Best option to fix a Radiant floor that's cold on one side?

Roobul · January 18

Hello,

My first floor is all radiant; 6 zones, 11 loops 1/2 PEX ranging from 50' to 300' in a 1-1/2" slab with spray foam underneath and tile on top. Largest zone has 5 loops. The first half of a loop the floor is warm, but the second half is cool. The source temp is 110* and return is only 75-80* at the manifold, so 30-35* Delta T. IBC 20-160G3 mod/con, primary/secondary loops separated by Califfi SEP4 LLH, radiant manifold fed with Taco 0015e3 pump.

I've tried the different pump settings, currently on 3 (full speed, non-variable). With several loops open (or the 5-loop zone) it just doesn't heat the downstream portion of the floor. From what I can tell from poking around here, the GPM per 1/2" loop should be 0.6, times 11 loops is 6.6 GPM (worst case, all loops open), with head pressure of the longest loop (300') at 4 feet, plus a foot for manifolds and fittings, 5 feet. 6.6 GPM @ 5 FT should easily be handled by the 0015e3 on Low setting (constant 5 ft pressure up to 12 GPM). According to the performance curve chart, it should handle it at any of the 3 settings).

I'm tempted to try a Delta T pump (like the Taco VT2218) which will push harder until the delta T set point (say, 17*), but I don't understand why the 0015e3 isn't getting it done.

Do I need to raise the source temperature?

Any advice would be appreciated. Thanks!

EdTheHeaterMan · January 18

Does the manifold have flow meters? Does the manifold have adjustable valve heads? Unless you can adjust the flow and see the flow rate, you are only guessing.

I believe that your 50 ft loop is getting more than its fair share of the pump's flow. The water is going to take the path of least resistance. That would be the shorter loop(s) leaving less GPM for all the longer loops.

Is this a newly commissioned system?
OR
Is this a new problem on a previously working system?

hot_rod · January 18

A pic of the manifold may help to see what you can do to balance the loops.

Roobul · January 18

Ed, the manifold does have flow meters with adjusters. Focusing on the 5-loop zone, the loops are all 300', I tweaked the adjusters to get the same flow into each loop (just under 1 GPM). The 50' loop heats the small bath quickly, no problem there--it doesn't stay open very long. Installed two years now. It works, I'm just trying to fine tune it, get the floors more evenly heated, squeeze some efficiency out of it. The contractor didn't do much in the way of design calculations, didn't install the outdoor reset sensor, put the same pumps on boiler and all zones, .... But the copper piping in the boiler looks nice!

The attached FLIR picture shows the middle of the living room where one loop ends and the next one begins. You can see the 8-9* temp difference (and feel it in your feet if you straddle the line). I wish they did a helix layout, but too late now. Would like to get it more even.

Roobul · January 18

Bob, it's a Blue Fin manifold, like the one in this picture. It's not so much that one loop is hotter than another, it's that each loop cools down too much by the end of that loop. They all start off nice and warm and end up 10* cooler 10' away. The delta T at the manifold is quite high, 30-35* so I'm guessing I need either more flow or higher injection temp...?

Roobul · January 18

Roobul said:
I tweaked the adjusters to get the same flow into each loop (just under 1 GPM).

Ed, I misspoke - the indicators are just under the first line on the scale, which is 0.2 GPM (not 1 GPM). So it appears I have 1 GPM feeding 5 loops in total. I don't get why that's so low with the pump on #3 if the theoretical head is 5 Ft.

hot_rod · January 18

If there was a design done, it would indicate the delta T you would see at steady state condition. How long did it run before you took the pic. Most residential systems design around a 10- 15 delta from one end of the loop to the other. At some point you should get there, maybe even tighter just before that zone satisfies. A bit of window in that room, any idea on the heat load?

