needs. A constant pressure ecm pump with 3 valves might be what you need. Or maybe you'd like three 0015e's better.

hot_rod

I agree with @Teemok A delta p circ with zone valves would make a system like this sing
The delta p function allows the circ to adjust to the different loads as valves open and close.

Teemok

Most zone valves will add significant head at higher flow rates. If you go valves make sure you add their head at zone flows to the calculation.

Teemok

If the boiler has too much flow through it with all zones calling and its a lot smaller than the total radiation it could run in a low temperature (condensing) condition for a while. Some form of boiler minimum temperature protection might be needed.

jeremypbeasley

I really appreciate all this feedback. That said, I’m pretty overwhelmed now. I’m a homeowner in a tough spot (37F last night and I’ve got 3 kids in the house) —not a professional by any means. 

It sounds like if I ran with my original plan, I need much smaller zone pumps than I thought. Which is great. Sure beats paying $500 per zone for a pump. 

But now I’m questioning going back to a one pump system and valves given what you’ve all said. Frankly, I don’t understand how the boiler would he under any less stress having one pump cranking with 3 open zone valves vs having 3 much smaller pumps going at it. The advantage isn’t obvious and I don’t entirely understand how a single pump, multi-zone systems would operate. 

That said, my boiler is 133 BTU (Weil-McLain CGa-5) and I’m concerned about overwhelming it or at the very least having a very inefficient system. 

Questions:

1.) If I choose a single pump for the whole system (using Taco valves), how do I size it? Is it just the sum of what each manifold would need inidivudally? So in my case 2.1GPM. Then consider the head loss for each. That’s 12.2 each. So triple that and the GPM? So 6.3GPM and 36.8 Head? That’s a huge pump. What am I doing wrong here? 

2.) With a single pump and zone valves, should I be concerned about my 133k BTU boiler being able to keep up? 

3.) Bonus Question: can someone outline what a Delta P setup would look like specifically? I’ve read for hours and there are so many different configurations, I’m quite lost.

Thank you all so much for your time!  

jeremypbeasley

Welp, I answered #1. When sizing a circulator for a zone valve system, you DO add up all the GPM needed for each manifold/zone but you do NOT add up the head loss for each. You only need to consider the head loss for the zone with the highest head loss. In my case, 12.2 on headless and 2.1GPM. 

I got this information from the extremely helpful Taco training video here. Skip to 44:00 if you want. 

https://youtu.be/PhNaYBuTVWE?si=_g_ZpZ2foWZktBoC

Rich_49

Well there's no surprise . John Barba may very well be the most interesting and engaging hydronics trainer there is .

hot_rod

Really what happens in a hydronic system is the radiators dictate the operating condition of the boiler. If the boiler is sized to or larger than all the radiators you have, there in no way you will stress the boiler.

Most cast boilers need little if any flow to operate. It is just a big cast iron vessel with a fire below. Once it gets to temperature it shuts off regardless of how much, or no flow.

Think of a delta P circulator as a pump with a cruise control. It will automatically increase or decrease the speed (gpm) as valves open and close. So you are never over-pumping the systems as you would with a basic single speed circ when only one zone is open.

Many times, system zones with pumps are over pumped as they rarely have balance valves installed to adjust the flow to the actual load. They get installed wired up and run wherever they want on the curve.

Teemok

"12.2 on headless and 2.1gpm" = I'm not sure what you are saying.

jeremypbeasley

Teemok said:

"12.2 on headless and 2.1gpm" = I'm not sure what you are saying.

Just a typo! 

I’m saying my pump requirements would be the following, assuming my calculations are correct. 

12.2 FT of head loss
2.1 GPM

Teemok

6-7 loops on each manifold right? Each loop needs 2.1gpm right? Break down how you get 12.2 ft of head. You might be right about that if it includes the boiler and distribution piping as well. Zone valves have a rating called the Cv. It is the flow in GPM that will pass through it with an added 2.31ft of head. The more flow you need to pass through it the higher the added head will be for just the zone valve itself.

