Understanding Head Loss at below max capacity

flyingmeatball

Sorry - this post got deleted when I tried to attach a file

I'm trying to understand how to think about circulator pump sizing for radiant heat design and the associated tradeoffs. I understand the theory of what is happening, but I'm trying to get an understanding of what will actually happen in the real world.

I'm working with simple single circuit design as an example - 258' of 1/2" PEX tubing under three different scenarios - a Delta T of 10, a Delta T of 20, and a Delta T of 10 with a higher outside temperature. (full specs of each scenario in the attached PDF)

Going from a Delta T of 20 down to 10 pushes the head loss from 9 to 27.9 (circuit A vs. circuit B in my example). If I'm designing this system, I probably choose a different pump at that point. But then imagine that instead of it being 0 degrees F outside, for a temp differential of 70, it's a more normal day at 40 degrees F outside. Now my temp differential is only 30 - the head loss at a delta T of 10 is only 8.2 - pretty close to the head loss on my cold day with a delta T of 20.

Would you typically pick a pump with a high head loss, knowing that the majority of the time you won't have that in your system, or would you pick a pump that will function more efficiently on a "normal" day, accepting that Delta T may creep up from 10 to 20 when it's really cold outside?

Any insights would be much appreciated!

Zman

Don't get too wrapped up with the delta T. It is a design target that will never be dead on.
When sizing your circs, I would recommend looking only at gpm required to meet your target delta T for your design day. If the delta T fluctuates a bit,don't worry about it

flyingmeatball

Thanks Zman - attached are the circuits as I have them right now. The delta T matters because it impacts how I size my pumps. You can see that manifold 2 has an estimated head loss of 23.2 - that feels pretty big to me. If I don't actually need a delta T of 12, but could instead go with a delta T of say 20, that head loss would drop to 10.2, which is a very different (cheaper) pump. Any input would be much appreciated.

Zman

How is the tubing being run?
In slab? Radiant plates? What spacing?

flyingmeatball

The first floor is poured in concrete (1.5" pour), flooring will be polished concrete or tile. Second floor in above the subfloor with aluminum heat fins under an engineered hardwood floor. Generally a 9" OC spacing. All of this + the insulation/windows are accounted for in the required load, which I comfortable with.

Rich_49

What are the R values under the slab and the upper floors ?

SuperTech

Have you given any thought to the Taco Viridian VT2218 variable Delta T pump? I have been impressed with it's capabilities.

flyingmeatball

I believe 2" of R-10 polystyrene sheets under the concrete pour (there is an unfinished basement below, not a true slab). but I'm at work and can't check now. The upper floors will be something like this:

https://www.radiantec.com/installation-manuals/installing-tubing-within-a-subfloor/

For the sake of discussion, lets assume that my heat loss calcs are correct - how should I go about sizing pumps when the delta T makes such a big difference in the head loss?

flyingmeatball

SuperTech said:

Have you given any thought to the Taco Viridian VT2218 variable Delta T pump? I have been impressed with it's capabilities.

I have not gotten to the selection of an actual pump yet - I'm trying to understand how I should be thinking about the sizing at this point. What makes that a good pump in a situation where delta T of 10 gives me a "high" head loss, but a delta T of 20 is more manageable?

Rich_49

Manageable is a word that has nothing to do with what you are doing except that a circ should be able to handle loads at and below design . What is important is choosing the right circ so that when the system is at low loads your head does not get too low and below the low end performance curve of the circ . When that happens you move more water than you need to , your Delta goes to hell , your boiler sees higher water temps than it would like , short cycling will take place for a longer part of the season and , well , it just makes for a bunch of parts instead of a system . Could you give us the square footage of the rooms as you have them called out on your prior attachment ? For the sake of this exercise we'll assume your loads are correct but I suspect they may be a bit high .

Rich_49

Told you it was RadianTec ZMan

Rich_49

What makes that a good pump for damn near every application is that it monitors supply and return and the Delta between them . When the circ slows down due to a narrowing Delta the heads will follow and vice versa . It has alot of range .