If just that one loop behaves like that, lower the flow to the other loops in the zone. If all on that manifold fall behind and it had an adequate purge, look for a flow restriction. Any Y strainers in the piping?
Is this manifold temperature controlled by a thermostatic mixing valve? Or direct from the Sep? All the flowsetters on the manifold wide open? All the white caps on the manifold removed, or full open

You should have plenty of circulator for 5- 300' loops.

mattmia2 · January 18

Maybe you need to set up the ODR and get it dialed in so it runs long cycles at lower temps, maybe it will be more even if you make the cycles longer so it can all heat up instead of part of it heating, satisfying the t-stat, shutting down and it all cooling off.

Roobul · January 19

hot_rod said:
If there was a design done...

Bob, there wasn't much design. It was more a matter of how many loops do we need to cover the floor, and which areas should operate together on the same t-stat. No calcs on heat loss, windows, ceiling heights, heat load, required BTUs, etc. More like "you need 11 loops to cover the square footage and a 20-160 MBTU boiler will be good in this medium size house." Having said that, the house does stay warm, the boiler usually condenses, no issues with DHW. I don't think it's way off, but needs some tweaking and maybe a larger variable speed circ pump. A 10-15* would be great and should make the floor heat evenly. Heat had been on for 4 or 5 weeks when the pic was taken.

Maybe I'm overreacting to the color variance on the FLIR image, but you can feel the difference if you straddle that line btwn the two loops. I thought about reversing the flow so the left side heats right to left, and the two hotter ends of the loops would be adjacent, cooling towards the perimeters.

Adequate purge is a possible question, as the pump is not whisper quiet; it has a slight rattle noise. No Y strainers, but there is a thermostatic mixing valve before the pump. For the 5-loop zone, the loops seemed to get slightly less flow as they start further from the source end of the manifold (i.e. 1 was more than 2, 2 was more than 3, etc.) so I tightened the flowsetters in 1/4 turn increments, from right to left for those 5 loops (1 = 1 turn, 2 = 3/4, 3 = 1/2, 4 = 1/4, 5 = 0), which made them flow evenly. I didn't pop the actuators off, but measured once they were fully opened, figuring that was real world conditions.

Do mixing valves restrict the flow significantly? That would explain it...

Roobul · January 19

mattmia2 said:
Maybe you need to set up the ODR and get it dialed in...

I did install the ODR sensor on the north wall of the boiler room and the boiler is recognizing the outside temp. Oddly, it shows about 4-8* higher than a thermometer--I can't imagine that could be warmed thru the wall (5/8 sheetrock, 3" CC spray foam, 1/2 sheathing, 1/2 styrofoam, siding), or that heat is travelling to the sensor thru the wires, but that's another story. Without the ODR, it assumes 32*F, and since I put it in we've been above and below that, so it's has altered the water temp a little. The lower temp for longer cycles is interesting, though the boiler often modulates down to 20 MBTU so it may cause it to cycle more...?

GroundUp · January 19

It sounds like a balancing issue more than anything. The hot loops (likely the shorter ones) need to be throttled back using the flowmeters on the supply manifold and the cooler loops (likely the longer ones) will need to be opened up further or even wide open. Keep adjusting until the return from each loop is the same temperature (this will require a separate thermometer of some sort- whether it's just a steady hand or your FLIR). Once that's been established, you should have the issue mostly if not completely taken care of.

Roobul · January 19

GroundUp said:
It sounds like a balancing issue more than anything.

The loop in the left side of the FLIR picture above is farthest from the start of the manifold so its flowsetter is backed all the way out (full flow). I tried closing off the other 4 loops so it was the only one running for several hours, and the temp at the end of that loop did increase (as did the flow). So it does seem that if I could increase the flow to all those 300' loops it would resolve the issue. When only one loop is open, the indicator does drop farther down, but it doesn't get pinned to the bottom (seems like less than maximum flow). I wonder if that also indicates a flow restriction (air or mixing valve constraint)?

mattmia2 · January 19

What is the delta t on the short loops? If it is too low then that will cause the boiler to cycle.

