Click here to Find a Contractor in your area.
In fairness to all, we don't discuss pricing on the Wall. Thanks for your cooperation.
Need to contact us? Visit

Hydronic Radiant Heat - Retrofit Help Please

We were flooded by Sandy and are in the midst of repairs/remodeling.


The room in question is on slab.  About six years ago, we had hydronic radiant heat installed for an oak floor.  The contractor installed the pex tubing (1/2") between sleepers that consisted of 2x4 lumber.  After the pex was installed, the cavities between the sleepers (where the pex is) were filled with a cement mix.  With the installation of a wood floor, the contractor nailed ¾ plywood to the sleepers and the oak floor was nailed into the plywood.  We lived with it for 6 years.  The heat did not work very well in that room, too much wood to get through.  The adjoining room (tiled kitchen with radiant) was far more comfortable.


Long story short, flood ruined wood floor.  We have now removed the oak floor and plywood, exposing the sleepers and concrete in which the pex tubing is encased.  We are going with porcelain tile instead of another oak floor.


Here lies the question.  If my tile guy pours 1.5 inches of mud on top of the existing base (which is the sleepers, mud and pex) will the heat make it to my tile floor?  I am concerned because the pex is already in 3.75 +/- inches of concrete.  Is adding another 1.5” plus the thickness of the tile going to make the system less effective (or not effective at all)??  The reason for going with tile instead of another wood floor is to get better heat in that room.


Any input would be appreciated!  The photo attached displays the floor as is, the tile guy would pour the mud on top of this floor.


  • IronmanIronman Member Posts: 3,421
    More Info Needed

    Was this installed on top of an existing slab?

    How much insulation is there beneath the slab? How much around the perimeter?

    What water temp is going to the slab when at design temp (coldest day outside)? How is the water temp controlled?

    It appears that there's a lot of glass in that room. Do you know what the heat loss is for it? This would have required a load calc.

    Without the above info, it's not possible to determine if the radiant is sufficient to heat the room.

    As far as adding more "mud" goes, that would not prevent the radiant from heating the room; it would only add more mass which would cause the response time to be longer for heating up and cooling off. Having tile, would reduce the "R" value of the floor covering which would cause more heat to make it into the room. A wood floor would have a higher "R" value and resist heat transfer from the pipes to the room greater than tile.
    Bob Boan

    You can choose to do what you want, but you cannot choose the consequences.
  • GordyGordy Member Posts: 7,104
    edited January 2013
    What ironman said

    It depends on how the math works out.

    How many runs of pex are in each sleeper bay?

    How deep was the pex in the pour? From your post it sounds like the bottom.

    I assume your tile guy wants the thick mud bed to make up for the 1 1/2" of flooring that existed before.

    So in essence your trading about r 1.5 for an r .6 or so depending on how well packed his mud bed will be.

    You gain in the aspect of a more conducting surface with less r value than wood.
  • rsc92rsc92 Member Posts: 50
    edited January 2013
    Answers to your questions

    Was this installed on top of an existing slab? - Yes, it was and the pex tubing was secured directly to the slab so it is at the bottom of the pour.  As such, the pex will be approximately 5.5 inches below the tile.  There is one round trip of pex in each bay/cavity.  The entire room is serviced by a single loop (which I have now learned is not ideal).

    How much insulation is there beneath the slab? How much around the perimeter?  I dont know the answer to either of these.  The slab was pre-existing.

    What water temp is going to the slab when at design temp (coldest day outside)? How is the water temp controlled?  The water temperature is controlled by a mixing valve before the manifold.  The water temperature going into the manifold is 125 degrees.  I can adjust this up or down..

    It appears that there's a lot of glass in that room. Do you know what the heat loss is for it? This would have required a load calc.  You are correct.  The room faces southwest and the entire wall is windows (approx 20 feet).

    My plumber just suggested that we abandon the existing pex and run two loops on top of what is there now.  He said the tile guy will embed the new pex in his 1.5 inches of mud.  This sounds great, but comes at a significant expense.  Any opinion as to whether or not the benefit will outweigh the cost????