When you design for a 10* Delta you should not experience a higher Delta . A lower Delta may happen but we want to limit that to the shortest amount of time during a heating season

SuperJ

Looks to me like you should consider using 3 instead of 2 circuits in the den, as it's your worst offender (maybe 3x 200' at a tighter spacing). Try and get your system designed so that it is "naturally" pretty balanced so you don't have to rely on power wasting balancing valves to make things work.

If your circuits have widely varying head pressure requirements you're going to have spend more time carefully balancing the system. You basically will have to restrict all the low pressure drop circuits (usually an adjustment on your manifold) to flow at the design flow rate (not over) with the pump running at enough pressure to satisfy the design flow rate of your highest pressure drop circuit. Flow will take the path of least resistance so you need to add resistance to force flow into the higher pressure drop circuits.

Are you running a single pump with valves on each circuit, or a valve on each manifold, or a pump per manifold? It may make sense to pump your high pressure drop manifold separately if it's an extreme outlier.

If it's new construction you could address the heat loss with better insulation and window strategies in the den. If you can knock 2000-3000btu/hr off that area the pressure drop won't be so bad.

flyingmeatball

Rich said:

Manageable is a word that has nothing to do with what you are doing except that a circ should be able to handle loads at and below design . What is important is choosing the right circ so that when the system is at low loads your head does not get too low and below the low end performance curve of the circ . When that happens you move more water than you need to , your Delta goes to hell , your boiler sees higher water temps than it would like , short cycling will take place for a longer part of the season and , well , it just makes for a bunch of parts instead of a system . Could you give us the square footage of the rooms as you have them called out on your prior attachment ? For the sake of this exercise we'll assume your loads are correct but I suspect they may be a bit high .

Thanks Rich - Added the square footage to each zone (manifold). We have high ceilings and significant fenestration (floor ceiling windows in some rooms), which is raising the load.

I can provide the LoopCAD file I'm working on if that would be more helpful.

flyingmeatball

SuperJ said:

Looks to me like you should consider using 3 instead of 2 circuits in the den, as it's your worst offender (maybe 3x 200' at a tighter spacing). Try and get your system designed so that it is "naturally" pretty balanced so you don't have to rely on power wasting balancing valves to make things work.

If your circuits have widely varying head pressure requirements you're going to have spend more time carefully balancing the system. You basically will have to restrict all the low pressure drop circuits (usually an adjustment on your manifold) to flow at the design flow rate (not over) with the pump running at enough pressure to satisfy the design flow rate of your highest pressure drop circuit. Flow will take the path of least resistance so you need to add resistance to force flow into the higher pressure drop circuits.

Are you running a single pump with valves on each circuit, or a valve on each manifold, or a pump per manifold? It may make sense to pump your high pressure drop manifold separately if it's an extreme outlier.

If it's new construction you could address the heat loss with better insulation and window strategies in the den. If you can knock 2000-3000btu/hr off that area the pressure drop won't be so bad.

Thanks SuperJ,

We are just wrapping up the carpentry phase of a gut reno (down to the foundation)/addition, so the fenestration/sq ft is fixed at this point. I was planning on running one pump per manifold.

If I split the den into three circuits, I don't have a good way to split the office, and it's on the same manifold. I would then have 200, 200, 200, 273, which seems worse from a balance perspective.

SuperJ

TLDR version. Like @Zman said, look at the pump curve and make sure you can move the designed flow (all circuit flows added up, at your worst pressure drop)

So you your case: 18.2GPM at 23.2 feet of head.

IMHO, you shouldn't need that much head (in a normal residential application) unless you solving someone else's problem after the fact. (probably stemming from something not being handled properly in the initial design.)

Rich_49

I need the room square footages , not by zone . Can we make that happen ? You can have more than one zone on a manifold .

flyingmeatball

Yep - here are the lists of the rooms with their approx square footages and what manifold they are attached to and the circuit if it is unique to a room

SuperTech

The Viridian VT2218 can handle up to 18 gpm maximum. It is really a great pump for many applications. I would definitely consider it

Rich_49

Check your inbox

SuperJ

> @SuperTech said:
> The Viridian VT2218 can handle up to 18 gpm maximum. It is really a great pump for many applications. I would definitely consider it

The 2218 does either 22fthd @ 0gpm (deadheaded), or 18gpm at 0fthd. Or anything in between if you draw a line between those two points. So it will only work if you can reduce the pressure drop requirement substantially and probably pump each manifold separately to get you back towards the middle of the pump curve. Taco may make a bigger delta P pump though that would be more appropriate.

https://www.hearth.com/talk/attachments/viridian-vt2218-curve-png.172055/

Zman

I see some issues with the calcs you have posted.