Teemok

2.1 gpm in 1/2" pex is pushing it. It can be done if needed but it means higher head (bigger pump). 1-1.5 is more realistic. 100ft of 1/2 pex @1.5gpm is 8ft of head. @2.0gpm is 13.4Ft When you add up all the flows(GPM), from all the loops, on all the manifolds, you will find the total GPM that the distribution piping will need to deliver when all zones are calling. At that GPM the distribution piping will add head to the longest loop head calculation. The boiler might add a little too. If you are using zone valves they will add head related to the zones flow rate to the longest loop head calculation as well.

jeremypbeasley

Teemok said:

6-7 loops on each manifold right?

Correct. 7-Loop Manifolds.

Teemok said:

Each loop needs 2.1gpm right?

Nope. ~2GPM was the total for each manifold. See the table I've created below. Keep in mind Zone 1 (the basement) isn't finished yet but the BTU load will be comparable to Zone 3 with much shorter runs.

Teemok said:

Break down how you get 12.2 ft of head. You might be right about that if it includes the boiler and distribution piping as well. Zone valves have a rating called the Cv. It is the flow in GPM that will pass through it with an added 2.31ft of head. The more flow you need to pass through it the higher the added head will be for just the zone valve itself.

I get 12.2 because the Uponor table here tells me that for every 100' of 1/2" hePEX at 180F (my max temp), it's 12.2 ft of head. Now, my assumption is that if I have say, a 40ft run, I can multiply 12.2 by .4 and get a more accurate number. Is that correct?

As for calculating head loss for manifolds, valves, and the boiler; I'm lost here. I've read so much contradictory information online.

I was planning on using Taco's 1" zone valve here. It has a 8.9CV rating.

How do I calculate the head loss for the above valve, my manifold, and the boiler.

If my table isn't flawed, how do I size a pump to meet the needs of 3 valved zones? I'm assuming the GPM is added up. So 2.0 + 2.8 + 2 (we'll assume this for Zone 1 / basement) = 6.8 GPM. There there's head loss. Is that added up like GPM? If not, my highest head loss is Zone 2 at 19.9, meaning I need a system circulator pump that can handle ~7 GPM and ~20 ft of head loss.

In that case, the Taco 009 1/8HP seems ideal. 10GPM + 36 ft of head loss.

Am I getting any closer here, folks? I appreciate you sticking with me!

Teemok

Edit: These calc's below are way off for you after seeing the charts you posted. The theory is good but your case is much easier.
The needed flow rates for radiators is based on their radiation capacity at peak supply water temperature. Your radiators may not need 2.1GPM each. (my misunderstanding) A radiator that loses 20 degrees F of water temperature from supply to return with a 2.1GPM flow is giving off 21,000 btu. That's a big radiator.
This might be the total for all the radiators on one floor? Small radiators could do with less flow so your real flow needs might be less than the calc below.
1.5gpm gives you a 1/2 pex flow velocity of 2.7ft/sec. good for air removal.
1.5gpm x 7 radiators is 10.5gpm per 1" manifold. A Caleffi 7.5CV zone valve flowing 10.5gpm will add 4.5Ft of head.
If your longest 1/2" radiator loop is 30' each way 60' total. we get .6ft of head per 10' @1.5gpm. 60' of 1/2"pex @ 1.5 gpm is 6ft of head. The valve + the loop=10.5ft.
Now add the 1" home run piping head while flowing 10.5 gpm at the total equivalent length of pipe (convert fitting to pipe length and add) to shared piping.
Then add the shared piping and boiler head when flowing 31.5 gpm (assuming all floors/zones are the same.
You had 12.2ft. You might be close depending on the boiler piping and runs to the manifold size and length.
The Taco 0034e seems like it might be oversized. The 0026e would be a compromise on cold starts but would work well once the on/off cycles of the zones fall out of sync. Or if true flow needs are lower it would be a great high efficiency single pump.
If you go individual pumps the up15-58 (this is typical/cheapest) might be slightly on the small side or it might be great if true needs are more like 8gpm at 12ft of head.
The up26-99 on medium would give lots of flexibility but costs much more.
Hope this helps some rather than confuses.

