Although floors are certainly the most often used room surface for site-built radiant panels, other room surfaces can also deliver very comparable comfort. By its very nature radiant heat travels equally well in any direction. This allows a heated wall or ceiling to bathe its room with radiant energy.

In some situations these "alternative surfaces" can correct for conditions that limit or even rule out floor heating. For example, if floor coverings with high thermal resistance are being specified the better approach may well be to forget about heating the floor and focus instead on walls or ceilings.

Many of you radiant specialists have undoubtedly heard all this before, but never taken the next step of hydronically warming a surface other than a floor. Until last summer I, too, had not looked beyond the soles of my shoes when site-built radiant panel heating was "in the cards." But then along came an opportunity that begged for a little experimentation.

Mirror mirror, on the wall, who's the best guinea pig of all?

Do unto thyself before specifying unto others.

I had several opportunities for such experimentation during the construction of a new office for Appropriate Designs last year. Since

I'd be spending a lot of time in the new building I could literally "sit in judgement" (like the U. S. Senators back in January) of how well the various experiments performed. Then I could decide which aspiring approaches would see the light of day in other buildings.

My very capable associate Harvey Youker was equally enthusiastic about venturing into uncharted territory with me. Together we designed and built our first radiant wall.

This month I'll describe how we built it; why we built that way; and how it's performing thus far. When I get some accurate performance numbers I'll pass them on in a future column.

Not Too Big ... Not Too Small ... Soon To Be ... A Radiant Wall

We began the job by cladding the interior side of the 2 by 4 truss webs with 7/16-inch oriented strand board (OSB). It's the same material often used for exterior wall sheathing on houses. The 4 by 8 ft. sheets were simply tipped up along their long side and nailed to the vertical truss webs. Besides serving as part of the heating system the OSB provided strong lateral bracing for the trusses.

Next we snapped horizontal chalk lines on the OSB starting 3-5/8-inch up from the floor and then vertically every 8 inches. These marked the top edges of the foam furring strips that would eventually support the heat transfer plates.

The furring strips were made by ripping 3/4-inch foil-faced polyisocyanurate insulation board into 7.25-inch wide strips using a table saw. We used a product called ThermaxR from the Celotex Corp. This material has a number of desirable properties for this application including:

• Very low thermal mass (less than 5 percent that of wood). This allows the radiant wall to achieve normal operating temperatures quickly following a cold start.

• Its R-value of 5.4 adds to that of the back-side insulation to help direct heat into the room, rather than out through the rear side of the wall.

• The aluminum foil facing enhances lateral heat conduction away from the tubing in concert with standard aluminum heat transfer plates. It also provides a good bonding surface for adhesives.

• The product remains dimensionally stable and does not outgas at temperatures as high as 250 degrees F, well above any temperature it would see in this application.

The foam furring strips were secured to the OSB using Franklin solvent-based contact adhesive. We used a standard paint roller to apply the adhesive to both the OSB, and rear side of the furring strips. After waiting five minutes for the adhesive to tack-up, the furring strips were simply aligned to the chalk lines and pressed against the OSB. If you've ever installed laminate countertop you know that once contact adhesive makes contact with itself, it's there to stay. Small 3/4-inch wide wood spacer strips set along the top edge of each successive row of furring strips made it easy to set the strips in place and keep them aligned.

Next we used a 3-inch wide paint roller to coat the upper half of each foam furring strip and one rear side of each aluminum heat transfer plate with contact adhesive.

This would hold the plates in place while the tubing and drywall was installed. If you try this DON'T apply contact adhesive on both rear sides of the plates. If you do the plate will bond in place so rigidly that the tubing channel can't expand as you're trying to press the tubing in place, especially if the plate is slightly bent. Believe it or not we actually anticipated this problem before we discovered it by accident.

The plates were set in place leaving about 1/2-inch between their ends for expansion. As the plates were pressed in place the side with the adhesive was held tight against the coated edge of the furring strip. This creates a slight gap between the groove in the plate and the opposite furring strip allowing room for the plate to expand as the tubing is pressed in place. We didn't anticipate this seemingly minor detail but quickly learned it improved our system.

At the ends of the wall the foam strips where cut short to allow space for the return bends. A vertical 3/4-inch by 2-inch wood strip was installed at each end of the wall to provide a solid fastening surface for the drywall.

Next came tubing installation. The 1/2-inch PEX-AL-PEX snapped right into the plates and was easy to hand bend at the ends of the wall.

