We are building a new home this summer with a geothermal system. We are from minnesota so it does get very cold. I want to heat the house with hydronic infloor heat.  The house has a basement which will have the tubes in the concrete floor.  The main level is where I'm not sure what to do.  I have read and read on which way is the best and all i do is get confused.  The main level is 1624 sq ft.  We would like some tile, some hardwood and some carpet.  What is the best way to put the tubing.  Some of the places say the staple up between the trusses but to me that seems like about the worst way.  Gypcrete but I read a lot of horror stories on putting tile and stuff on it ???  What are some other options warmboard?  Looks pretty sweet but pretty expensive.  What do some of you people on the forum have, do you like it??? What would you do different?  Just looking for some opinions.

Thanks for you time

We design and install every available radiant floor, wall and ceiling systems. What may work in one room may be inadequate in another. I have 4 different radiant floors systems in my own home. Each has its own attributes. Which radiant floor system is best for your home depends on the heat loads, floor coverings, heat source and most important; the design and skill-set of your installer.

All starts with a room-by-room ACCA Manual 'J' heat load analysis. Anything else might as well be a lot of hot air.

Relative to hydronic radiant floor heating system heat transfer efficiency and going from best to worst...slab-on-grade, thin-slab (gypsum), above-floor tube & plate, and below-floor tube & plate.

You will also lose additional heat transfer efficiency as you increase the floor R-value by adding tile, hardwood, and carpet. If the R-value of these floor additions exceeds about 2, you will not radiate very much heat from these areas...and you should consider these areas "blocked" relative to providing radiant heat gain. This just means that your design will have to provide more upward heat gain from the unblocked areas to satisfy the room/building heat load requirements.

All of this is good to keep in mind relative to heating system efficiency, however, in the end your hydronic radiant floor heating system design should properly address all these design factors.

As Badger from MN indicated, the first step is doing a proper heat loss analysis. We also have information and software on our website that many people have found educational and useful too.

blimes - I hear you with regards to gypcrete. In its proper place and with a proper mix, it is an acceptable solution. Still I don't like it. We will engineer a floor to accept either regular concrete or lite weight concrete. Often the joist design does not need to change to add the weight of concrete because the floor trusses have already been stiffened to take out vibration. The concrete layer dampens the floor so the stiffness for vibration is not needed and can be used to pick up the additional weight. When we place 1.5" concrete over pex we normally add poly fibres to the concrete and we put a second base plate under the walls to accommodate the thickness of the floor without compromising the ceiling height or the drywall layout.
Porcelain or quarry tile adds almost no r value to the floor so tile and acid finished concrete are preferred floor finishes. Unless you are going with a super insulated home, Mn should give you sufficient heat loss to keep the floors warm with out have to go with dust trapping carpet. A few throw rugs maybe?
Needless to say, get the proper design done for the hydronic system. Maybe someone you actually knows the area well?

Relative to hydronic radiant floor heating system heat transfer efficiency and going from best to worst...slab-on-grade, thin-slab (gypsum), above-floor tube & plate, and below-floor tube & plate

Aluminum plates are a better heat conductor than concrete and so with the same covering, I don't believe this. Do you have figures?

for tube and plate, sailaway is right in most cases, because tubing and plate contact is not perfect so theoretical conductive efficiencies are not reached. Plates are important but it's not perfect.

For "fitted" systems like Warmboard, raupanel, or thermofin plates this is much less true, and then you get into total coverage area. If you compare any of those high-grade aluminumized panels to a 'crete pour, they stack up very well, and even exceed 'crete outputs with equivalent floor coverings. if you do a nailed wood floor so the 'crete needs sleepers, and it's even more true.

that isn't to say that 'crete is bad. but it's not always better.

Posted By jonr on 24 Mar 2013 08:49 AM

Relative to hydronic radiant floor heating system heat transfer efficiency and going from best to worst...slab-on-grade, thin-slab (gypsum), above-floor tube & plate, and below-floor tube & plate

Aluminum plates are a better heat conductor than concrete and so with the same covering, I don't believe this. Do you have figures?

Yes, please consult Siegenthaler, Chapter 10, Figure 10-123.

Per the figure for the same 12" spacing and no additional material (e.g., tile, wood, light carpet) on top of the hydronic floor assembly, the heat transfer coefficients for slab-on-grade, thin-slab, above-floor tube & plate, and below-floor tube & plate are 0.882, 0.787, 0.55, and 0.36 respectively. Or in percent relative to slab-on-grade, 100%, 89%, 62%, and 41% respectively.

This also becomes very apparent if you do a design for a zone with given heat gain requirement and you only vary the hydronic floor assembly. The required boiler supply temp to achieve the required upward heat gain will increase significantly as you go from slab-on-grade to below-floor tube & plate.