Your 3500 sq ft house almost certainly does not have a heat loss of 102,650. Where did that number come from?

I am suspect of the panel outputs and flow rates on the loop calcs. The sheet shows a double in flow rate almost doubling the output. My calc shows only a 10% increase with a doubled flow.

Trying to push more than 1 gpm through a 250' 1/2" loop for no good reason is a waste of energy.

flyingmeatball

I've got the house loaded into LoopCAD - that's where I'm getting the 102,650. I also have had two Manual J's calculated - they came in at 104K and 109K. I'll message those two you privately if you'd like.

The 1gpm was because of the delta T of 10 in a high heat loss area (large windows, unconditioned on 5 sides). That was my original question of how to size the pumps and whether a delta T other than 10 was palatable, or how to design for expected operating conditions, not design conditions.

Rich_49

SuperJ said:

> @SuperTech said:

> The Viridian VT2218 can handle up to 18 gpm maximum. It is really a great pump for many applications. I would definitely consider it



The 2218 does either 22fthd @ 0gpm (deadheaded), or 18gpm at 0fthd. Or anything in between if you draw a line between those two points. So it will only work if you can reduce the pressure drop requirement substantially and probably pump each manifold separately to get you back towards the middle of the pump curve. Taco may make a bigger delta P pump though that would be more appropriate.



https://www.hearth.com/talk/attachments/viridian-vt2218-curve-png.172055/

You do not have to worry about the middle of the pump curve as it is an ECM pump that operates at varying points based on system requirements . middle of the curve is solely for PSC circs .

A look at the 2 depictions on this page shows a shaded area . The VT2218 will operate at infinite points between the high and low end curves anywhere within the shaded area .
http://www.tacocomfort.com/products/variable_speed_products/vt2218/index.html

SuperJ

The design point isn't even close to curve though. The pump doesn't really have much flow capability until the pressure is below 18-16 total head.

I'm not saying the VT2218 is a bad pump, just not a great pump for this job as it stands today. But, if the design is altered, it may be suitable (likely one on each manifold).
Altering the design is probably a good idea, since big variable speed pumps get expensive fast compared to the little guys.

SuperJ

flyingmeatball said:

If I split the den into three circuits, I don't have a good way to split the office, and it's on the same manifold. I would then have 200, 200, 200, 273, which seems worse from a balance perspective.

The long circuit isn't as bad as it looks since the flow rate requirement is lower in the office. The goal would be to get the den circuits down to a similar pressure drop as the office. Balance is more about getting a similar pressure drop across your circuits, than similar lengths or flowrates. A long circuit with a low flowrate may be ok.

At the risk of sounding like a complete hack... you could split a den circuit in 2 right below the manifolds (if you can't get a bigger manifold). You'd want to do this where the connection is exposed though, do a reverse return, and as close to the manifolds as possible to keep the common 1/2" to a minimum length. (Not buried in a wall or floor somewhere).

Rich_49

This whole thread was about someone wanting a bit of insight into head and flows . Not to beat him up because he is making a good , honest attempt at designing his own system . The loop configuration has to be worked through to make it make sense all the time and be part of a system .

flyingmeatball

Thanks everyone for the advice. Really appreciate all the input I've received. I'm feeling pretty good about my design so far. I will probably open a new threat related to the boiler room when i get the circulators/manifolds/valves/boiler/etc all laid out, but for now I've attached the circuits I landed on - moved a few around to reduce my head loss. Feel free to leave any last comments/updates if something looks terribly off. Thanks again!

-Karl

Zman

I think you got it. If the average homeowner (and many pros) put a fraction the effort you have into this, the world would be a more comfortable and efficient place.