    Thanks for all of your help.

  • GordyGordy Member Posts: 7,104
    edited January 2013
    Existing concrete slab

    I,gues start off is the system operational in that room now? if so is the floor warm, or keeping room to temp? if yes then the mud bed will not add much rvalue compared to the wood floor.

    If the slab was existing the system was built on I would say the slab is not insulated.

    Are you using a modulating condensing boiler or conventional type?

    Is your kitchen floor on the same type of sleeper setup?

    If it were mine I would do loops on top of what is there now. With 1/2" xps insulation. Do 8" centers on the tubing. 1" mud bed, and tile.

    Or if tile installer wants a thicker mud bed go with fan fold 1/4" and then 1 1/4 mud bed.

    There are actually a lot of combinations to do a sleeper over the top of what is there, and get insulation in there. This will help response time in a huge way, and lower supply temperature so long as the room load will,allow it. Tile is installed all the time on 1/4" cement board.

    Another combination would be 3/4" xps insulation 3/8" pex with 1/2" plywood sleepers screwed to existing 2x4's 1/4" hardibacker, and tile. Forgot to include a vapor barrier over existing slab unless there was one installed in the previous sleeper system.
  • rsc92rsc92 Member Posts: 50
    More info

    System is operational and maintaining room temperature. In fact it is working much better than it did when the oak floor was in place.

    We have a conventional gas fired boiler.

    The kitchen is over a crawl space. In that installation the pex is secured to the sub-floor. There is about 1.25 inches of mud and then the installer used thinset to secure the tile. Like I said earlier, that works great. I should note that the subfloor is not insulated in the kitchen and it still works fine.

    My installer wants 1.5 inches of mud, but we don't. We would like the height of this floor to match the adjoining floors. We would like the total height of the mud and tile to be 1.5 inches. With that said, I don't think we will have room for insulation and a new pex run. Any thoughts?

    I dont have a vapor barrier. How important is that with tile vs wood? I know we should have had it with the wood floor and that is probably why we had some minor cupping almost immediately after the floor was installed and the radiant was turned on.
  • GordyGordy Member Posts: 7,104
    edited January 2013
    Vapor barrier

    Is very cheap moisture control to the room. With wood a bigger deal. With tile not so much.

    Being that the system is working now may not be an indicator that it will heat the space at design temps though.

    If the tile installer wants 1 1/2" it's his call to give you a finish product he will warranty.

    What type of tile is going in? Thickness?

    Conclusion is this if you want to stay with what you have then all you can do is increase water supply temps. The tile verses the original 1 1/2" of wood will improve things, but may not perform well at design temps which means you will have to increase supply temp. About 140 is max - 150

    There are products that will get you insulation tubing, and tile in that 1 1/2 space, but the cost will be higher most definitely. Your installer needs to be familiar with them though.

    Keep in mind the goal is to have a radiant system that is efficient in both heating the space, and operating costs. If you are running higher water temps, and it takes a longer period of time to heat the space that's not efficient, and harder to control. That's what is in the minds of the people giving advice here.
  • MikeGMikeG Member Posts: 138
    Remove the old

    If you are considering installing new pex loops in mud, why not remove the old mud and tube.  You could at least install 1 1/2" extruded foam and vapor barrier in the bays between the sleepers. This gets you some insulation on the slab.  Or take it all up and put the foam down, with a mud pour to the required floor height.   Do your pex, mud and tile over that to the required thickness.  This would also remove some excess weight from the floor, think structural here.  Are we talking gypcrete or regular cement type mud.  The tile could be set into thinset.  If you are thinking of using any of the cement board products, check the manufatures specs on required thinkness for floor use.  I know they use 1/4" but I wouldn't unless the everything was solid underneath.  A slab is pretty good but it also has to do with how even it is.  Some some thoughts.   Mike
  • rsc92rsc92 Member Posts: 50

    Thanks for the input.  It is very much appreciated. 
  • RobGRobG Member Posts: 1,850
    More BTU's

    What about adding a couple of loops of 3/8 pex on top of the existing slab, doing an over-pour then tile. Don't abandon the existing system, just supplement it. You would then have two layers of tubing heating that room.