Teemok

Just saw the chart you posted. 23.6Ft of head for 100ft of pex is very high. You are only flowing 0.6 gpm. It is closer to 1.5Ft for 100ft of pipe

Teemok

I don't know the the flow specs on your radiators or your home runs equivalent lenghts or pipe sizes from manifold to the boiler but looking at the numbers on your chart makes me think that 3 Grundfos up15-58 pumps is all you need. Flow rates will be higher than the chart states likely by double or more and that's good for air removal. Balance flows at the manifolds for radiator sizes if you can. Above 1 gpm for big ones and around 1 for the smaller ones. Restrict as little as possible. The chunk of cast iron boiler will be fine with 8 gpm from a single zone call. I hope a cast iron expert chimes in. It's been 25 years since I installed cast iron. I think hot_rod was saying that lower flows would be ok previously on this thread.

Rich_49

Which Rococo rads are you using first off ? Are you using zone valves and 1 circ or 3 circs ? What type boiler do you have ? Why did you choose a 180* SWT as opposed to something else ? Bare with me here , we'll get you where you need to be .

jeremypbeasley

Rich_49 said:

Which Rococo rads are you using first off ?

I'm rebuilding an existing system. The radiators are all original to the house, made by American Radiator Company, their Rococo models. A company in the UK makes carbon copies of these units and that is where I'm getting the BTU ratings for each unit. Link to my spreadsheet that I took a screenshot above, here: https://docs.google.com/spreadsheets/d/1tDbQUl-_PsNaXY287UobXhK4MHdBz1AC6_ZLtPxq8EY/edit#gid=0

The boiler is an existing 133 BTU Weil-McLain CGa-5.

All other plumbing has been removed along with the asbestos it was covered in. I kept the circ pump just in case.

Rich_49 said:

Are you using zone valves and 1 circ or 3 circs?

3 zone valves (1" Zone Sentry Valve, Normally Open (Threaded)) and one circ pump.

Rich_49 said:

Why did you choose a 180* SWT as opposed to something else ?

Because it was the max the Uponor hePEX head loss specs listed and I know I won't be much hotter than that because the hePEX is only rated to 200F. Beyond that, I have no more rationale.

Rich_49 said:

Bare with me here, we'll get you where you need to be.

Thank you so much!

jeremypbeasley

Teemok said:

Edit: These calc's below are way off for you after seeing the charts you posted. The theory is good but your case is much easier.

All good! Yeah, that table has all the good stuff.

Teemok said:

The needed flow rates for radiators is based on their radiation capacity at peak supply water temperature. Your radiators may not need 2.1GPM each. (my misunderstanding) A radiator that loses 20 degrees F of water temperature from supply to return with a 2.1GPM flow is giving off 21,000 btu. That's a big radiator. This might be the total for all the radiators on one floor? Small radiators could do with less flow so your real flow needs might be less than the calc below.

Yes, 2.1 GPM was for the whole zone, adding up the GPM need for each radiator in that zone. The average across those 6 radiators is .29 GPM per radiator.

Teemok said:

A Caleffi 7.5CV zone valve flowing 10.5gpm will add 4.5Ft of head.

Can you explain how I get from the CV rating to the amount of head loss from the zone valve I need to add to the overall head loss? I don't follow your math on this so far.

Teemok said:

We get .6ft of head per 10' @1.5gpm.

How did you get this number? My head loss is just coming from the tables Uponor provides for hePEX here.

Teemok said:

Then add the shared piping and boiler head...

How do I find out what head loss my boiler itself introduces? Also the manifolds? I see nothing in either of their specs about pressure or head loss.

Teemok said:

Hope this helps some rather than confuses.

Definitely helping!