So let's just run an example to see how the supply temp varies when we only vary the hydronic floor assembly. For our example let's use 100% water for the hydronic fluid, a design outdoor temp of 20 deg F, a design indoor temp of 70 deg F, an allowed circuit temp drop of 10 deg F, an unblocked heated area of 500 SF, a zone total heat loss of 7000 BTU/H, an exposed floor heat loss is 400 BTU/H, 0.5" PEX tube with a spacing of 12", and no additional material (e.g., tile, wood, light carpet) on top of the hydronic floor assembly. The required floor surface temp required to provide the required 6600 BTU of upward heat gain is 76.6 deg F. The required boiler supply temps to make this happen with slab-on-grade, thin-slab, above-floor tube & plate, and below-floor tube & plate are 90.7, 92.4, 99.5, and 111.8 deg F respectively.

You may verify these results by either using the software on our website or by using Siegenthaler's Hydronic Design Studio Pro V.2.0.

Posted By NRT.Rob on 24 Mar 2013 12:52 PM
for tube and plate, sailaway is right in most cases, because tubing and plate contact is not perfect so theoretical conductive efficiencies are not reached. Plates are important but it's not perfect.

For "fitted" systems like Warmboard, raupanel, or thermofin plates this is much less true, and then you get into total coverage area. If you compare any of those high-grade aluminumized panels to a 'crete pour, they stack up very well, and even exceed 'crete outputs with equivalent floor coverings. if you do a nailed wood floor so the 'crete needs sleepers, and it's even more true.

that isn't to say that 'crete is bad. but it's not always better.

Yes, we have also found that Warmboard approaches thin-slab relative to heat transfer performance. However, Warmboard is pretty pricy compared to conventional thin-set installation, but perhaps is easier to install and perhaps more robust too. We worked with a person in NJ after the hurricane who needed to install hydronic radiant heating first and also had to raise his home to met the new code requirements. I forget the reason why he had to accomplish the logistics in this strange order, maybe structural or occupancy timing issues. Anyhow, he used Warmboard because of having less likelihood of damaging the hydronic floor assembly during the subsequent house raising.

I would say it exceeds thinslab somewhat significantly. Costs comparisons are highly volatile though... 'crete can be expensive in many locations and cheap in others, and the size of the job matters too.

I would not compare at equivalent on centers either. Warmboard is an even temperature at 12" o.c. 'crete in a thin pour is not and requires a 9" o.c. install or tighter in all residential cases if even floor temps are a consideration, as do most other panels by design. so some cost is made up in material as well.

Yes, a 9" spacing for thin-set is pretty typical...and sometimes even 6" is appropriate. Sieg's figure 10-123 doesn't have a 9" spacing option for above-floor tube & plate and below-floor tube & plate (only has 8" and 12" given conventional joist spacing) so I had to use 12" for my example. You need to keep the spacing the same to do a proper comparison relative to heat transfer. Decreasing the spacing will increase the heat transfer for all the hydronic floor assemblies...just not equally. We have not done a proper side-by-side comparison of Warmboard versus standard thin-set to say one way or the other with any degree of certainty, however, I suspect you are likely right Rob.

Looks pretty sweet but pretty expensive

I am just finishing my own residence in which the goal was to optimize the efficiency of the heating system. The project was overseen by both architects and solar engineers. Again and again and again, both building and HVAC contractors tried to turn us to the tube and plate route or similar. Every time we investigated them, we found that they were SUBSTANTIALLY more expensive than concrete slab for what you get. Although I am sure there are situations, including retrofitting, which warrant it, they are just another product someone wants to market. If you want a hard-surfaced, efficient radiator floor, the way to go is concrete. For a first floor over a basement, look at a product called QuadDeck. Goes down fast, lay your tubing, pour your concrete. Done.

I would wager that people who are not already dealing in large volumes of concrete may not find the comparison quite the same as you do, at least not reliably in all locations.

I also don't know many residences with suspended full slabs like some ICF homes do. From what little i know of costs, the quad deck type products are much more expensive than typical subfloor framing. so you have to compare more than just pipe vs panels. the total floor system is important. I haven't actually done a cost comparison though so I'd be happy to sit back and learn about this with real numbers.

> we have also found that Warmboard approaches thin-slab relative to heat transfer performance.

Here are comments (caution: from Warmboard) regarding their product's significant superiority over thin slabs. I would expect tight fitting, ~100% coverage aluminum plates to perform as well as Warmboard. And as they point out, high mass (thick slab) creates problems with slow response (ie, under and overshoot).

http://warmboardradiantheat.blogspot.com/

I haven't actually done a cost comparison though so I'd be happy to sit back and learn about this with real numbers.

The cost of my Insuldeck was less than $10/sf. That's formed, braced, poured and finished.
When we looked at Warmboard, the price of the board alone was $6/sf, not to mention building the floor.

Jon, unless I miss my guess, the Warmboard folks like to compare their product to "thin slab systems". Looking at the data, I suspect that is gypcrete which does not transfer heat quite as well as good old concrete.