    Just a thought.

  • GordyGordy Member Posts: 7,104

    Thought about that. Which is the supplemental bottom layer in my opinion. My opinion was if your going to do another round of pex may as well try to get it insulated with quicker response. Seems the tile guy wants his 1 1/2" of mud though.
  • ChrisChris Member Posts: 3,056
    edited January 2013
    Come on Guys

    This isn't rocket science. Guy needs to do a heat loss period. He is just adding more mass and he is lowering the existing floor r-value. May take a little longer to respond but could always add in a floor sensor to maintain a min slab temp to overcome response time. Simple radiant loss and use existing loops lengths/spacing to get flow rate, pump head and water temp required. Floor is going to heat the space provided he has enough tubing to overcome the loss.
    "The bitter taste of a poor installation remains much longer than the sweet taste of the lowest price."
  • GordyGordy Member Posts: 7,104
    Not rocket science Chris

    There are barriers. It seems between the tile installer , and HO there are certain obstacles.

    You can suggest all kinds of right ways, but if the money is not there?

    Slab sensors do not make existing more efficient it just tames the beast a bit. Need insulation, and yes control strategy and yes a heat loss calc.
  • rsc92rsc92 Member Posts: 50
    edited January 2013
    Our Compromise

    This is the compromise.  We are going to abandon the old pex and add a new loop of pex on top of the slab with sleppers (shown in my original post).  The tile guy will come in with his mud and attempt to keep the floors all the same height, if he cant, then we will just have to install saddles.

    If the tile guy goes with 1.5" of mud, then there will be somewhere between 3/4" and 1" of mud over the pex.  If he does it to meet the existing floor heights, then the total depth of the mud would be 1 1/8".  As such, we would only have approximately 1/4" to 5/8" of mud over the pex.

    I know the recommended amount of mud is 3/4" over the pex, but what is the drawback of having less - say 3/8" over the top of the pex.

    I know this compromise leaves no room for insulation under the pex, but I just cant go through the mess and expense of ripping up the cement, sleepers and old pex.  Hopefully the existing 4+/-" of cement and sleepers over the slab will provide some r value.
  • ChrisChris Member Posts: 3,056
    edited January 2013
    Through This

    Entire thread you have never once mentioned that

    anyone has calculated a radiant heat loss and used

    your existing loop lengths to even see what water temps

    you would need.

    You are only adding mass which effects response

    time. This can easily be addressed via slab sensor.

    You are on the path of a big mistake. Radiant isn't about

    just throwing tubing in a slab and it works. The last installation

    proved that.

    Get a radiant heat loss done and plug in the existing

    loop lengths!
    "The bitter taste of a poor installation remains much longer than the sweet taste of the lowest price."
  • rsc92rsc92 Member Posts: 50
    edited January 2013
    Heat Loss

    The plumber did a calculation today. I don't know what it told him, but We don't know the exact length of the existing pex, though, I could guess. He says a new run of the correct length is the most efficient solution since the pex will be much closer to the surface. He says we will use fewer btus than if we run hotter water through the old pex which will be some 5.75 inches from the surface.

    Makes sense and we will also add a slab sensor to pair up with the existing tekmar t-stat.yy
  • ChrisChris Member Posts: 3,056
    Blind Leading the Blind

    I never rant in this way. So sorry but your plumber is

    clueless. Depth of tubing has nothing to do with water temp.

    PS. Expect a cracked floor if you don't maintain that min of 3/4".
    "The bitter taste of a poor installation remains much longer than the sweet taste of the lowest price."
  • rsc92rsc92 Member Posts: 50

    Clueless, maybe! Blind, I don't think so.

    Perhaps, I am not explaining properly. What the plumber told me is that since we don't know how much pex we have in the floor already, he cannot assure us that it will generate the necessary BTUs to heat the room on the coldest days. 