Why is there a thermostatic mixing valve with a mod con? You should just be supplying the temp you need out of the mod con. Is it a high limit only for when the boiler is fired hotter for dhw?

Roobul · January 19

mattmia2 said:
What is the delta t on the short loops?

If only one or two short loops are open, the delta T is a bit less, but they seem to satisfy the t-stat without too much cycling. Yes, the mixing valve is to prevent high temp from going to radiant when it follows the DHW or another high heat (non radiant) zone. The mixing valve is set to 125* whereas the target temp for radiant is ~110*. In this case, the boiler does cycle because it has maybe 140* supply temp when the target drops to 110*, so it shuts off for a few minutes until the return temp comes down.

hot_rod · January 19

Yeah the Infrared pics can vary, especially when shooting at an angle. Stand right in front of the loops, now that you know where they are.

Do mixing valves restrict the flow significantly? That would explain it...

Some mixing valves have checks in the H and C port, and usually a small strainer to protect the check. The strainer is sometimes built into the sealing washer, so it cannot be left out, easily.

If it is an ASSE 1070, or a dual listed 1017/ 1070 mix valve it must have checks.
See if maybe that is causing a restriction. You don't need or want the checks or strainers for hydronic use.

A 3/4 or 1" thermostatic will be in the 2.7- 3 Cv range, so 5 gpm is about all you want to try and flow through them before you get into a higher head circulator requirement.

If all 11 loops are from one mixing valve?
11 X .65 gpm = 7.15 gpm

7.15 gpm through a 3 Cv valve , shown below.

Is this the circulator? If so speed 3, on the 0015 is just getting you there. What brand and model mix valve do you have?

Roobul · January 19

hot_rod said:
If it is an ASSE 1070, or a dual listed 1017/ 1070 mix valve it must have checks.

The mixing valve is a Taco 5000-3 ASSE 1017 on 1" piping, and it feeds the one manifold for all 11 loops. Same as the one on DHW. Specs show max 20 GPM, Cv 2.3, check valves included (I don't know if they're installed).

Taco 0015e3 circ performance curves attached

Roobul · January 19

That Figure 7-5a is exactly what I have (the one with the "X" drawn thru it!)
The Taco 5120 is a high flow mixing valve with Cv 4.1 and GPM 27.2 -- maybe that would help?

mattmia2 · January 19

Does the boiler have multiple DH calls so it can have a low temp call for the radiant?

Roobul · January 19

mattmia2 said:
Does the boiler have multiple DH calls

I'm afraid I don't understand the question. The DHW is a zone on the boiler with a special profile specifically for DHW, which I assume has a specific ramp up, delta T and target temp just for DHW. There are two other heating zones on the boiler that use high heat: an air handler in the garage, a CI radiator in the mudroom. Either of these 3 will get the water up to 140-160* (max temp is set to 180), and when they shut off, it's common for the Radiant to kick in and the mixing valve temporarily caps it at 125* until the supply temp drops. I don't know if a temporary injection of 160* would damage the floors before dropping to 125 in a couple minutes, but the mixing valve prevents this.

mattmia2 · January 19

So it the boiler has separate calls for each zone instead of a zone controller handing the multiple zones and sending a single domestic heating call to the boiler? What I'm getting at is do the low temp and high temp zones have a way to make different calls to the boiler to supply different temps?

Roobul · January 19

mattmia2 said:
do the low temp and high temp zones have a way to make different calls to the boiler?

