In-Floor Heating System Not Meeting Set-Point Temperature
Hello there,
Going through the first winter in a new house and trying to diagnose an issue with the heating system keeping up – appreciate your advice!
We had a bit of a cold snap and reached about 1 F (-17 C) this week, maybe a bit cold overnight, and I start to notice the system unable to keep the main area warm.
20 year-old wood frame house is approx 2500 sqft excl. garage, has hydronic radiant in-floor on the main level (zone 1, about 800 sqft), hydronic radiant in-floor in the partially finished basement (zone 2), and hydronic baseboards in the upper level (split into zones 3 and 4). The in-floor is supplied via a taco radiant mixing block that has a setpoint of 110F (no outdoor reset), mixed from water heated to 150-160F with a 120kBTU oil boiler. It is embedded in concrete slab.
I am most concerned with zone 1 (main level) – trying to get that to 22 C without success – seems to stall at 18 C or 18.5 C. I should note that we’re quite comfortable at 18 C but I worry that it’s indicative of a design problem or losses.
I’ve had a look at other folks with this issue – I believe I’ve eliminated the following possibilities:
a) that the thermostat is the cause: as it seems to be keeping the zone valve on as far as I can tell (though I haven’t tried shorting the terminals for an extended duration I have not observed it turning off and every time I go to check the taco supply is close to setpoint).
b) that there is insufficient GPM: I used a thermocouple on the copper pipe for supply/return and it indicated a 10F drop (95 F return at 105 F supply) after the system had been running for several hours. I gather that this is a reasonable temp drop that would indicate there is reasonable flow.
c) leakage from slab to outside: my thermal camera does not show any observable hot spots on the exterior siding
Some things I think *might* be contributing to this:
a) Floor covering: The ceramic floor was covered with wood by previous owners (but no rugs/carpet) which may be insulating preventing some conductive heat transfer. Note that the floor temperature seems to be 22-23 C according to IR spot thermometer so I would imagine there is still sufficient transfer going on.
b) That there is insufficient zone isolation: we are leaving the basement unheated at the moment and it’s settling at around 14 C. The part of the basement that is finished has insulation on top, however there is an approx 300 sqft storage/utility room below the main area where the ceiling is uninsulated and I can feel a bit of heat loss to that area from bottom of main level. Door is also open to basement permanently at the moment. The upper level zones (bedrooms) are accessed via open staircase and we also leave at pretty low setpoints (14 C) – doors mostly closed though.
c) The main area has lots of windows: There’s about 200 sqft of windows, about half of that south facing which can heat the place up nicely on a sunny day but perhaps a source of losses? Double glazed but can feel the convective loops near them at night when window coverings off.
d) There is air infiltration: either through oven exhaust vent, wood stove chimney (can hear a slight whistle) or other.
e) insufficient loop area: there are only 3 supply lines running to the 800 sqft main area (zone 1), perhaps there should be more?
Any thoughts on this?
As far as solutions go – since we’re comfortable I don’t know if we need to solve it I just want to make sure there aren’t any major sources of losses that could be resolved.
I imagine we could raise BTU output of the floor with outdoor reset to bring water temp up but there are a couple issues there 1) I see the wood floor already a bit warped/split (perhaps from when this was tried before?) and 2) I plan to replace oil boiler with heat pump water heater and it would not be efficient to run that at higher temps so trying to avoid that path.
Thanks very much!
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
I have a Taco RMB that heats our radiant floor system here in Zone 6A, it has outdoor reset with a range of 75 to 95F. Have you tried turning down the water temperature?
I would also insulate under the floor in order to have heat flow upwards. During our remodel — where we installed the Warmboard, we used CCF to seal the air leaks and reduce the heat loss. Generally you want R20 under a radiant floor.
Wit all due respect William , turning the temp down will do nothing but make the issue worse . The outdoor reset should have been used to vary SWTs throughout the heating season .
Thanks for your reply William! I'm also in Zone 6A
Hrm I had never thought to actually turn the water temp down, would that not just decrease BTU output and make it struggle more?
I guess it would also reduce the burden on the variable injection pump in taco RMB but it doesn't seem to fail to hit the current 110F setpoint so I'm not sure that's the issue.
I will definitely give the temp reduction thing a try overnight - currently I think the sunshine is influencing the temperatures. Unfortunately (?) it's starting to warm up outside a bit so my test conditions are gone =(
I think we'll definitely insulate under floor in the unfinished basement area at some point for sure.
Turning down the water temperature is a diagnostic procedure. Sometimes thermostats are misplaced , so they cut off heat flow when the thermostat reaches the set point, where a nearby room has not yet reached temperature.
Sometimes relocating a thermostat solves the problem.
That is a good diagnostic tool to have, thank you - I also have an interest in getting the system functional using lower water temps due to impending heat pump installation.
In this case thermostat did not appear to ever reach the setpoint and as far as I could see was not turning off the zone valve prematurely due to any malfunction, though I noted that I didn't properly eliminate that possibility by manually enabling the zone valves - I may try that in the next cold snap
Thermostat is out of the sunshine but is in an area that would be influenced by the floor heat so it is a possibility I will keep in mind when further tuning - might get a couple more thermometers placed in the area to make sure temperature is consistent and see how lower water temps influence that consistency in general as an experiment (or start out with lower water temps when heat pump installed)
I think the floor covering is likely the issue. You're right around design temp and the floor is coming up just short. You could try a combination of the following:
1. Try to reduce heat loss a tad using air sealing.
2. increase flow (right now average water temperature is 100, so doubling flow should bring that up a couple degrees). If the set point is 22 C, you'll need higher than 22 C at the floor. This might be extremely easy or might involve a new circulator.
