Does Tube Size Matter?<br>John Siegenthaler

Figure 1

In radiant heating design, one manufacturer calls it the question that's always asked: "Do you get more heat out of a 5/8-inch pipe than a 1/2-inch pipe?"

Another manufacturer suggests there's a gain in heat output of about 2 percent using 5/8-inch tubing rather than 1/2-inch tubing.

Other manufacturers, although offering a variety of tubing sizes for sale, don't even discuss the thermal implications of using one size versus another.

What do you think? Does it make sense to increase the tube size in radiant floor slabs to get more heat output? You likely suspect that thermal performance increases along with tube size. But how much, and is it worth it? The more systems you design the more questions, like this one, you come across. Inquiring minds want to know the tradeoffs so they can make informed decisions, but where do those minds turn for information?

Models Vs. Myths: To get a handle on how tube diameter effects the thermal performance, I simulated the performance of a typical floor slab using the same heat transfer software I've shown you in past columns, such as "Depth Perception" in the May 2000 issue of PM.

This time I set up models of PEX tubing of various standard diameters spaced 12-inch on center in a 4-inch thick concrete slab. The models were tested at water temperatures of 80, 100, and 120 degrees F. The room above the slab was assumed to be at 70 degrees F. The slab was covered with 3/8-inch laminated hardwood flooring having an R-value of 0.43.

Models were constructed for tubing diameters ranging from 3/8-inch to 3/4-inch nominal I.D. The inside and outside diameters of the tubes modeled are as follows:

Each time a simulation is run it generates hundreds (sometimes thousands) of temperatures at different locations within the floor assembly. These temperatures can be grouped together to form "isotherms" (e.g. lines of approximately the same temperature) as shown in Figure 1.

At every location within the floor assembly, heat moves perpendicular to the isotherms, like "waves" of heat flowing away from the tube.

The isotherms in Figure 1 reveal how the concrete slab, having relatively high thermal conductivity, acts like a wick to move heat laterally away from the tubing and spread it through the slab.

The 3/8-inch oak flooring on the other hand has considerably higher thermal resistance than the concrete. The isotherms within the oak flooring are almost parallel to the floor surface. Heat is moving mostly upward rather than sideways in this part of the floor assembly. Similarly, heat flows almost straight down through the higher R-value underside insulation to the assumed 65-degree F soil below.

The temperatures along the floor surface determine the upward heat output of the floor. The simulations performed generated temperatures every 1/2-inch along the floor surface, from directly above the tube centerline to halfway between adjacent tubes.

Figure 2

The Results Please

Data from the simulations was used to make a graph showing the estimated upward heat-output of the floor versus water temperature for a room air temperature of 70 degrees F. A line showing this relationship for each tube diameter is shown in Figure 2.

To help put these numbers into perspective I divided the predicted heat-output for each tube size by the output of 1/2-inch tubing at the same water temperature. These numbers show the performance of the tube relative to that of commonly used 1/2-inch tubing. Numbers greater than 1 indicate more heat output, while numbers less than 1 indicate less.

The difference in heat output between 3/8-inch tubing and 1/2-inch tubing is minimal. For the situations modeled, 3/8-inch tubing puts out about 98 percent of the heat of 1/2-inch tubing.

The difference between 1/2-inch and 5/8-inch tubing is more pronounced. Again, for the water temperatures and materials modeled, the slab with 5/8-inch tubing averaged about 6.4 percent more heat than the one using 1/2-inch tubing. The output is slightly higher (7 percent) at the low water temperature (80 degrees F), and slightly lower (5.9 percent) at the higher water temperature (120 degrees F).

To check the effect of finish flooring, I changed the hardwood to concrete and reran the simulations. The average gain in heat output using 5/8-inch rather than 1/2-inch tubing increases to about 9 percent, a difference that's not trivial.

The difference in heat output between 5/8-inch and 3/4-inch tubing is also predicted to be quite small. When averaged over all water temperatures, the gain in heat output is only about 2.4-percent. The gain is slightly greater (5.4 percent) at 120-degree F water temperature, and slightly less (0.9 percent) at 80-degree F water temperature.

What That Means: Here's what I suggest these numbers are telling us:

The heat output gained by using 1/2-inch rather than 3/8-inch tubing, or 3/4-inch instead of 5/8-inch tubing is relatively small, averaging about 2 percent to 3 percent. A very slight change in water temperature could produce the same change in output -- so could a slight change in tube spacing, tube depth, or floor covering resistance, for that matter.
Therefore, changing from 3/8-inch to 1/2-inch tubing, or from 5/8-inch to 3/4-inch tubing, solely to boost heat output, is hard to justify.
The gain in heat output using 5/8-inch rather than 1/2-inch tubing is more noticeable, about 6 percent to 9 percent depending on flooring.
Still, for systems using a conventional boiler, it's a stretch to conclude that absorbing the higher costs associated with installing 5/8-inch tubing, solely for a gain in heat output, is better than boosting the supply water temperature or trying to get the tubing positioned at the mid-depth of the slab.

For systems using a condensing boiler, geothermal heat pump or solar energy collectors, it's important to keep the water temperature as low as possible to maximize the efficiency of the heat source.

For an output of 25 Btu/hr./sq. ft., the average water temperature of the circuit using 5/8-inch tubing would be about 2.5 degrees lower than a circuit using 1/2-inch tubing.

Although the gain in heat source efficiency associated with this temperature reduction is probably small, it should remain for the life of the heat source. A prudent designer would weigh the reduced operating cost over the life of the system against the higher initial cost of the large tubing.
Changing from 5/8-inch to 3/4-inch tubing based on the gain in heat transfer is also hard to justify based on the numbers obtained from these simulations. Such a size change would instead be driven by flow resistance considerations.
Depending on which manufacturer's data you reference, 3/4-inch tubing has from one-half to less than one-third of the flow resistance of 5/8-inch tubing. This can make a big difference in pump sizing for snowmelting systems.

Endless Questions

Although thermal simulation is not guaranteed to yield exactly the same results as field measured performance, it is a reliable tool for relative performance comparisons. Hopefully the numbers it produced for this column will help you put the thermal implications of tube size selection into perspective.

There are still plenty of questions related to the thermal performance of hydronic heating systems that simulations like this can help answer. For example: How does the thermal performance of a heated slab change for different tubing materials? What are the performance implications of using concrete with 20 percent higher thermal conductivity? What's the difference in heat loss for 1 inch versus 2 inches of under-slab insulation?

In future columns, I hope to show you more simulations that address questions like these. After all, when we as an industry stop asking these questions, stop looking for answers through reasonable means, stop trying to use these answers to improve what we do, myths, hype and complacency develop to fill in the gaps. As professionals we should do our best to see that this doesn't happen.