    We already have a slab sensor in the existing setup.  That reads 85 degrees and the room is comfortable right now at around 70 degrees.  This is with 125 degree water feeding the loop.

    He told me that adding the extra mud and tile will require a longer time to heat the floor and may also require us to run water hotter than 125 degrees into that loop.  He told me he could not guarantee it would be satisfactory. 

    By installing new pex he can be certain that the correct linear footage is installed and is properly configured. 

    Fact is whatever is in the floor now is working and working fine.  It has been very cold here the past few days (low in the teens high in the low 20s).  It does not get much colder than that here. 

    Are you suggesting that he is wasting my money? That we can simply put a new slab sensor in the mud job the tile guy is installing and we would be done?  Am I correct that what you say is that adding the additional mass does not correlate to heat loss? I would think that my boiler will work much harder to heat all that extra mud and it would work far less if the pex were only an inch or so below the tile..

    Thanks for the heads up on the 3/4" of mud.  I guess we will have saddles.
  • ChrisChris Member Posts: 3,056
    edited January 2013
    I'm Just Saying

    Make sure you cannot use the existing radiant. Figuring out the length isn't that hard. Weigh the water. There is .92 gallons of water per 100' of 1/2" pex. One gallon of water weighs 8.33lbs

    All radiant software allows you to manipulate loops length. Do the heat loss, put in the existing loop length and see where you stand. You've stated that the prior system worked but not to expectation. You have lowered the finished floor r-value by installing tile.

    Is 15 minutes worth of homework worth the extra cost? Based on the pic and what I've read in your posts the room is 14x20. I ran the heat loss. Using these values with slab

    perm and below insulation R-Value of 2, Wall R-Value 11, Ceil R-Value 21

    Outside Wall 14'

    6x7 Door

    3x7 Window

    3x7 Window

    Outside Wall 20'

    100 Sqft of Glass

    Outside Wall 14 (Can't see it so no doors or windows)

    Heat Loss = 13,117 btu/hr

    Just saying. Slab thickness is mass and has nothing to do with water temp. Yes is may respond slow but forget zoning it with the thermostat and just maintain a slab temp. If we know for instance that with 111 degree water I can heat the room to the design setpoint with a surface temp of 77 degrees that's all that matters.
    "The bitter taste of a poor installation remains much longer than the sweet taste of the lowest price."
  • rsc92rsc92 Member Posts: 50
    Thanks Chris

    Your assumptions are fairly accurate.  Based on your numbers how many linear feet of pex shoudl be in my installation to be capable of generating the 13,117 btu/hr?

    Based on my crude (very crude) estimate, i would think we have somewhere between 220 and 260 linear feet of 1/2 pex.   If necessary, i can drain that loop and weigh the water.  Great idea!

    Thanks for the advice.  I would be very interested to know how many feet of pex we really need in that room. 
  • ChrisChris Member Posts: 3,056
    One Loop

    Of 260' would require a 20 degree delta-t (we like 10 for radiant slab) flow rate is 1.3gpm with a pump head of 14.2'. Water temp required 123. Not ideal but is workable. This is based off of assumptions as to the heat loss. See 15 minutes of work.
    "The bitter taste of a poor installation remains much longer than the sweet taste of the lowest price."
  • rsc92rsc92 Member Posts: 50
    edited January 2013
    Need more input - or Confirmation!


    I did a test today.  Set my minimum slab temperature at 88 degrees.  The room is fairly comfortable at 72 degrees by the t-stat and a bit cooler by the windows.  Sun is down and outside temperature is in the low 30's upper 20's.

    I told my plumber about my research.  He insists that the room will be more comfortable if we abandon the old pex and install new pex on top of the slab.

    Based on my experiment, i dont see how that would make a difference.  If the slab is 88 degrees, it is 88 degrees.  It should not matter where in the slab the pex is.  Now mind you my slab sensor is at the bottom of the slab (i will put a new slab sensor in on top of this slab before the mud job), so I may not need to have the t-stat min temp for the slab that high.   I do recognize that the extra mud and tile will add more mass, but if i can heat the room now, I suspect i will be able to do it after the tile job too. As many have pointed out, the r value of the mud and tile is very low.  BTW water temp feeding this zone is 125 degrees on the supply side.