3. slightly increase water temperature to in-floor, maybe to 115. That's still within the range of an air/ground-to-water heat pump and shouldn't affect the floors.
4. Heat the basement zone, which shouldn't take much additional energy if it's at 14 C unheated. Similar to insulating the basement ceiling, just easier.
If the loops are about 800 feet, there are big enough and that's not really a correctable mistake at this point. If none of the above work, you could add a different low temperature emitter to the main floor to help. A panel radiator or low temperature baseboard would work. Last option would be to reduce heat loss via better windows.
Three things to look at: increasing the output of the floor, getting heat from other sources, and reducing the demand for heat.
Two things determine the heat output of a radiant floor: the area, and the surface temperature. You're locked into your floor area now so really the only dial to turn is the surface temperature, and the only adjustment that seems to make sense is the water temperature.
Sometimes the thing to do is to add more heat elsewhere, in the form of radiators or conductors, either in the same room or a neighboring room.
The other side of the equation is the demand for heat. You can reduce heat loss by better insulation and air sealing.
The floor temperature of 71-73 F is considerably lower than the 95 F water return temperature. Without knowing the details of where the tubing is and how the floor is constructed I don't know why the difference is so big. Insulation underneath might or might not help that, depending on what is limited that, but it might help a lot.
I would also check on air sealing, e.g., getting a blower door test. When it's harder to heat lower levels, that can be a symptom of cold air leaking in at lower levels and hot air leaving at upper levels. You want to seal both ends of that leakage path.
+1 on Charlie's comments. This is especially true with the unconditioned basement below. Cold air leaking into the basement and warm air going out the attic will travel through the main floor, cooling it. Seal air leaks on all levels of the house. This is often some of the easiest and cheapest energy upgrades. Plus, air sealing and insulation lower your heating demand, which saves money for the life of the house. Increasing water temperature or adding radiation may work, but also increases heating costs.
Thanks so much for all the tips -
Hm it seems like I should really try to figure out why the floor surface temperature is so different from the return. Just going by the attached photo of the floor (with my hand for reference as I couldn't get the scale to show up), do any issues stand out? If it's a slab (as I have assumed) should the heat be more distributed throughout the floor?
I'll note that we did have a blower door test done last fall which came in at 4.44 ACH or
equivalent leakage area of 278 square inches. Since then I sealed up some obvious places, e.g., there was some venting in the basement and old unused dryer vent unsealed, so my hope is that brought the figure down. I also upgraded the kitchen extractor fan exhaust system to something a bit more heavy duty - my hope is that it brought the number down a bit but tough to know until I get the follow-up audit next fall - I will make sure attic hatch is airtight.
To get heat to flow you need to have a temperature difference. In this system you've got four surfaces -- the pipe that is at about 38C (average of inlet and outlet), the surface of the floor that is at 23C, the room air that is at 18C and the outside air that is at -17C. And you've got three interfaces: pipe to floor, floor to room, room to outside. At a steady state the heat flow through each of those interfaces is going to be the same.
The simple model that is usually used in home heating is that heat flow is directly proportional to temperature difference. So right now you've got a temperature difference between the room and the outdoors of 35C. If you want the room to be at 22C instead of 18C you need to increase that difference to 39C, which means the amount of heat flowing into the room needs to increase by 39/35= 11.4%.
Right now the floor is at 23C and the room is at 18C, a difference of 5C. To get 11.4% more heat that difference needs to increase by 11.4%, or to 5.6C. If the interior is at 22C that means the floor needs to be at 27.6C (or 82F).
Right now the floor is at 23C and the water is at 40C, a difference of 17C. To get 11.4% more heat that difference needs to increase to 18.9C. Also your floor needs to be at 27.6 so your water needs to be at 46.5 (or 116F).
In terms of what you can change, the room to outside interface is dictated by your insulation level and construction, you might be able to improve that a little. If the limiting factor is the floor temperature that your floor finish can withstand that's a hard limit, there's no way to get more heat out of the floor without making it warmer. The 27C/82F I calculate as the desired temperature is not hot at all, it would get warmer than that when the sun shines on it in the summer. Even the water temperature of 46.5C/116F is within the range of normal for building materials, a house could get that hot in the summer if it was closed up with no AC on a sunny day.
So try setting the water at 120F send, 110F return.
Important to note that a temp of 120 is compatible with air to water heat pumps, whether directly or paired with an electric resistance supplemental element for the coldest hours.
Hi Paul & DCContrarian
DCContrarian - thanks your way to estimate the heat flow between interfaces is really helpful, I thought it would be more complicated. I guess that this depends on the interface areas being similar? I suppose the floor area and wall area are pretty close.
I will try the 120F setting to see if it helps, however further adding to my confusion is the fact that it's currently a balmy -6 deg C outside and it's still failing to bring the main area above 18.5 deg C so I suspect something else might be wrong as it seems that inside/outside temp differential should allow the current 110F supply temp to do the job.
Thanks for your note about the air to water heat pump Paul - yes the unit I will install can supply at 120F, its CoP suffers at this output temp so ideally I'd run it lower but it's good to know that this might be an option if the higher temps are indeed the solution.
Correct, COP will be terrible at 120 (and above using backup). An outdoor reset will help you here.
"Thanks your way to estimate the heat flow between interfaces is really helpful, I thought it would be more complicated. I guess that this depends on the interface areas being similar? I suppose the floor area and wall area are pretty close."
