HPAC Magazine

Selective Coverage

May 1, 2012 | By John Siegenthaler

Radiant panels do not have to cover an entire floor, wall, or ceiling.

Radiant panel heating has matured from being the darling of the hydronics industry in the 1990s, into a respected technology that can provide excellent comfort in a wide range of applications.

Any of you reading this article have probably designed, and/or installed several radiant panel systems. In many cases, those systems involved covering an entire floor area with some type of radiant panel construction detail: slab-on-grade, thin-slab, tube and plate, etc. This has become standard practice in the industry and works well when radiant floor heating is installed in houses with average heating loads. However, as the design heating load per unit of floor area decreases, so does the average floor surface temperature. In very well insulated houses, the average floor surface temperature may only be a few degrees above the room air temperature. The reason is that the floor does not need to get any warmer in order to satisfy the heating load as determined by the setting of the room’s thermostat. Forcing the floor to operate at higher temperatures would quickly overheat the space, and probably lead to energy waste in the form of open windows.

From the standpoint of thermal efficiency of the heat source, lower surface temperatures are not a problem. Heat sources such as condensing boilers, hydronic heat pumps and solar thermal subsystems will all operate at high efficiency in combination with low water temperatures. Remember that the lower the water temperature is, the higher the heat source efficiency.

The “problem” is that the owner’s reasonable expectation of warm floors may not be realized. And as most of you can appreciate, unfulfilled customer expectations are a problem, even when the heating system is working at peak efficiency.


There are several alternatives that provide a reasonable balance between heat source efficiency and the owner’s desire for warm surfaces. One is to make the surface area of the radiant panel smaller by not covering the entire floor area with tubing.

For example: Imagine a room with a design heating load of 3000 Btuh, and a corresponding indoor temperature of 70F. The room measures 20 feet by 30 feet. If the entire floor area is covered with radiant panel, the upward heat flux requirement at design load would be:

Formula 1

The average floor surface temperature can be estimated using Formula 2.

Formula 2


Ts = average floor surface temperature (ºF)

q = upward heat flux (Btuh/ft2)

T air = room air temperature (ºF)

Thus, for the stated example:

This temperature is a few degrees lower than the normal skin temperature for hands and feet. The infrared thermograph of a normal (thermally comfortable) hand in Figure 1 shows fingertip temperatures in the low to mid 80s.

A floor surface at 75F surface will feel slightly cool to the touch of this hand, even though that floor is releasing sufficient heat to maintain the room at 70F.

Figure 2 shows an infrared thermograph of a radiantly heated floor – you can easily see the embedded tubing. After the hand of Figure 1 was pressed against the floor for a few seconds, it clearly shows the residual heat absorbed from the hand into the floor.

Also keep in mind that the 75F average floor surface temperature only exists on a design day, when outside temperatures are at or close to their lowest values. This average floor surface temperature will be even lower under partial load conditions.

The issue now becomes one of customer expectation. If the customer was informed that the floors would not feel warm, even though interior setpoint temperature would still be maintained, and if they understood and agreed to this operating condition, there should not be any unfulfilled expectations. However, if the customer cannot get their brain past all those cozy barefoot advertisements for radiant floor heating and still expects warm floors regardless of load, the result is likely to be serious disappointment.

The retort “but I paid for warm floors…” will surely be heard and your prospects for a good customer relationship are headed south. The fact that the modulating/condensing boiler you just installed is operating at 97 per cent rather than 92 per cent thermal efficiency is probably not going to smooth things over.


If the size of the radiant panel in the previous example were cut in half, the necessary upward heat flux would change from 10 to 20 Btuh/ft2. This would bring the average floor surface temperature on a design day from 75F up to 80F. Although still a tad low, such a temperature may appease those looking for barefoot-friendly floors. Reducing the panel area to one third of the room’s floor area would boost the average floor surface temperature to about 85F, a recommended maximum for floors in which there is prolonged foot contact.