    Am I correct that the only drawback of leaving everything alone and putting my new floor down ontop of the existing slab without any new pex loops is that I will use more natural gas to heat this room since i will inevitably be wasting some energy heating the slab and losing heat to the sides and bottom?

    Would the room be any more comfortable if the pex was higher up.  I really dont think it would be.  In fact, I dont see how it could be.  If we get the slab to the required temperature to maintain the room at the set temp, then the overall comfort should be no different.

    What it comes down to, I think is energy savings (or waste)....

    My plumber wants around $1,500 to run the new pex.  Based on his heat loss calculation (which he did not share with me), he said that i need 2 loops spaced 8" on center.  This would require a new mainfold because I dont have room on my existing manifold for the extra loop.  I think it will take a very long time to make up that $1500 in energy savings.  Perhaps he is right and it is the technically better way, but from a dollars and cents standpoint. I dont see how..

    Am I right????
  • ChrisChris Member Posts: 3,056
    edited January 2013
    Set Point Is What It Is

    If your reaching set point how is comfort changing. 70 degrees is 70 degrees no matter what water temp is running through the slab. Yes I would install a new sensor but because you have added mass it may take a little longer to respond but it also takes longer to cool because of the added mass. If you keep a slab min it will cut the response time to reach set point down. The added mass will make for longer cycles. In the end it will be a wash.

    Your only adding a .5 R-Value with tile vs the 1.68 you had prior. By the way 72 is hot for radiant. We design for 68. By the way 12" centers will work just fine. What is the surface temp of the slab?
    "The bitter taste of a poor installation remains much longer than the sweet taste of the lowest price."
  • rsc92rsc92 Member Posts: 50
    edited January 2013
    Thanks Chris

    Thanks for the input.

    I think we are going to leave well enough alone.

    I dont know the surface temperature of the slab, I dont have a good instrument to read it properly, but it feels warm to the touch, very warm, not hot.

    My wife is cold all the time - hence 72 degrees. It is very dry here and we dont have a humdiifier (i know we should get one).

    I guess my plumber is just shooting for overkill or wasting my $$

    Thanks for the sound advice.  Any chance you are in the NY area?
  • ChrisChris Member Posts: 3,056
    Your Plumber

    Is being cautious. Don't think any ill intent. Yes I live in NY and yes I visit the island but I am not on the contractor side. You should invest in an infrared laser. They are not expensive.
    "The bitter taste of a poor installation remains much longer than the sweet taste of the lowest price."
  • rsc92rsc92 Member Posts: 50
    Slab Temperatures


    Took your advice and went to Depot and got an infrared thermometer.  Interesting results.  My in slab sensor reads 85 degrees.

    The surface temperature of the slab goes from 70 degrees in the bay at the outside wall to about 95 degrees in the bay closest to the inside wall.

    I know the pex runs up and down the bays.  Based on the temperature readings i am getting, would you agree that the water flow begins at the inside wall and loops around the sleepers and ends at the outside wall before returning to the boiler?

    If I were to reverse the supply and return at the manifold, i wonder if we would get better (more even) heating.  This way, the hotter water would feed the outside bays first and as it cools going through the slab would reach the interior wall.   I would think that the bays closest to the exterior wall would warm up a bit more and the bays closest to the interior wall would not get quite so hot.  The other question is does it matter??  Since our couch is along the exterior wall, it may improve comfort in the room by better tackling the drafts from the windows and exterior wall.