Since we're going off of observed behavior it doesn't matter what the surfaces are made of. The observed behavior tells us what's actually going on in your house.
Ok, let's plug -6C into my numbers from post #8.
If you think of the whole system as an assembly from pipe at 40C to outdoors at -17C the temperature drop over each component will be equal to its share of the thermal resistance. So if the floor is at 23C the pipe/floor interface is 30% of the total resistance. Similarly the air at 18 means the inside air/floor interface is 9% and the inside air/outside air interface is 61%.
In the simple model those percentages should be constant. So if it's -6C outside and the water in the pipe is still at 40C, the total drop is 46C. if 30% of that is between floor and pipe the floor should be at 26.2C, and if 61% of that is between inside air and outside air the inside air should be at 22.2C.
So you should double-check what's happening with floor temperatures and water temperatures, because you should be seeing higher floor temperatures and indoor air temperatures. And you're definitely not seeing higher air temperatures which makes it unlikely you're seeing higher floor temperatures either.
The most likely reason is that your water temperature is dropping. Which means something is throttling it back.
A 10 degree drop over the zone seems a little low. What type of pump do you have on that zone? Or do you have one pump delivering to all the zones? Obviously on very cold days the zones are working overtime and more heat may be being delivered to the other zones depending on how your piping arrangement is whereby the heat is taking the least path of resistance. With that minimal temperature difference, the water may be moving to fast through the radiant zone, so even if you increase the temperature it still won't have time to transfer the heat to the floor. A simple test would be to close the return line valve slightly to slow the water in that zone and see what happens.....or if you have a pump with variable speeds, adjust that.
A 10 degree delta T is preferred for floors - you don’t want a large gradient on a surface you walk across. Radiators would be different. The flow rate should stay the same or be increased to increase the average water temperature. The transfer is from tube to concrete to floor, so the higher the tube wall temperature is, the higher the rest will be.
Paul, I believe he states the temp. drop in the return water is 10 degrees, not the delta t of the floor, so that would be 10 deg/ 800 ft^2 or 0.0125 deg/ft^2, not very much of a gradient, which is why it is taking so long for the floor to heat up. We don't know the flow rate so hard to figure that into the numbers at this time, but that minimal loss equates to minimal transfer....
Correct 10 degree delta t supply - return. Floors are advised to keep about that delta T so one part of the floor (loop end) isn’t noticeably colder than another, balanced against pump energy use. I’m not sure what your coefficient means. In no way will exclusively raising delta T (cutting flow rate thereby decreasing surface temp) increase floor heat output. Increasing flow rate will slightly increase surface temp, but not as much as increasing supply temp by 10 degrees and keeping flow rate the same.
Hi Paul & Tom
It's using the taco radiant mixing block for that zone (that is also supplying the other infloor zones garage/basement but they are currently not in use)
I believe the pump has a constant speed but secondary variable injection pump (my layman's understanding of how it achieves desired supply water temp) - I will keep your note about adjusting ball valve on the return to control the deltaT as something to try just to see what happens.
Per the manual I believe the mixing block has the ability to regulate DeltaT however I don't think it is set up for this as there is no temperature sensor on the return connected to the unit.
BTW it's possible my measurement is not the most accurate (using thermocouple on side of copper pipe for supply/return) it was just to give an idea.
Paul, one of the heat rate equations is q = -h A (T2 -T1) which shows that both delta T and flow, derived from the heat transfer coefficient h, effect the heat rate q. Understanding how h works is crucial to good heat transfer, gain or removal.....
Proper installation of a radiant zone, eg. layout, is what keeps the floor evenly hot....You can see in his picture how they must have ran a continuous back and forth loop and how it cools as it get closer to the wall. If they laid it out starting at the outside working inwards you would see a different pattern.
I suspect a previous owner may have cranked the flow rate up to try and get the average temperature hotter.
DCC, I think that’s also probably right. You’re correct there’s some to gain there, but not much vs. raising the temp.
Tom, increasing the fluid Delta T will not increase surface temp independent of raising the supply temp, so that equation doesn’t apply here. We went to increase the average water temperature in the loops. By your logic, a delta T of 40 would have twice the output of delta T of 20, which would be 2x delta 10, despite lowering AWT from 100 to 95 to 85.
See Paul, you don't understand h. T1 and T2 are not input and output, they are the temperatures of surfaces one and two . If you increase the tubing temperature, you will increase the temperature difference thus increasing q, the heat transfer rate, in which time is a factor and where h comes in when convection is introduced .....Conductivity of the the materials and how tightly packed together they are also contribute...That equation is q = -k A (T2 - T1)...both equations equally apply....
We’re talking different things. Yes, you’re right that a larger difference between floor and ambient air temp will lead to more heat output. I’ve been saying that all along. But the poster said he had a 10 Delta T between the supply and return tubing, which is just right for this application. You suggested increasing that Delta T, not the surface Delta T.
Paul, I think we are just misunderstanding each other slightly. I am not saying to increase the delta T in and out, since we cannot do that, that is up to the system installed and how it operates. If we could maintain a constant water temperature in the loop that would be wonderful and easy for calculations, but the delta T in and out is indicative to how much heat is being transferred or taken from the water. If we make adjustments to perform better, more heat will be transferred resulting in a higher delta T in and out. So I am simply saying, the bigger drop in water temp., in and out, is relative to how much heat is being sent into the floor and room. If the temp. came back the same, then it's not losing any heat...indicating bad conduction and/or convection.