The concept of not covering the entire floor area with tubing was common in the days when copper tubing was used for radiant floor heating installations. Each radiant panel was sized to the room load assuming a specific upward heat flux and specified supply water temperature. A room with half the load of another would get half as many square feet of panel area. Assuming floor coverings of comparable R-value, this approach allows the system to work with a single supply water temperature and eliminates multiple mixing devices.

I used this approach when designing the floor heating system in my own house in 1979. Figure 3 and 3a show images of the floor heating panel in our dining area. The panel was constructed using  3/8 -inch copper tubing because PEX tubing was not available in North America at the time. We installed the radiant panel under the eventual location of the dining table, right where feet rest on the floor. It feels great on a cold winter morning.


Another approach is radiant wall or ceiling heating. Since occupants do not rest their feet on them, radiant wall and ceiling panels are not constrained to a maximum average surface temperature of 85F. Instead, a practical surface temperature limitation for a gypsum surface is about 120F. This temperature limit is based on avoidance of long-term degradation of drywall joint compound.  The heat output of a radiant wall can be estimated using Formula 3.

Formula 3

The heat output of a radiant ceiling can be estimated using Formula 4.

Formula 4

The variables in both these formulas are as follows:

q = outward heat flux (Btuh/ft2)

Ts = average surface temperature (ºF)

Tr = average room air temperature (ºF)

Formula 3 implies that a radiant wall with an average surface temperature of 100F, releasing heat into a 70F room, would yield an output of: 1.8 x (100-70)=54 Btuh/ft2. A radiant ceiling operating under the same conditions would yield 1.6 x (100-70)=48 Btuh/ft2. These are both significantly higher than what is possible with floor heating assuming the latter is constrained to an average surface temperature of no more than 85F based on physiological comfort. At 85F average surface temperature, a heated floor releases about 30 Btuh/ft2 into a room with an air temperature of 70F.

The higher outputs associated with radiant wall and ceiling surfaces imply that smaller panel areas are possible. Smaller panels mean less materials and lower installation cost.


If you are designing radiant panels for very well insulated buildings there is a good chance that a high efficiency hydronic heat source will also be used. The thermal efficiency of condensing boilers, solar thermal collectors and hydronic heat pumps is dependent on the water temperature at which the heat distribution system operates. The lower the better. My suggestion is to select and size the heat emitter so that the supply water temperature
does not exceed 120F on a design day. This temperature is attainable by all of the aforementioned heat sources, and allows a reasonable compromise between surface area, surface temperature and heat source efficiency.

The radiant wall construction shown in Figure 4 has proven to be a good performer on the basis of heat output versus water temperature. Its output can be estimated using Formula 5.

Formula 5


q = heat output of wall (Btuh/ft2)

Twater = average water temperature in wall panel (ºF)

Troom = room air temperature (ºF)

Figure 5 shows this type of radiant wall during operation. The upper image is what your eyes see. The lower image is what a thermographic camera sees. Notice how well the aluminum heat transfer plates are dispersing heat away from the tubing. The greenish/yellow stripes are two-inch wide areas between these aluminum plates.

If you were to use the same construction shown in Figure 4 for a radiant ceiling, the panel’s heat output can be estimated using Formula 6:

Formula 6:


q = heat output of wall (Btuh/ft2)

Twater = average water temperature in ceiling panel (ºF)

Troom = room air temperature (ºF)


It is hard to overestimate the importance of warm surfaces in conveying physical, as well as psychological comfort. This basic human need can be balanced with the desire for low energy use and high equipment efficiency by using smaller radiant panels. As you plan future systems for very energy efficient homes, be sure to have a conversation on these tradeoffs so customer expectations will be met. <>

John Siegenthaler, P.E. is the author of Modern Hydronic Heating (the third edition of this book is now available). For reference information and software to assist in hydronic system design visit www.hydronicpros.com.



Stories continue below