Teemok

The chart you are reading is for 1/4" tubing. Head loss is figured by pipe size and type, equivalent of length of pipe at a flow rate. If your 1" home runs are 32' each way and have 6-90's total. A 90 is equivalent to a length of pipe say 5'. So we figure 2x32'+(6x5') = 94' of 1" copper. Find a chart or calculator to find head of 94' of pipe at the gpm the zones flow rate. Say 8gpm. Your chart flow rates are low. Shared piping uses all the flow from all zones going through the equivalent length of that size pipe. The boiler head will be low. Most of the time only one or two zones will be calling/pumping/flowing at a time. Toss 2-3Ft in. You will likely find a Grundfos ups15-58 will do for each individual zone. It's a 3 speed pump using high speed. It is the most affordable good performing standard design residential sized pump. If you zone with pumps you don't need to deal with zone valve Cv values. If you want to play with numbers with a valve zoning design. A 1" Calleffi valve has a Cv rating of 7.5. This means it will flow 7.5GPM with a 1 psi pressure loss or head loss through the valve. Convert: 1 PSI = 2.312 ft of head. Flow 15GPM and head goes up to near 2PSI or 2x2.31=4.62ft of head. Flow 8GPM and you can extrapolate head as just above 2.31', say 2.45' for just the valve. 2.5' would be a safe number to use for a 1" Caleffi 7.5Cv zone valve flowing 8GPM. So loop longest length head at 1-1.5 GPM = (X)
Home run at the whole manifold flow rate (8-9GPM) =(Y) Boiler and near pipe guess is 2ft of head. If you want to use zone valves rather than individual pumps add the valve head as 2.5ft. So we add X+Y+2'+2.5'= total head for the longest loop. Maybe 12' of head @ 8-9GPM look at the ups15-58 pump curve. It flows just under 8GPM at 12' of head. If your real head is less than 12' it will flow 9-10 GPM. If it's higher than 12' than you will be splitting up 7GPM between the 7 radiators. So about 1 gpm per radiator at worst case. Now you do the math.

jeremypbeasley

Thanks so much for all this info!

Teemok said:

The chart you are reading is for 1/4" tubing. Head loss is figured by pipe size and type, equivalent of length of pipe at a flow rate.

Ah, it looks like you're just looking at the first page. Page 6 has the tables for 1/2" tubing. Easy to do!

Teemok said:

Your chart flow rates are low. Shared piping uses all the flow from all zones going through the equivalent length of that size pipe. The boiler head will be low. Most of the time only one or two zones will be calling/pumping/flowing at a time. Toss 2-3Ft in.

Makes sense. I'm not accounting for anything but the PEX. Account for all this makes my head hurt, but I'm trying! Here's the rest of the head loss I'm accounting for:

Boiler: 3 ft (Because you said so )

Radiator connections:
63, 1/2" 90 deg Elbows (3 x 21 radiators)
42 1/2" 8" Nipples (2 for each radiator)

According to this each ft of SCH 40 1/.2" steel pipe is .0021 ft of head loss at 1 GPM. So my nipples add up to 28' ft of pipe and thus .0588 ft of head loss. A 1/2" 90 steel elbow, according to this, is equivalent to 1.5 ft of pipe. With 63 elbows total, that's equivalent 94.5 ft of 1/2" steel pipe and .198 ft of head loss.

21, 1.2" radiator ball valves:
I can't find any tables that specify the length equivalent of a ball valve, but let's assume it's similar to the gate valve in the table above. 21 valves, equal to .35 ft each or 7.35 ft total. 7.35 x .0021 ft of head loss = .0154 ft of head loss.

3, 1" Zone valves: I still don't understand how exactly to get from CV to head loss for a unit, but here's what I did.

My Taco 1" Zone valve has a CV rating of 8.9. I used this calculator to take the 8.9 CV, the 3 water multiplier (for 3 x 60F= 180F), and the 2.8 GPM for that zone. It shows me .84 ft of head loss.

3, Manifolds: I can't seem to find how these contribute or how much to head loss.

Teemok said:

Now you do the math.

Ok, now it comes to adding this up.