Ah, yes, the IBC boiler has a built-in 4-zone controller and each zone is configured for its load. 1=DHW, 2=CI Rad, 3=Air Handler, 4=Radiant (the radiant manifold actuators are controlled by an external Taco zone controller, but it in turn tells the boiler that radiant is being called for on #4). So when zones 1-3 stop, the water in the primary is 140-180, too hot for radiant, so the boiler drops to 20 MBTUs, and eventually shuts off until the return temp drops below 110. It's during this ~5 minute transition that very hot water would be sent to the radiant without a mixing valve.

hot_rod · January 19

Roobul said:

hot_rod said:
If it is an ASSE 1070, or a dual listed 1017/ 1070 mix valve it must have checks.
The mixing valve is a Taco 5000-3 ASSE 1017 on 1" piping, and it feeds the one manifold for all 11 loops. Same as the one on DHW. Specs show max 20 GPM, Cv 2.3, check valves included (I don't know if they're installed).

Taco 0015e3 circ performance curves attached

Once again assuming you are wanting .65 gpm per loop 7.1 gpm total

Here is what a 2.3 Cv mixing valve pressure drop will be. 22' head.

Unless you wire that pump to a 240V circuit, I doubt it will get the flow you need?

hot_rod · January 19

Roobul said:
That Figure 7-5a is exactly what I have (the one with the "X" drawn thru it!)
The Taco 5120 is a high flow mixing valve with Cv 4.1 and GPM 27.2 -- maybe that would help?

Give that manufacturer a call tell them you bought that 5120 valve, 4.1 Cv for a radiant system that needs 27 gpm. Ask them what circulator you need.

Here is what I come up with.
27 gpm, 4.1 Cv valve

Even on a DHW system, lets say you have 60 psi in the building, your taking a 43 psi drop across the valve, when flowing 27 gpm now you got 17 psi left for you DHW system.

Design engineers typical like to stay under a 10 psid

Comes a point when a thermostatic mixing valve is the wrong answer. Even with a high Cv 2", it will not mix accurately when you have small gpm loads.
A ball, vane, globe, or butterfly type motorized valve is a better option.

That is why you see hi-low thermostatic mixing stations on large DHW requirement applications. A small mixer, low Cv, paralleled with the high Cv valve.

Roobul · January 20

hot_rod said:

Give that manufacturer a call

The 27 GPM was the valve rating. I only need 7.1 GPM. Using the formula it appears 7.1 GPM and Cv 4.1 would be 3 ft — is that right? If so, assuming the radiant head is 5, the 0015e3 would pump 10 GPM @ 8 ft.

The system in my house is @ 18 psi.

hot_rod · January 20

Ignore the system pressure, it has nothing to do with the pump calculation. The static pressure is what fill the water to the highest level in the building. 12 psi should be plenty, but 18 is fine.

To flow 7 gpm through a 2.3 cv mixing valve, the pump needs to be sized for 7 gpm at 22’ of head.

The pump you showed above is capable of 18’

Here is the calculation result again.

Do you currently have the 2.3 installed as you mentioned above, or the 4.1 Cv?

Roobul · January 20

hot_rod said:
Do you currently have the 2.3 installed as you mentioned above, or the 4.1 Cv?

I have the 2.3 Cv (5000-3) valve installed, so if I understand this correctly, the pump is fighting 9.5 ft of pressure just to get through the mix valve, and then ~5 ft for the radiant loops, and at 14.5 ft I'm only getting 5.5 GPM flow on the High (3) setting. I don't know what I'd get on Medium (2), which shows a constant P of 10 ft, but the pump speed and flow do seem to increase when I switch it from 2 to 3 (does this further imply I have more than 10 ft head?)

If I replaced the 2.3 Cv (5000-3) with the 4.3 Cv (5120) valve, would that mean the pump would be fighting only 3 ft at the valve plus ~5 ft for the radiant loops, for a total of 8 ft, and get 8.5 GPM on Medium (2) and 10 GPM on High (3)?

If so, then it sounds like you've identified the problem as the mix valve and not the pump.