Half-inch drywall was installed over the entire assembly. To eliminate butt joints we cut the 4 by 8 sheets in half and installed them with their tapered edges vertical. Because of the OSB backer sheet, a 2-inch drywall screw could "catch" anywhere. We set up a grid for the screws spaced 12 inches apart horizontally, and 8 inches vertically. We were careful to snap horizontal chalk lines half way between (rather than directly over) the tubing. Why? Well, can you imagine getting this far only to shoot several rows of drywall screws right through the tubing because of our habitual tendency to drive fasteners through lines!

The final step was to install 4-inch fiberglass insulation on the rear side of the OSB. In our case we could reach the back side of the interior knee wall from the first floor of the building. Obviously this won't always be the case. Some situations could require the back side insulation to be the first material to be installed.

Figure 1 shows a detailed crossection of the system we used. Figure 2 is an elevation view of the wall showing how the tubing was routed.

Receptacle Remedy

Water for the radiant wall is supplied from a manifold controlled using a variable speed injection mixing system. To date the mixing system has been operating in setpoint mode rather than outdoor reset mode (although it's capable of either). Whenever the room thermostat calls for heat a circulator supplies water at 125 degrees F to the wall circuit. The tubing is connected so that water enters at the top of the wall and moves down as it cools. We did this to make the wall a counter-flow heat exchanger as far as convection was concerned. The flow can easily be reversed at the manifold, which no doubt will be part of the continuing experiment.

I've run several informal tests this past winter during which the wall was turned on from a cold start while I listened for expansionsounds. Thus far I can honestly tell you I've not heard a single "tick" from the wall. Heat output is noticeable about 10 minutes after a cold start (when supplying water at 125 degree F). The wall appears to reach a steady operating condition about 30 minutes after start up. At that point it's hard to detect much variation in surface temperature using the highly scientific method of running my hand slowly across the wall. Inevitably instruments will find some variation, but my subjective side concludes it's not enough to cause any complaints.

The heated wall has been in operation for one heating season. Thus far I've not noticed any problems such as screw pops, expansion noises, hairline cracks or smells from outgassing materials. I'm convinced the system works well and am ready to use it again.

This system is well suited for rooms with chair rails. I'm sure most of you know what chair rails are. For those that don't, image a hardwood molding about 2 inches wide and 2 inches high that's fastened to the walls about 36 inches above the floor. They could make a perfect divider strip between a slightly thicker heated wall below, and a standard wall above. No need to shim out the upper studding to eliminate a step in the wall. Heating the lower portion of the wall also favors radiant output into the lower portion of the room where it's most desired. It's also a pretty good bet that people won't be driving nails to hang pictures below the chair rails.

Points Worth Pondering

• This system has very low thermal mass allowing it to respond quickly to load changes. This is nice for spaces that are normally kept at reduced temperatures, but often need to be brought up to normal comfort conditions in a short time. The low mass characteristic can also help prevent overheating in rooms spaces with large solar (or other internal) heat gains.

• The surface temperature of a radiant wall can be higher than that of a radiant floor without producing discomfort. A rule of thumb used in the radiant industry suggests keeping drywall surfaces below 120 degrees F to prevent cracking or discoloration at joints. If we assumed an average surface temperature of 110 degrees F at design conditions, a radiant wall should be able to deliver about 60 Btu/sq. ft./hr. into a room at 68 degrees F. That's almost double the permissible heat output of a radiant floor if the average surface temperature is not to exceed 85 degrees F to prevent the "hot foot" effect.

  Wherever there are strengths, there are weaknesses. Here're some things to be cautious about:

• A heated interior surface on an exterior wall could substantially increase the wall's exterior heat loss. To counteract this the R-value of such a wall should be increased by about 50 percent to keep back side losses comparable to those of a non-heated wall. The system described adds about 5.8 to the R-value of the wall it's applied to.

• It's wise to ask clients about furniture placement relative to where a radiant wall is being considered. For example, placing upolstered furniture up against a heated wall could block a substantial portion of its heat output.

• Locations where nails, screws, or other pointy objects are likely to be driven into walls are obviously not good areas to install water-filled tubing. I'd like to say that making occupants aware of the fact there's tubing in the walls would prevent the obvious problem, but you'll have to judge that one for yourself.

I'll have more information on the radiant wall in the months ahead. In the meantime feel free to try it yourself. Who knows, it may even become a key part of you radiant repertoire.