I've read that the pump requirements are NOT determined by adding the head loss of every loop, otherwise, that would make the pump needed for a single circ gigantic (200-300 ft of head loss!!). I'm still not clear on what head loss to add up though. If I'm just accounting for the head loss in PEX, my longest loop on Zone 2 is 19.9 ft and Zone 3 at 14.2 ft.

Am I just picking the 19.9 ft, adding the head loss for the zone valve (.84 ft), elbows, brass pipe at the radiator (.004 ft), valve at the radiator (.00007 ft) , and the boiler (3 ft)? If so, we're looking at ~24 ft of head loss.

Did I do that right?

If so, the UPS 15-58 FC maxes out at 19.31 ft of head loss. That makes me think the Taco 0011 Cast Iron Circulator, 1/8 HP might be a better fit since its max GPM is 10 and head loss is 36 ft., both plenty above what I need.

Rich_49

The Cv number simply tells you how many gallons per minute through the DEVICE impart 1 PSI of pressure drop . Cv is not a linear number either , with greater than listed flows it increases exponentially , proceed with caution , I have cleaned up after many " smart contractors " who made this mistake . 1 PSI = 2.31' head . Look up any episodes of Taco after dark where Cv is covered and invest the time .

Rich_49

Please share your room by room heat loss calcs with us . It very well sounds like you'll possibly never need 180* water

hot_rod

Here is a Cv calculator.

Use column 3 if you know the Cv and the flowrate you are considering. Other columns for troubleshooting and determining Cv of an unknown device.o

Teemok

The calculation you want is just the longest loop. 19ft sounds very high. You already have Taco zone valves? I said 3ft of head for the boiler not as a fact but as a number for a calc. example. You never gave an exact boiler model number. Cast iron boiler head loss is normally very low. @8GPM it might be 0.5' or less. When @24GPM it will be more depending on your model. Your near boiler piping size is? The piping from the manifolds is how long and what size, how many 90's and joins together with the other manifold pipes where and how? Total nipples and 90's is not needed just the fittings involved in the longest loop. Think of the water traveling a commute. Leaving the boiler is a freeway, it might be big but if there's variable traffic you could get some head loss. Once it reaches the pipe to the manifold it's an avenue of some size and length twisty or not but designed for the fixed amount of trafic. Then it hits the manifold, low head loss. Then it enters the longest pex loop, (a small road with a few bends) Then it makes it to work, the radiator (low head). The resistance to flow to get there is most of the time similar to the way back. The complete loop is what we are considering. Road sizes, types, turns, restrictions and traffic conditions make up the effort needed to make the commute. Head loss is friction against the movement of the commute (Flow). Head loss is energy that must be spent in order to get a desired flow rate using that particular long path. You don't add up all the possible routes you could take, just the longest path considering all the worst case traffic conditions and known restrictions of that route. Flow is the combination of a velocity and a volume. The pump makes a pressure difference that creates flow. The head loss, volume of water(GPM) and velocity (Ft/sec.) are factors of the pressure difference made by the pump and the resistance to flow in the path of circulation. If a path narrows velocity goes up, as velocity goes up friction/head loss increases and flow volume decreases. There is a system energy of flow balance that exists for every circuit. We are estimating the energy needed to create the right volume of flow in the longest route. If the pump can do that it can move water in the other paths. Now we have a pressure vs volume and motor size compromise that identifies our pump specs.

Teemok

You can do all your calculations for the flow of 0.2-0.6gpm per radiator but you are designing for the absolute lowest functional limit at 180F.
Increasing flow will give you some room to lower water temperature and have some extra capacity to cover a miscalculation. We don't design for absolute minimums unless that's the requirement. We want to use the least expensive, durable and easily available materials to achieve a good compromise between performance, investment and efficiency. Goldilocks choice is not too big not too small. Costs, durability and efficiency all at reasonable levels.

jeremypbeasley

Rich_49 said:

Please share your room by room heat loss calcs with us . It very well sounds like you'll possibly never need 180* water

Seems like it! 