The reason slab-on-grade transfers heat approximately 11% better than thin-slab is because of the concrete and steel composition. For those of us that integrate both passive solar and hydronic radiant heating systems, the high mass provided by a thick slab allows it to be used as an effective thermal mass...to absorb and store heat for a long time...which minimizes temp variations in the building...which improves the overall comfort level. However, a hydronic heating system has to continually supply Warmboard with hot water in order to maintain an even temp warm floor. We consider Warmboard to be okay for remodels, but we would not recommend it for new construction either...if cost and heating system efficiency are high on your priority list.  Sometimes customers just want hydronic radiant floor heating in their existing building, they are clueless and care less about the pros/cons of the different design options, and cost is not an issue for them...then Warmboard becomes an acceptable option.

thick slab allows it to be used as an effective thermal mass...to absorb and store heat for a long time...which minimizes temp variations in the building...which improves the overall comfort level

A high mass radiator (eg, thick slab) does the exact opposite - it causes temperature variations (slow response, overshoot) in the building. And the energy savings available from passive interior mass (non active radiators) only occur when you allow significant interior temperature changes. In both cases, LESS comfortable.

"Mn should give you sufficient heat loss to keep the floors warm with out have to go with dust trapping carpet."

The location does not determine the heat loss so much as the construction.

Here in Minnesota we have cold weather and it does tend to keep our floors warm more of the time.

Sailor is right, but for all practical purposes, the differences in design water temperature in the example sighted is of little consequence since all can be achieved in the typical range of most common heat sources including condensing boilers and heat pumps.

So the real challenge is finding the most efficient type of radiant floor in terms of operation and installation. We have the edge on specification of the many options since we design, install and service hydronic radiant floor heating system the Minneapolis/St. Paul area. Much of our work is no troubled boiler/radiant floor systems that suffer from a lot of theory and little experience. 12" PEX tubing spacing is a little "fat" for most residential applications. As for low-mass vs. high mass radiant panels. The structure matters. Without mixing it up with solar (changing controls strategies entirely) high and low-mass panels can be easily controlled to overcome temperature swings in both slab and ambient air (remembering Rob's lectures on the efficacy of the esteemed Tekmar 522). This

Room-by-room heat loads using dedicated radiant floor software. The designer should provide a CAD drawing before the job starts.

http://www.badgerboilerservice.com/images/SampleHeatLoadAnalysis.pdf

It comes down to the specific job and local resources, as Gypcrete® started here in the Twin Cities, it is competitive and concrete is common everywhere. With the addition of outdoor reset and new construction techniques the "mass" of emitters, be they concrete or radiant ceilings is of little concern since the boiler will modulate to keep the design water temperature matched ot the heat load. Fitting the right system to the customer's specific needs including economy and efficiency is what radiant floor designers do.

I have to strongly disagree Jonr. The design objective of a hydronic radiant floor heating system is to get the floor surface to the exact temp required to provide the required zone radiant heat gain to precisely offset the zone heat loss and to maintain this floor surface temp with minimum system effort and maximum system efficiency until the zone heat loss changes…which is typically very slowly as a result of outdoor temp changes. Once a high mass floor slab reaches the required surface temp, it tends to stay there and it is easy for the system to keep it there…and the customer will not notice any variation in floor surface temp like they would for a cycling Warmboard system. Under and overshoot has not been a problem since the days we stopped putting metal pipe in slabs back when micro processor based control systems first became widespread.

Surface temperature requirements change with load. high mass slows down how fast you can adjust your surface temperature requirements. That is not a benefit. I have worked on many a system in which high mass emitters were not properly managed to change temp with load and in fact, in many cases they simply cannot because, you know, physics. Under and overshoot is a control problem, largely... with proper thermostats and controls most any issue can be dealt with. But it's a requirement low mass emitters do not have... they can be operated in a simpler fashion without concern for overshoot/undershoot.

Mass can provide bufferring benefit, but if you need bufferring, putting that mass in the water, in a tank, is much better than in the emitter. no sacrifice of controllability in that case and no consideration for zone sizing either. Simple and very effective.

Mass is fine and it can be worked with, but I don't regard it as superior in any respect to a low mass system. It's just a different method, that requires more attention to detail than low mass systems do.

Yes, I would agree that a low mass emitter requires less attention to control design detail than a high mass emitter.

As I mentioned previously, when you integrate passive solar and hydronic radiant floor heating systems, using passive thermal mass (4-6” thick masonry floor/walls) is typically the first step to provide the required buffering once you exceed a specific passive solar heat gain. One also has to factor in system complexity and maintenance when doing a design so passive thermal mass may turn out to be the best only step too. Warmboard is not going to provide any significant buffering, which is desired/required, for this situation.

Yes, when you get even more aggressive about capturing passive solar heat gain, it is better to a use an active thermal mass (to date, our designs have been 4-6” thick masonry floor/walls with circulating hydronic fluid). This is another advantage I see with vertical ICF compared to horizontal ICF. Rob, have you used Warmboard for active thermal mass and, if so, can you share how well you found it to perform?