  • ChrisChris Member Posts: 3,056

    If you reverse them the hotter water will be flowing where the heat loss is most on the outside walls. Concerned with the 90 degree surface temp. That is rather hot for tile.
    "The bitter taste of a poor installation remains much longer than the sweet taste of the lowest price."
  • rsc92rsc92 Member Posts: 50
    Even temp??

    if I reverse the supply and return do you think the 90+ degree temp i am reading on the bay closest to the inside will will temper since the return will be cooled by the poorly insulated outside wall/slab?  Should make the time to heat the slab initially a bit longer, but i think it would even out the temperatures a bit.  I would think it is unlikely that the bay on the outside wall reach that high of a sustained temperature.

    Is there risk to damaging the mud or tile with a surface temp that high?
  • ChrisChris Member Posts: 3,056
    Why Not

    Switching them now before you lay the tile. What is the pump model currently installed for the radiant?
    "The bitter taste of a poor installation remains much longer than the sweet taste of the lowest price."
  • rsc92rsc92 Member Posts: 50
    That is the plan

  • ChrisChris Member Posts: 3,056

    What make and model circulator pump on the eixsting radiant?
    "The bitter taste of a poor installation remains much longer than the sweet taste of the lowest price."
  • rsc92rsc92 Member Posts: 50

    Armstrong Astro-30
  • ChrisChris Member Posts: 3,056
    Someone Did

    A little homework because that pump fits right in the wheelhouse for a flow rate of 1.3gpm @ 14' of head. So you have the right pump and heating to set point. Sounds to me like the existing radiant is just fine. I am a little concerned with the high surface temp. Knock that water temp down to 120 and see how the floor reacts.

    A little homework goes a long way.
    "The bitter taste of a poor installation remains much longer than the sweet taste of the lowest price."
  • ChrisChris Member Posts: 3,056
    How Did the Experiment

    Work out for you?
    "The bitter taste of a poor installation remains much longer than the sweet taste of the lowest price."
  • LudwigLudwig Member Posts: 8
    edited January 2013
    How are you making out

    How is your problem working out?
  • rsc92rsc92 Member Posts: 50
    Will Keep You Posted

    I did not have a chance to reverse the supply and return for that loop at the mainfold.  I hope to get to it this weekend.  I did lower the supply temperature coming into the manifold to 120 and that resulted in a minor change in floor temperature in that room (and in the other rooms fed off of that manifold). 

    Once I have a chance to reverse the supply/return on the manifold, I think that will result in a more even floor temperature, but only time will tell.  I will keep you posted. Thanks again for the input.
  • ChrisChris Member Posts: 3,056
    Drop It Again

    Drop it to 115. How is the comfort and are you still hitting set point?
    "The bitter taste of a poor installation remains much longer than the sweet taste of the lowest price."
  • rsc92rsc92 Member Posts: 50
    Water Temp

    We are hitting setpoint and it is comfortable, but, it is mild here now.  The next few days will be the test.  Will keep you posted.
  • ChrisChris Member Posts: 3,056
    You May Want to Consider

    Changing that Thermostatic valve out for a Taco I-Valve once you make the decision as to which way you are going. Will not require any piping change and will add outdoor reset to the radiant. Better investment then adding the second pour and a whole lot less expensive.
    "The bitter taste of a poor installation remains much longer than the sweet taste of the lowest price."
  • ChrisChris Member Posts: 3,056
    Any Updates?

    Just checking in to see how things are..
    "The bitter taste of a poor installation remains much longer than the sweet taste of the lowest price."
  • rsc92rsc92 Member Posts: 50
    Reversed the Flow

    Yesterday, I swapped the supply and return at the manifold and the (surface) floor temperature was much more even.  Around 82 degrees in all bays with the exception of the bay closest to the exterior wall. That bay read around 75 degrees.  This was accomplished with input water temperature at 120 degrees.  Exterior temperature has been in the upper 20's low 30's.  So this was a good indicator. 

    Tile guy is supposed to start floor at the end of the week.  Just turned off the radiant in that room so that the slab can cool.  Tile guy wants it off so that mud does not dry too quickly. 

    Hopefully, I can get the radiant up and running by mid next week.  That will be the real test, but I fully expect it to work just fine.
This discussion has been closed.


It looks like you're new here. If you want to get involved, click one of these buttons!