Roobul · January 20

The Caleffi Series 5231 MixCal+ shows a Cv of 7.6 which would impose only 0.87 ft of head at 7.1 GPM, plus the ~5 ft -- if I'm getting this right, that would let me run the pump on Low (1).

hot_rod · January 20

Roobul said:

hot_rod said:
Do you currently have the 2.3 installed as you mentioned above, or the 4.1 Cv?
I have the 2.3 Cv (5000-3) valve installed, so if I understand this correctly, the pump is fighting 9.5 ft of pressure just to get through the mix valve, and then ~5 ft for the radiant loops, and at 14.5 ft I'm only getting 5.5 GPM flow on the High (3) setting. I don't know what I'd get on Medium (2), which shows a constant P of 10 ft, but the pump speed and flow do seem to increase when I switch it from 2 to 3 (does this further imply I have more than 10 ft head?)

If I replaced the 2.3 Cv (5000-3) with the 4.3 Cv (5120) valve, would that mean the pump would be fighting only 3 ft at the valve plus ~5 ft for the radiant loops, for a total of 8 ft, and get 8.5 GPM on Medium (2) and 10 GPM on High (3)?

If so, then it sounds like you've identified the problem as the mix valve and not the pump.

Nope, still off on your numbers.
Flowing 7 gpm through that 2.3 valve is 22’ of head, see the attachment
I don’t see how you are coming up with 9.5’?

The very left of the pump curve is called shut off head, where the circ is no longer able tomove any flow
18’ on the left axis, follow it down to the bottom line to arrive at zero

That pump cannot over come the pressure drop of the 2.3 cv valve at a 7 gpm requirement. Bottom line in your system, you are moving a smaller flow through the loops, maybe .20 gpm per loop?
Which may be why the loop in the pic is a lower temperature.

So move over to 5 gpm on the bottom axis, run straight up to just under 16’. So as you pump less gpm the pressure drop decreases. Does that make sense?

Ya need a higher cv valve or higher head circ. IF you need or want 7 gpm.

I can send the Excel spread sheet that does those calculations for you so you don’t have to do the math or use the formula.

Cv is a flow coefficient applied to any/ all devices. Check valves, ball valves, y strainers, air, dirt seperators zone valves, on and on. It is not specific to three way thermostatic valves.

Look up a 3/4 ball valve or check valve, for example. It may have a cv if 13 Plug 13 into the known cv, plug in whatever flow number you are trying to achieve. You get the answer in head, and psi.

Roobul · January 20

Oh, I was using the formula for PSI, not feet. My bad. (7.1 GPM / 2.3 Cv)**2 = 9.5 psi = 22 feet.

So let’s say I’m getting 0.2 GPM per loop x 11 loops = (2.2 GPM / 2.3 Cv)**2 = 0.91 psi = 2.1 feet for valve. Add 5 for the loops. At 7.1 feet the pump would produce 10.5 GPM.

However, if I’m getting 0.45 gpm per loop x 11 = (5 / 2.3)**2 = 4.72 psi = 10.9 ft + 5 ft for loops = 15.9 ft, at which the pump gives 5 gpm (a match). So at 5 gpm across 11 loops I’m getting about 0.45 gpm per loop which kinda fits with the rooms maintaining temperature but the floors being uneven.

With the 4.3 Cv valve, I’d get the 7.1 GPM at 6.2 + 5 = 11.2 ft, which works on the performance curve chart for High (3).

With Medium (2) the pump varies speed to maintain 10 ft pressure. If the valve and loops have 11.2 ft, what does the pump produce at that “10 feet” mode? (The curve only shows 10’ straight across)

Thank you for your patience and walking me thru this! It’s quite interesting.

hot_rod · January 20

Unless the loops are the same length, or within a few feet of one another, that weak loop could be longer, so it is getting less flow, moving less heat. Or it has a partial kink on one of the end loops?

What you don’t know or aren't telling us

is how many gpm you actually need. The .65 I used is the recommendation from
the RPA quick calculator, RadPad for a 1/2" loop

A room with a high load, lots of cold glass for example, may need .75 or even 1 gpm per loop to heat adequately.
It some cases radiant floor heat alone may not be adequate. That is why the heat load number is so important to getting the expected result.