Yes you're entirely right! Just a misunderstanding. We can increase flow and decrease delta T which would in fact slightly increase output by doing exactly what you recommend: increasing the Delta T of the surface.
In my cottage, with in-floor sensors, I need to set the thermostats to 25-26 celcius in order for the rooms temperature to be at 21-22 degrees. And water temperature leaving the electric boiler is around 140 F and coming back at around 120 F. All my loops are below 220 feet.
Thanks Stephane! Is that typical for when the outside temps are -15 to -20 C?
it is a brand new system. I live in Quebec and we are getting -25C to -30C at night these days and the system does not have any issue to keep room temperature at 21-22C. The house is very air tight and built using ICF all the way to the roof, 3-pane windows so maybe it helps :)
I'll bet you are glad to have that! Yes I am in the maritimes, we didn't get quite as cold as you outside =)
I suspect additional air sealing to improve the 4.4 ACH rating will help a lot, I am guessing your rating is much lower than this. Not in the budget to replace the windows unfortunately haha
If the original floor covering was changed to a surface with a higher R value ( tile < hardwood -wood ) you'll need to increase the SWT ( supply water temp) . Do not worry about damaging the floor because of the radiant , 100% of hardwood issues are , and were caused by moisture , PERIOD . The maximum water temp recommendations have nothing to do with damaging hardwood as much as creating uncomfortable conditions for the occupants .
Use the outdoor reset and program the RMB properly or at a minimum raise the temp through the programming capabilities of the RMB . If the radiant zone is in 1 area , once the surface temp has reached "X" the thermostat will satisfy and shut off flow to that zone . Once flow is ceased there is no chance of "overheating" the floor .
Thanks Richard that is really good to know re: wood flooring ratings. The floor is a bit messed here (laid over uneven tile) so I suspect I'll replace it at some point down the line and it is good to have this fact in mind.
Enabling the outdoor reset is a good idea I suspect I will eventually do this to increase the comfort once I sort out the issue (as it's still struggling to bring up heat even with outdoor temperature warmer so something else must be awry)
Alrighty some further investigation reveals some new information:
-It seems as though the bottom OSB surface of the heated floor is considerably warmer (about +3 deg C) than the upper side - it is 24-26 C on the bottom vs 21-23 C on the top. I would guess that the bamboo then has higher R value than the OSB and/or radiant plumbing is closer to the bottom surface, so it's heating downwards.
-After running the radiant for a while trying to get the temp up today, in the unfinished portion of the basement the air temp is about 18 deg C at head height, I had noted 14 C however that it is in the finished/insulated area on the thermostat but the two areas are quite different. I believe this difference in temp is due to waste heat from the furnace as well as heat from the underside of the heated floor . The exposed foundation walls are sitting at about 10-12 C (and lower in above ground portion) and some harvested by the heat pump water heater.
This unfinished area represents a significant portion of the footprint, maybe about half of the main floor area (400 sqft) so I suspect I'm losing a bunch of energy through these interfaces (bottom of floor to unfinished basement and unfinished basement to foundation walls)
So if I were to better isolate the unfinished area of the basement by insulating the unfinished basement ceilings I suppose I would eliminate a bunch of losses - and I am thinking this may also increase the slab temperature overall and therefore the surface temp of the bamboo flooring and help to solve the issue without increasing water temps -
...if this is correct, the only thing I can't quite figure out is that, in the area above the *finished* portion the floor temps aren't noticeably higher at all despite that area being insulated and plumbed line density appears to be the same - perhaps since there's nothing really impeding airflow to the unfinished portion through the joist bays?
If this is correct, thoughts on how to best insulate? Other posts seem to indicate just using batts is a good way to go if foam not an option... that would be ideal as unfinished portion has a lot of utilities run.
Insulating the bottom isn't going to send more heat up. Heat flow is determined by temperature delta, and insulating the bottom doesn't make the top any hotter. In theory reducing heat loss makes your average water temperature higher, but your temperature drop is so small already -- only 10F -- that there's only so much you can gain, maybe a degree or so, which isn't going to be significant.
Hrm my thinking was that insulating that bottom portion increases the temperature that the slab stabilizes at in equilibrium since (we reduce the losses to the basement/underside or effectively eliminate the interface), and that would bring the avg temperature of the whole slab (and hence surface on the top) up - is that not the case or not significant?
Apologies I think you addressed that point in a different way by saying that the temp drop might be reduced by an insignificant amount by insulating
I really like DCContrarian's simplified thermal resistor model, because it makes it easy to get some rough numbers for things and get a better understanding that way. But it's important to recognize that it's a simplified model, and when we start wondering about what goes on inside one piece of that system--the floor, we need a better model. And when we consider it more carefully, we see that insulation underneath can help, more than just for reducing the 10 F drop over the length of the tubing.
As you can see in your thermal image, the lateral heat spreading is imperfect. Right above the tube, the floor is hotter than halfway between them. That situation can be improved by insulation underneath. If we think of it in thermal resistance terms again, there's some amount of heat flowing laterally out of the tubing and through the concrete, and there's some thermal resistance that flow goes through. The temperature drop from the tube to halfway between two tubes is R*Q, where R is the thermal resistance and Q is the heat flow. Now, about half the heat flow then turns upward and flows into the room, and the other half turns downward and flows into the basement. If you added insulation, that would remove the heat flow downward, so Q would be roughly cut in half. So the temperature drop from tube to midpoint between tubes would be cut in half. So the variation in temperature over the surface would be cut in half--it would be closer to the higher temperature you see right above the tubing. With that higher average surface temperature, you get more heat flow into the room.