Here’s my attempt at a heat loss calcs for each room. https://docs.google.com/spreadsheets/d/11osKwGUeEeKfwiDMhCqQ4W1v_IVLdIYLWpo1lYZwDAQ/edit#gid=0

What I can’t figure out for heat loss are a few things: 

- How to account for air infiltration. Everything I read either says set up a blow fan (no thanks) or measure the CFM from your forced air system to calculate Air Change per Minute.

- How to account for the ceilings. Do I treat that just like a wall but with a higher R value given the blown insulation in the attic? Do I count the ceiling in the basement or main floor where there is a heated space above it?

Either way, the total heat loss in each room compared to the BTU output at 170F for each rooms radiator is about 47% on average. That tells me I have plenty of room to lower temp from 180F and would never have to go that high. 

Is that correct? 

The question remains, what temp would I need to operate at? I have no idea how to calculate that yet it’s what I need to get an accurate head loss calc on my system.

jeremypbeasley

Teemok said:

The calculation you want is just the longest loop.

Understood.

Teemok said:

19ft sounds very high.

I think what happened here is that I was calculating the head loss in the PEX itself incorrectly.

The Uponor 1.2" hePEX table tells me that at 180F, every 100' I'm going to get 12.2 ft of head loss at 2.1 GPM. But that's the GPM for the whole manifold, not the loop itself. Most of my loops only need .2-.6 GPM. See my new calcs below.





My formula for getting that GPM is: GPM = 0.002*BTU/(Temperature Drop, F)

For me that’s: 0.002 x 3690 BTU (Existing radiator's raiting at 170F) = 7.38. Then divide that by 20 and you get .4 GPM.

Max head loss on any loop is 2.6-3.0 ft with just the PEX.

Did I do that correctly?

Now to calculate the other components!

Teemok said:

You already have Taco zone valves?

Nope. Haven't even installed anything around the boiler yet.

Teemok said:

I said 3ft of head for the boiler not as a fact but as a number for a calc. example. You never gave an exact boiler model number. Cast iron boiler head loss is normally very low. @8GPM it might be 0.5' or less. When @24GPM it will be more depending on your model.

My boiler is 133 BTU (Weil-McLain CGa-5). Here's the manual! I can't find anything about head loss or pressure drop. I'll assume .5 for now since my whole system should be ~7GPM.

Teemok said:

Your near boiler piping size is? The piping from the manifolds is how long and what size, how many 90's and joins together with the other manifold pipes where and how? Total nipples and 90's is not needed just the fittings involved in the longest loop.

Fittings on the boiler itself are 1 1/4". Planning on 1" steel pipe between boiler and manifolds.

From the boiler to the manifolds Zone 2 and 3 is about 14' of pipe. To Zone 1 it's just 6'.

Each radiator has a 1/2" or 3/4" nipple 8" long. Each radiator has two 90 deg street elbows, an angled radiator valve, and one 90 deg elbow on the return side.

Manifold are each 7-loop 1" Uponor models. Specs here: https://www.supplyhouse.com/Uponor-Wirsbo-A2700702-7-Loop-1-Stainless-Steel-Radiant-Heat-Manifold-Assembly-w-Flow-Meter

Teemok said:

Think of the water traveling a commute...

This was super helpful.

Teemok said:

You can do all your calculations for the flow of 0.2-0.6gpm per radiator but you are designing for the absolute lowest functional limit at 180F. Increasing flow will give you some room to lower water temperature and have some extra capacity to cover a miscalculation. We don't design for absolute minimums unless that's the requirement. We want to use the least expensive, durable and easily available materials to achieve a good compromise between performance, investment and efficiency. Goldilocks choice is not too big not too small. Costs, durability and efficiency all at reasonable levels.

Makes sense to me!