For this troubleshoot, keep the pump in speed 3 to do your data gathering.

If you have ruled out any flow restrictions, plugged strainers, etc, etc.

My answer is the same, a bigger pump, or a higher CV valve.

Don’t over do the valve. A large valve with a high CV, 17 for example, would need a minimum of 5 gpm, always, to control accurately, you cannot provide that in this small system.

If you want to stay with a thermostatic mixing valve, A valve with a 4-5 Cv should get you over the hump with the pump you have.

The Caleffi 5231 in 1" is a 4.4.

So the "arbitrary" 7 gpm

through this 4.4 valve, under 6' of head across the valve.

Use the attached spreadsheet to avoid math errors, 3rd column

This issue may clear up some of the concepts and technology around circulation and circulators.

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

Roobul · January 20

The loops on the 5-loop zone are pretty close in length, and I do see the same heat fading at the ends of the others.

What I don’t know is how many GPM I need 🙂 As noted earlier, there was little design before installation, pretty much “the boiler will make plenty of DHW and we’ll cover the floors with loops at 8 inches.” If I get to where some are too hot I can throttle them at the manifold, but not enough heat is… well here we are.

I like the higher cv valve. A 4.4 should give me enough flow to get the heat across the ends of the loops.

Thanks again for your input (and the spreadsheet). It’s been a good crash course but it’s starting to make sense.

EdTheHeaterMan · January 20

OK it is time for this one:

Best option to fix a Radiant floor that's cold on one side?

Move to the warmer side of the room!

Roobul · January 23

hot_rod said:

A valve with a 4-5 Cv should get you over the hump with the pump you have.
The Caleffi 5231 in 1" is a 4.4

It looks like the 5231 is actually 4.4 GPM and 7 Cv, but that should work even better.

JohnFX · January 27

According to Uponor Design Manual, to push 0.6gpm in 300 feet of 1/2" PEX at 110f requires approx 20.4 foot pump head. Add to that your other losses and you probably need 22-25 foot head total. The 0015e will not do that.

Daveinscranton · January 27

This is a brute force/not terribly elegant thought. Maybe remove your cold floor loops completely from the manifold. Get another (separate) manifold, with a dedicated 3 speed circulator, wire the pump and controls in a way that makes sense and be done with it. The idea being that your problematic loops of pex are on their own manifold and pump.

The control of that pump you can make as complicated or as simple as you like.

I am probably misunderstanding the problem.

hot_rod · January 28

JohnFX said:
According to Uponor Design Manual, to push 0.6gpm in 300 feet of 1/2" PEX at 110f requires approx 20.4 foot pump head. Add to that your other losses and you probably need 22-25 foot head total. The 0015e will not do that.

Are you sure about that pressure drop? I see .01606 as the multiplier for 1/2” Pex 120F so 4.8’

JohnFX · January 28

MY Bad! 4.84'

Zman · February 1

What are the measured supply and return temps on your loops? A temp drop is normal, it would be ideal if the temp drop is 10 degrees or less.
The tubing is only part of the puzzle with system resistance. The mixing valve, manifold balancers, check valves, and even the length of straight pipe going into the circulator all play into the puzzle.

4Johnpipe · February 1

Pictures of the near boiler piping are needed. If the heat is being left in the room thats not a bd thing.

Issues can only be flow, insulation below or air.

SKYPAINTER · February 1

Is the pump installed on the "mix" side of the 3-way or is it pumping "into" the valve? Always want the pump on the mix side, pumping away from 3-way.

propmanage · February 1

If you close off the hot loops do the cooler loops heat up. Are you sure you have no air?

Best option to fix a Radiant floor that's cold on one side?

Comments

Best option to fix a Radiant floor that's cold on one side?

Categories