You would not double the heat flow into the room--the simplified model is useful for helping understand that. But you would increase it, without increasing water temperature, and thus you'd make the system more compatible with a future heat pump.
There may be some control issue as well, but the insulation will help more than the oversimplified model would indicate.
Thanks Charlie for expanding on this - I agree that the thermal resistance model is very helpful for the big picture
Your explanation supports a bit of what I would hope for, i.e., that heat spread more evenly across the slab.
I would plan to insulate regardless of this issue to eliminate the loss into that area (it seems problematic to insulate the foundation walls themselves so the plan would be to keep that unfinished utility/storage room unconditioned)
Heat flows from hot to cool. Insulation doesn't "bend" it, it just makes it flow more slowly. Adding insulation just makes less heat flow where you added the insulation, it doesn't make more heat flow elsewhere.
DCContrarian,
There are two ways to really prove my point. One would be to draw the more complex 2D network of thermal resistors approximating this network, and show how changing the the ones reprenenting the under-floor insulation affects the heat flow in others. The other is to do a 2D finite-element simulation in THERM or similar. If you don't get if from the simpler intuitive explanation, I might not be able to convince you without going to those lengths.
But I'll try once more to explain it intuitively, using your framing:
" Adding insulation just makes less heat flow where you added the insulation,"
Here's how using that framework leads to better heat delivery to the room above:
1. Part of that "less flow" from adding the under-floor insulation is reduced lateral heat flow in the floor material.
2. Reduced lateral heat flow means less lateral temperature drop.
3. Less lateral temperature drop means a higher average floor temperature.
4. A higher average floor temperature means more heat delivered to the room above.
If you still disagree, which of those four steps do you think is incorrect?
I think we're mostly agreeing*, I think where we disagree is the magnitude of the effect. Let's say we have a slab with water entering at 105F and leaving at 95F. Let's also stipulate that the slab is perfectly symmetrical, with 50% of the heat going up and 50% down, and that we can insulate it perfectly so that no heat now goes down. What happens? Half as much** heat flows out of the slab, so half as much flows in at steady state. The incoming water is still going to be at 105F, but with half as much heat flowing out the outgoing water will be at 100F. So the average water temperature goes from 100F to 102.5F.
If you look at my posts #8 and #30, the ultimate heat flow out of the floor --and in the whole system -- is determined by the temperature difference between the water and the outside air. So with an outdoor temperature of 1F the difference between an average water temperature of 100F and 102.5F is 2.5%. Plugging 102.5F into the formula from post #30 I get a floor temperature about 1.75F higher and room temperature 1.5F higher.
A few observations:
1. This is assuming 100% insulation, which you never see in reality.
2. The same result could be obtained simply by raising the water temperature 2.5 degrees.
3. The reason to insulate the backside of radiant systems is efficiency, it keeps heat from going where it's not needed. It's not going to overcome a defective design.
* You got me going with "about half the heat flow then turns upward and flows into the room, and the other half turns downward and flows into the basement" which isn't at all how it works, but I now realize that's a useful approximation.
** An astute reader may note that the heat flow increases so it isn't exactly half. But it's around 2.5%, so half is an adequate first approximation.
There are two effects that allow the insulation underneath to enhance the heat output from the floor surface. One is the reduced temperature drop of the water as it flows through. I agree with your analysis of that effect. The other is the effect of the 2D heat flow, laterally out from the tubing through the concrete, and then, yes, turning 90 degrees and flowing up into the air, or turning 90 degrees and flowing down into the basement. That 90 degree turn happens with or without the insulation underneath. That more complex behavior is harder to analyze with ultra simple models. The insulation doesn't cause the 90 degree turn. It just removes some of the heat flow in the concrete, with is a shared path between the heat flows up and down. If you actually want to understand that, I can sketch a diagram of the resistor network.
Charlie --
Please do, because I'm not seeing it.
Might want to post at the bottom of the thread because we're tapped out on replies here.
Any type insulation will do just for the sake of creating resistance to downward losses and to force the heat upward where we want it .
Heat does not rise !
All radiant whether it be walls , ceiling or floor requires insulation on the opposite side to create the resistance . Do not be concerned with shiny surfaces on the insulation or any other nonsense , all you need is R Value ( resistance) .
The heat flow from the top is determined by the surface temperature of the top. Already the water temperature drop is low -- only 10F -- so reducing heat losses to the bottom isn't going to change the temperature of the top of the floor appreciably, so it won't affect the heat flow from the top.
Yeah I suppose the insulation can't hurt to reduce losses, but I see the point that all it can do is increase the return temps (and hence average temp) which isn't significant for just a 10F drop as you say DCContrarian
One possibility is that I am not measuring supply/return temps properly - the mixing block changes the supply temp quite rapidly when it decides to inject so I might be a bit off- if it was indeed 20F then perhaps reducing those losses in the slab would be a more significant factor, i.e., avg fluid temp would be influenced more
I see bamboo flooring and heated floors and my alarm bells go off.
Bamboo has pretty bad dimensional stability, add in dryer winter air plus heat and you are asking for trouble. I've had no end of issues with my floating heated bamboo floor, the only way I've managed to keep it together is by keeping the air in the house around 50%RH, even then I get some areas where there are pretty big gaps in the winter.
If you are increasing your slab temperature, keep an eye on your flooring. If it is a floating floor, once the click joint comes apart, about the only way to get it back together is by taking it up and relaying it. If it is glued/nailed down this is less of an issue, mostly just an eye sore.