Rich_49

That looks better than previous postings . Care to let us in on what circ/s you decided on and how you'll group them with manifolds ?

flyingmeatball

I'm going to group the two manifolds that are in the basement together on one circ, then another for the two upstairs manifolds. That should give me a decent split. The only concern I have is that the three circuits in the den which are low-mass (sandwich style) are on the same manifold/pump as the rest of the circuits that are in the high-mass concrete. I couldn't come up with a better way to do it. Those three circuits will be zoned together with zone caps.

I haven't fully decided on the exact circulators to use, but I'm leaning towards fixed flow grundfos units. I'm worried that if something goes wrong I won't be able to troubleshoot a variable circulator because multiple things are changing.

I'm also leaning towards Watts radiant manifolds with Everhot Pex-B tubing. Is that an acceptable setup? I don't need the best thing ever, but I don't want shoddy quality either. I've already done a ton of due diligence on Pex-A, and just can't find any hard and fast data saying it is conclusively better than the cheaper Pex-B. I suspect there might be some opinions floating around out there on that...

SuperJ

Nice progress on pressure drop and reducing design temperatures.
Most variable speed circulators let you run in a constant low/med/high speed as well as constant/proportional pressure. So troubleshooting shouldn't be any more difficult. If you're zoning out individual circuits I highly recommend ECM circulators.

flyingmeatball

Thanks @SuperJ - one more question to the team, is it kosher to mix radiant wall panels, like a runtal, with in-floor radiant heating? The runtal asks for 180F water, which I'm obviously not going to run - do I basically have to run a whole new line of copper to wherever the panels are going to be or is there a common workaround? My contractor is pushing for radiant panels in the sunroom and the den, where there are 5 sides of heat loss and are relatively remote from the central manifolds. I'm trying to get my head around the fixed costs that would be required to have both radiant panels AND PEX in the floors.

Rich_49

Size the Runtal for the lower water temps you'll use for radiant . By radiant panels , do you mean panel rads ? Runtal is not the only manufacturer of this type of product , they are at the top in price though

SuperJ

I have some Purmo panel radiators in my house. I love the way they heat, seems to be next best thing to infloor heat.
All the manufactures should offer sizing charts for various temperatures, sometimes it's just a table of multipliers, other times they will offer 2 or 4 ratings for each product at different water temperatures and flows. The 180F default sizing is just how the represent their products by default. Showing capacities at high temps makes their products look cheaper and impressive.

Here is a table I used to size (it's Purmo specific, but you should be find the equivalent information for any decent panel rad). (you choose the entering/leaving/room temperatures and the table provides a sizing multiplier to correct for lower/higher temperatures).


You can use a an outdoor reset to reduce the water temperatures as it gets warmer out for both your floors and panel rads. Since you are just supplementing with the rads they probably aren't going to end up being that big. But you'll have to get larger manifolds to feed the panels rads as well as the floor. I have TRV (thermostatic valves) on my rads and love it.

I'm not sure what is more cost effective, to upgrade your infloor or add the panel rads. One nice thing about panel rads is that they are more responsive than infloor heating, they heat up quickly and cool off quickly.

Zman

If the Panel Rads are going to be on the same manifold/circ as a the infloor, it would be a good idea to use a simple circuit setter like Caleffi quick setter to control the flow to the lower resistance radiators.

flyingmeatball

I would have a separate manifold for radiators. Runthal is indeed quite $$$, I can't seem to find Purmo in the USA - know how to find a distributor for them?

SuperJ

I wasn’t necessarily vouching for Purmo. But they work well for me. I’m in Canada. My source imports them directly from Austria or wherever they are made. Price seems decent for what you get. They have an integrated balancing valve as well. There are many similar panel rads. Like Buderus, myson stelrad.

https://www.supplyhouse.com/Buderus-Panel-Radiators-13370000

https://www.homedepot.com/s/myson%20baseboard?searchtype=suggest&NCNI-5

However by the time you add additional manifolds, rads, tubing and labor. I think it may be simpler and more cost effective to not use panel rads and increase your infloor btu/ft2. Tighten spacing and use top grade heat spreaders and a thermally conductive grouting. Maybe increase tubing diameter carefully evaluate floor coverings.