Given my needs of ~7 GPM (each zones total GPM added up) and ~10 ft of head loss (3 longest loops in each zone added together) I may just try to use my existing pump. It's a Taco 007-F5 Cast Iron Circulator and you can see it's performance specs below. Seems well within what I need.

hot_rod

A delta P circ sure would be a nice upgrade on a zoned system like that. All the various flow rates would sort themselves out

jeremypbeasley

hot_rod said:

A delta P circ sure would be a nice upgrade on a zoned system like that. All the various flow rates would sort themselves out

Can you explain what you mean by "various flow rates would sort themselves out"?

I'm dubious of the "magic" of variable speed pumps. Barba from Taco makes a pretty good case against them, which is strange given Taco makes them! https://www.youtube.com/watch?v=IGBGgGRlhE8&ab_channel=TacoComfortSolutions

Teemok

"Given my needs of ~7 GPM (each zones total GPM added up) and ~10 ft of head loss (3 longest loops in each zone added together)" ?????
Each zone might be close to 7 gpm. (manifold and 1" home runs @ 7 gpm is great) Add them up as though it's morning traffic where all the 1" home runs join together with the 1-1/4" (all zones on at once). 21 gpm for all three is ok in the 1-1/4" freeway. Head is the longest loop only not the 3 longest added. Your description is of a fairly low head system with each zone @ 7-8 gpm max. This is done with either with a single ecm pump and zone valves or a pump sized for each zone. Hydronics people will argue about pumps vs valves all day long.
Imho, three ups15-58s is the common, least expensive, simplistic, 3speed flexible, durable, easy and cheep to repair option. Three ecm's if seeking electrical efficiency crumbs and you filtering for magnetite with a mag filter. A single ecm and zone valves with a mag filter, maybe quieter, more electrically efficient, but more parts, more expensive. Designer makes the call. Read or youtube some Valves vs Pumps stuff.

jeremypbeasley

@Teemok I think you may have misread. Understandable given how ungodly long this thread is now!

Check out this spreadsheet for all my math. https://docs.google.com/spreadsheets/d/1tDbQUl-_PsNaXY287UobXhK4MHdBz1AC6_ZLtPxq8EY/edit?usp=sharing

I decided to go forward with a single pump and 3 Taco zone valves.

I have 3 zones that are 2.X GPM each and 7 GPM when you add all 3 zones together.

I understand what you're saying about head loss being JUST the longest run, not the longest run in each zone added up.

In that case, my pump should be at least 7GPM and ~3 ft max head loss.

I just dropped 4k on PEX, manifolds, zone valves so money is tight. Seems like I could reuse my existing pump for now and upgrade to a 3 speed later, yeah? Here are the specs for the Taco 007 (existing pump) and the Grundfos UPS15-58FC you recommended.

hot_rod

Here is another look at his electronic circulators work

https://www.caleffi.com/sites/default/files/media/external-file/Idronics_16_NA_Circulation%20in%20hydronic%20systems.pdf


But if the technology frightens or concerns you, stick with a PSC circ😉

Teemok

Got it, you already made a call on valves. My only concern is your flow math is for the absolute minimum flow rates need to achieve the output of the radiators you have listed. 0.2gpm is not a good flow rate it's a minimum. A 007 will work well with a single zone calling with flow the flow meters open balanced but when the second and third zones call the loop flow rates will drop because the pump is small for all manifolds at once. No need for a 3 speed pump when it's undersized on high. The longer runs will take longer to heat and some radiators outputs might suffer a bit when all zones are calling at once. You can set your manifold flow meters for the low flows you calc'ed to try to balance what that 007 can move. Otherwise you will have high flows in the short loops and no flow in the others. You will need to choke down the shorter loops a lot and all loops some.
Flows will be what they are and zones will open and close out of sync with each other. Once the spaces start to heat up and satisfy many of the calls for heat will be one zone or two overlapping giving better flow potential. You have calculated 2.0gpm on one zone. If you do choke balance for that low a flow the boiler may tend to cycle when just that zone calls or any one zone calls. What I'm saying is.... you can choke tune to get by with the 007 but it's not ideal. If you go bigger 3 speed or ECM in the future remember to re-balance the manifold flow meters for higher individual loop flow rates that the new pump can provide. I hope you will be warm soon.