Also from your thermal picture I see a fair bit of striping and pretty large spacing between pipes. If this is a staple up, you need some better heat spreaders. If it is embedded in concrete, not much you can do. Be mindful that some areas might run much hotter than other spots which could make the flooring issue worse.
Hi Akos - that probably explains the issues with the floor as it is split in some parts (and soft spots in others where I believe surface it was laid over was uneven - floor joists are fine) - in a way it is comforting to know you've had issues as well but I empathize haha. In any case we plan to eventually replace the floor with something more suitable for radiant but for now just dealing with it... exposing the tile underneath might solve all my woes but it is way uglier than the slightly wonky bamboo =)
Yeah I was thinking as well the heat doesn't appear to spread as much as I would have thought - I am 99% sure it's not a staple up - I have not cut anything open to verify there is a slab but it's all OSB on the bottom and knocking on it anywhere it definitely feels like there's a solid mass of concrete on the other side of the board - I can't imagine what else it could be... joists spaced at 12".
Perhaps the floor had something like 1/2" Hardiebacker board on it slotted for the tubing to reside in then the tile was laid on top. It would seem pretty concrete like especially with the joists on 12" centers.
I'll investigate this by cutting a hole soon but I'm thinking that it would still sound a wee bit hollow when knocking if there was hardieboard - this feels like I'm knocking on the foundation walls when I tap on the OSB.
I believe it's also further evidenced by the long thermal time constant that I've now realized
Hi all -
Thanks for the very informative responses here
So I woke up this morning to the main area at its setpoint of 22C with an outside temp of -7C... floor temps on main level sitting at 25-27C (with underside still a bit warmer at 27-29C)
It would seem that I underestimated thermal inertia of the slab were a bit off - I knew it should take a lot longer than a forced air system to bring the room temps up a few deg but didn't think it would be on the order of 24h (up to now we usually keep it at 15 overnight and 17-18C through the day with thermostat set to raise temp about 3h before waking, and always seem to wake up to it at the setpoint)
Probably TMI but I think what happened was that on Sun-Mon the temps were milder (reaching a bit above zero), but we were also running gravely low on oil so had the woodstove running in an attempt to conserve usage in case we could not get a delivery prior to the impending cold snap. The wood stove has no issue bringing up temp of the room, so I think on those mild days we were barely triggering the floor heating at all and the slab may have "cooled" to what I imagine was an average of room above/basement temperature (we keep room above at 17-18C and basement is 14-15 so say slab settled at 16-17 degC, though I hadn't measured at that point)
The cold snap occurred quite quickly bringing temps to -15 to -20 and we got the oil delivered so weren't shy about using the boiler - I suspect at that point it was a combination of the slab starting off relatively cool and the energy flux to outside that made it seem to take way longer than usual to heat up. It warmed up a bit outside last night (to -1 or -2 C) and I suppose at that point the system was able to catch up in a shorter time frame.
We actually added quite a bit of blown-in insulation recently bringing attic to R60, and did a bunch of air sealing as I said - while we don't have a previous winter to compare against, I wonder if perhaps the fact that we still use the stove occasionally and get a lot of heat from the sun that brings it above the in-floor thermostat setpoint and now stays put longer than it did previously (due to efficiency improvements) has brought the system out of whack by allowing slab temp to drop.. in that case perhaps we can solve this to some degree by using a thermostat that supports temperature monitoring on the slab and keeps it from dropping below a certain level.
So anyhow, some lessons learned from a former "forced-air heater person"
a) Don't underestimate the thermal inertia of the slab.. patience
b) Good way to estimate heat flux is by considering temperature differences between all interfaces
c) Bamboo floors have poor dimensional stability, evidenced by my floor and Akos', and also seems to be better insulator than OSB just based on fact surface temp on bottom of slab hotter than top
d) Can use supply/return temps of radiant loop to diagnose whether losses/transfer are reasonable/expected, and 10F is a good target for radiant floor consistency, but can alter by changing flow rates (by adjusting variable speed pump or close return ball valve a bit)
e) Insulating the bottom of the floor (facing a cold area) may reduce this temp difference and improve uniformity of slab temp which may help with BTU output from floor to living area - to be confirmed with FEA or empirically
f) Increased water temp should raise floor surface temp and hence BTU output/floor area. Outdoor reset might be helpful to allow for increased water/floor temp (and hence reduced slab heating time) during cold snap.
g) Air-water minisplits can handle 120F but CoP suffers, better to design for lower water temp
h) Air infiltration from basement rising to 2nd floor may contribute to heat losses and should continue to investigate air sealing improvements to help with this
Hmm what else did I miss
I very much appreciate all the help and discussion - I'll try and follow up if I discover anything useful for folks who might encounter similar issues
Your last comment says alot . With radiant especially a slab if it is in fact a slab . You do not want to use setbacks at all . These systems run best at steady state .Especially if you turn them down at night and that night is very cold , it takes an immense amount of energy to heat that mass back up and then for it to contribute to the room it is heating . This is also why ODR should be used , it keeps the mass more in line with the required temps for that span of time .
These issues are why I and several others have started really suggesting radiant ceilings . Mass is small , R value of drywall and other materials is low and furniture , carpets and the like have no effect on output .
Hot always goes to cold regardless of whether it is above or below the space , think about the sun . Is the top always facing Earth ? You must not forget that radiant heat does not so much heat the air as maintain a mean radiant temperature , ie; outside walls , furniture , glass . If they are at a maintained MRT the odds that you'll be cold within them disappears .
That's a lot of good stuff you learned! I think all of your points are correct. The only one I'd modify is the ball valve--closing that off a bit can be an interesting experiment but I don't see any reason to leave it anything but wide open, as you'd get worse uniformity over the tubing run, and you are decreasing the flow without decreasing pump power much if any.
And I agree with Rich that radiant ceilings are a great approach that should be done more!
You're not decreasing flow, you are manipulating pressures and velocities.....
I want to print out and frame your post.
For years it was considered best practice to put radiant flooring in a massive slab, the bigger the better. "Thermal mass" was considered necessary for the proper operation. Your experience shows what should have been apparent to anyone who looked at the science -- comfort is maximized when your heating system is designed to be as responsive as possible, and responsiveness is maximized by minimizing heat capacity. You want that sucker to heat up as soon as the thermostat clicks on and turn off when the thermostat clicks off.
Longtime readers of this forum know I have something of a bee in my bonnet about the term "thermal mass." It's not a term that is used in science or engineering, but the real reason is that when someone uses that term it's an immediate red flag to me that the person does not have a grip on the basic science of heating and cooling.
OK, rant over. Glad your problem is solved and we could be of assistance in solving it.
Yeah for sure, this has been great discussion for me and I have a few ways to help improve/tune the system if I need to (I expect I'll need to when I get that heat pump installed)
Just to make sure I don't have a misconception here - I understood and use the term thermal mass/inertia as just a shorthand for referring to time constant associated with the slab heating up given same source (water temp) and load (outside temp). Regardless of whether a large time constant is good design or not it's still a major factor of heating/cooling systems isn't it?
The discussion about high responsiveness and comfort (especially in radiant ceilings) is an interesting one - this 24 hour response time of my system is a bit much haha. How about individually controlled graphene panels that track your movement and are manufactured to emit IR wavelength tuned to skin-tone for max absorption!
"Regardless of whether a large time constant is good design or not it's still a major factor of heating/cooling systems isn't it?"
Interestingly it's not a factor in engineering and design at all. Manual J is the primary tool for calculating heating and cooling loads and its formulas don't consider heat capacity at all.
It is implicit in the Manual J calculations though that buildings will have some heat capacity. Heating and cooling systems are sized to maintain temperature between the first and 99th percentile of temperature. There are about 88 hours a year when the temperature is above that range, and 88 hours a year when it's below. The assumption is that those "excursions" will be short -- an hour or two at a time -- and the building will have enough heat capacity that the indoor temperature won't get too uncomfortable, even though the HVAC isn't keeping up.
Deleted
And the name for that property is "heat capacity."
"Mass" is the property of matter that causes it to resist acceleration when a force is applied. Isaac Newton coined the term and that's the definition he gave. Using the term to describe any other property of matter is incorrect.
Deleted
What else did you miss? You forgot to snatch the pebble from my hand....
I want to highlight something I wrote in post #51:
If you look at my posts #8 and #30, the ultimate heat flow out of the floor --and in the whole system -- is determined by the temperature difference between the water and the outside air.
I want to highlight this because it's something that I had never realized until I wrote those posts, I've never seen it anywhere else. It's non-intuitive, and it's useful for understanding the performance of radiant systems -- which for some reason are really hard for people to grasp.
Yup....T1 and T2....you can choose any two anywhere within the system....Simple.
For DCContrarian, here's a super approximate 2D lumped model for 1 sq foot of floor, with a tube on the left edge of the region modeled and the right edge of the region modeled halfway between two tubes. The values are thermal resistances, not R-values. The vertical ones each represent 1/3 of a square foot, so there thermal resistance is 3X the R-value.
The result is 18% more heat flow to the room above when R-10 insulation is added below. That's not a real number, since I don't know the materials well enough to put in accurate values, and dividing it into three sections isn't fine enough to be accurate, but that's enough to show you what effect I'm talking about. And that's neglecting the difference in water return temperature.
This is the simplified way I think about this setup.
Because of the large tube spacing and thin concrete slab, the sideways heat transfer along the slab is limiting. This is why you see the pronounced striping in the thermal image. By insulating and reducing the losses to the basement, this thin slab will run a bit warmer and deliver more heat to the living space.
If this was a better heated floor setup with proper spacing, this sideways heat transfer would not be the limiting factor and insulating the basement would not matter.
P.S. To the OP. With an uninsulated basement ceiling and the basement is this cold, you have some serious air leaks you need to deal with. My uninsulated basement with 18" thick stone walls stays much closer to the house temperature just from the losses of the hydronic plumbing. Before insulating the ceiling, fix these air leaks, it will make the house warmer and will make a pretty big difference on your energy use especially if heating with oil.
Yeah I reckon the system isn't super well designed, I am OK with that now that I see it works OK and is comfy. Incidentally the now-finished former garage (now a bonus room, ~400sqft with 10ft ceilings) has only one loop of in-floor and is slab-on-grade and I had no luck getting any kind of temperature response at all after 24h in that room (I tried when it was milder outside). I've not tried leaving it on for multiple days though, maybe it would eventually be OK. Previous owners had put a small mini-split which heats the area just fine though, I just mention this point to indicate that it may not have been a well conceived design when house was built (circa 2000 but in rural area with no building code mandate). In the main area you can also see some gaps in the loop coverage, etc.
Just to clarify re: the basement, once the system had been running a while the unfinished portion of the basement got to about 19-20 C with the upper level at 22 C - that would be from waste heat from boiler and underside of floor I expect. There is a heat pump water heater also cooling that space periodically. It definitely needs some air sealing down there but maybe not that bad (sill plate is quite uniform).
The finished portion which I originally quoted basement temp based on was the one that was fairly cool (14-15 C) - the walls separating finished/unfinished portion are insulated and ceiling also insulated from radiant above, so no direct heat source in finished portion except an open door between finished/unfinished portion and any flow through insulated interfaces - I would think fairly low BTU input to that area which to me explains the lower temps.
OK, I see that's a scenario where adding insulation does increase the top surface temperature. But I think it's unrealistic because in your model the bottom air temperature doesn't change. So basically the bottom is modeled as an infinite perfect heat sink. In your example I see the lower surface going from 30 to 24 when the insulation is added, with no change in the lower air temperature.
Now, I like to model the earth as an infinite heat sink at constant temperature so maybe this is a good model for slab on grade. But if below is a basement or crawl space that's not a good model, there's going to be a significant drop in temperature when the source of heat for the space is reduced. If you take that same resistor model and instead of keeping the bottom temperature constant, you reduce it, the middle of slab and top of slab temperatures don't go up as much or at all.
You are correct, if the goal is to model what happens in the basement. That wasn't my goal. And as you put more and more insulation there, what happens in the basement becomes less relevant.
Thats great, I wasn't sure how to set up the equivalent circuit for this situation. I'm more comfortable with circuits so it helps a lot.
So this is just for one of the three loops and representing flow for a small section of the run e.g. 1 sqft right? In that case do these labels make sense for your diagram?
V1 - water temp provided by source
V2 - basement room temp
V3 - heated room temp
R3,R5,R8 - floor covering on top (ceramic with 3/4" bamboo overlay)
*Corrected to R3,R5,R8 - resistance in transfer from top surface to room
R2,R6,R7 - upper part of slab
*Corrected to: R2,R6,R7 - upper part of slab and floor covering
R1,R4 - lateral spread through slab in that 1 sqft region
R11,R12,R13 - lower part of slab plus insulation
*Corrected to: R11,R12,R13 - lower part of slab plus insulation and OSB covering
R9, R10, R14 - bottom slab covering (OSB)
*Corrected to: R9, R10, R14 - resistance in transfer from bottom surface to room
So to restate what I think you've said above, changing the basement temp would increase flow through R1, R4 and drop avg temp on those labeled nodes and hence affect overall output to the V3 sink but if increasing insulation that basement temp has less and less of an impact on overall flow through R1, R4 and would make those temps more uniform
Your labels are mostly right, but R2, 6 and 7 are inclusive of the upper part of the slab, and R3, 5 and 8 are the heat transfer from the surface of the floor into the room. Similarly, 11,12 and 13 are the lower part of the slab, the OSB, and the insulation, all the materials on the bottom. And 9, 10, 14 are the heat transfer between the bottom surface and the basement.
Lowering basement temperature would increase flow through R1 and R4, but I think the more important conclusion is that adding insulation means less flow through R and R4, which means more uniform temperature across the top surface, and a higher average temperature across the top surface. And that one you have a good amount of insulation there, the basement temperature doesn't have much effect.
Thanks Charlie I've made those changes in my Post 66 in case anyone re-reading misses this correction
Perhaps I keep missing it, but what is your in floor tube diameter and what specific pump is supplying the main run in the 800sf area? Obviously a drop of 10*F at a .3 gpm flow indicate half the heat flow of a 10*F drop in a .6gpm flow. Thanks, Hugh
I will admit I have not read all 74 posts I noticed you mentioned a wood stove in your first post.
I was wondering how often you use the stove?
Seems to me if the stove gets used often its heat could shut down the boiler before the boiler can get the slab warm enough to heat the home.
Seems to me the boiler needs 72 hours of uninterrupted run time before you can say it can’t make the set point.
Walta
You win the gold star. Just read post #45 where he explains he had turned off the boiler and was using the wood stove, and it took a day or two after turning the boiler on for the slab to get up to temperature. You're exactly right on both accounts.
While the concrete in the floor certainly doesn't respond like air in a forced air system, I think there's a tendency to over imagine the heat needed to increase its temperature. If the system uses an 1 1/2" concrete slab here's the quick and dirty math on changing its temperature.
Assumptions and calculations
concrete - 150lb per cubic foot
concrete's specific heat .2btus/lb
floor area considered 800sf
mass of concrete - 800 X (150/ 8) = 15,000lb of concrete
15,000lb X .2btu/lb = btus to warm whole mass 1*F = 3,000btu
So to warm the mass up 10*F only requires the addition of 30,000 btus from the 150,000 BTU boiler.
Will this create a lag? It certainly will but not one that would take days to overcome.
YMMV,
Hugh
That number, 30 kBTU/10F rise, provides useful perspective. But it's a little hard to estimate how much it matters. First, the rate that the system allows heating the slab is not the full 150 kBTU/h of the boiler. We don't know the flow rate in the tubing, but suppose it was 2 gpm. At a 10 F ΔT in the water, that's 10 kBTU maximum water delivery rate. And if the slab was at 70 F, or even lower given the cold basement below, while the wood stove was on, it might need to rise 25 F before the floor is putting out its maximum heat heat. And as the floor starts to get warmer than the room, the heat delivered from the water is split between warming the slab and warming the room. So it could take on the order of 8-12 hours to warm the slab. Then the slab needs to warm the room, and there's more thermal mass there.
So while the numbers you provide are a useful contribution to understanding it, people shouldn't conclude from them that the heat capacity of the slab would only delay things by tens of minutes.