I would consider radiant floors (or walls or ceilings) everywhere and use the ERV ducts for latent loads (ie, DOAS). It should be more efficient and simpler than a mix of radiant and air handlers. I'd consider avoiding staple up radiant in favor of plates or tubes closer to the surface. I would also compare to a traditional WtA ducted system and to multi-splits everywhere (I know that you said you don't like either, but always nice to know). I would try to avoid mixing systems (only because I value simplicity).

An extra advantage of DOAS and large radiators is that the supply tubing never needs to be cool enough to cause condensation. Eg, 60F or more. And no condensate plumbing (outside of the mechanical room).

what total cooling capacity should I be looking for, or how much "padding" do I want to add?

With your heating load, it looks like you are going to have plenty of padding, whether you want it or not.

After reading through your thread, it appears that a 5 ton w-w will carry your load, if your load numbers are correct. I agree that you should consider putting top of the floor radiant in your house on 1st and 2nd floor with aluminum transfer plates, something like the Raupanel system comes to mind. It is important to understand the concept of going for the lowest possible supply temperatures, every degree of lower supply temps will render your system 1.5% more efficient. 90F going to the floors instead of 120F will cause the geo system 45%less efficient, for life! Stable up even with heat transfer plates will require significantly higher temperatures and make your geosystem significantly less efficient.


The same word of caution about your radiant cooling. You life in an area with high humidity. and with a pretty high cooling load. Your heatgain numbers "feel" off, in GA you should not have twice as much heatloss than heatgain. You need to check your cooling numbers. It is also hard to imagine for me that you can carry more than 1/3 of your cooling load with radiant cooling before you run into trouble with condensation, plus cold floors are uncomfortable to walk on barefoot. Why don't you have 2 air handler, one for each floor, fed by chilled water from the geo system, and call it a day?

With the weather in GA, I would expect 2x heating vs cooling to be a good starting point for sensible load. But solar gain, internal gains, aux heat and latent load should move the numbers somewhat closer together.

If you use fans coils, you are eliminating the efficiency/delta-T advantages of large radiators. Probably for the entire house.

Posted By jonr on 28 Apr 2013 02:29 PM
If you use fans coils, you are eliminating the efficiency/delta-T advantages of large radiators. Probably for the entire house.

Doesn't one need to account for the internal gain, solar again and latent load? So if his heating load for a 4500 sqf house south of Atlanta is 44KBTU/H, his cooling load should be at least in the same ballpark with solar gain, latent load and internal gains. How are you dehumidify the air on humid days and at the same time provide enough cooling with the floors without condensation? I have seen this approach fail multiple times by installers who underestimated the condensation issue here in Buffalo NY. How do you make this work then south of Atlanta GA, where cooling loads and humidity are significantly higher?

his cooling load should be at least in the same ballpark with solar gain, latent load and internal gains

I've been reviewing this myself to see why the heat loss is so much higher than the heat gain. It appears that my two biggest "losers" are a masonry fireplace/chimney (not set in stone yet, no pun intended) and the north and south walls of the great room. If I remove the reflective coating from all the windows on those walls, the heat gain skyrockets, even with low E double pane glass.

The other thing that seems to contribute a LOT to the heat loss is the 2 french doors on the south wall. again, Low E with reflective coating. very low gain, but high losses.

I can drive up the latent load in the calcs artificially, by assuming an "average" envelope and increasing the infiltration, but that's one of the things the house has been designed against. What I don't want is the grossly oversized systems that some have tried to sell me. This is the whole reason I started doing so much of the legwork myself.

The point is that even for your calcs of 44KBTU/H for the heatloss with radiant you need a 5 ton heatpump, no matter if your heatgain is 20,000 or 40,000, your geothermal system size is the same.

The second point is that your geo capacity and efficiency will dramatically different if you design the floors for 90F supply temp, or 120F design temp (under the floor staple up, even with plates). You should consider putting the radiant system on top of the subfloor with aluminum plates.
The 3rd but most important point is that you will not be able to control your humidity in Atlanta GA climate and cool it with the radiant floors alone, without air handlers, no matter what the final cooling load (20 kbtu/h or 40kbtu/h) is.
http://welserver.com/WEL0384/

You need a design similar to WEL0384, where a heatpump conditions a buffer tank for either heating (radiant) or cooling (air handlers).
That keeps it simple and ensures that you don't get any condensation issues.

Infiltration is the key - if it is low, then so is latent load. It is well proven that DOAS systems can work, even in humid climates. So I wouldn't let "rules of thumb" trump the numbers. I'd also give some thought to unusual events. Say you are cooking during a party for 30 people. But with the extra sensible capacity you will have and a portable dehumidifier or two, you should be fine.

You should consider putting the radiant system on top of the subfloor with aluminum plates.

this should be easy enough to accommodate; without getting the topic at hand too convoluted, any recommended installation methods? I'm looking into warmboard, but that seems like it could be a bit cost prohibitive.

You need a design similar to WEL0384

this looks a lot like my original plan, with a slightly different delivery scheme to the loads. I had the heating and cooling side split by diverter valves from the tank. this may be a more efficient approach, and allow the AHUs to share the load and provide air circulation.

y'all have sure enough given me a bit to think about. By all means it's appreciated, keep it coming.

I'm looking into warmboard, but that seems like it could be a bit cost prohibitive.

I would also look at RHT Floor Panel System and any of the systems that embed tubes in plaster on the walls or ceiling.

As mentioned infiltration is the latent issue along with respiration of the occupants.The other point that I do not believe has been addressed is exhaust rates. What are you using for your kitchen hood, what is the cfm,how are you planning to make-up the air exhausted.The kitchen fan will change your load calc and your equipment selection.

Posted By danreed76 on 28 Apr 2013 03:37 PM

his cooling load should be at least in the same ballpark with solar gain, latent load and internal gains

I've been reviewing this myself to see why the heat loss is so much higher than the heat gain. It appears that my two biggest "losers" are a masonry fireplace/chimney (not set in stone yet, no pun intended) and the north and south walls of the great room. If I remove the reflective coating from all the windows on those walls, the heat gain skyrockets, even with low E double pane glass.

The other thing that seems to contribute a LOT to the heat loss is the 2 french doors on the south wall. again, Low E with reflective coating. very low gain, but high losses.

I can drive up the latent load in the calcs artificially, by assuming an "average" envelope and increasing the infiltration, but that's one of the things the house has been designed against. What I don't want is the grossly oversized systems that some have tried to sell me. This is the whole reason I started doing so much of the legwork myself.

Okay, now I can't stop digging into this... I went back through my data, and my biggest heat loss for a single component is the basement slab (i forgot to calculate underslab insulation into the works). My question now becomes, do I go with the code minimum (r-5 2' in from the edge of the slab) or go ahead and insulate under the whole slab? I guess the question really is do I lose benefit in my cooling (longer cooling season than heating) by insulating the whole slab, or do I save more in heat loss by insulating it? The tables (based on Atlanta, GA as a reference city) say we have a 64F degree ground temp. While digging the basement I measured 62F.

I'd start with 2" of foam under the entire slab and then try some variations to see if you like the payback. Note that radiant heat in the slab makes the insulation even more important (for btu and response time reasons).

Another big item I found is that if I take the reflective value of the window coatings out of the equation, the heat gain comes up nearly 12,000 BTU/h. Are these coatings really that good, or just another hyped up product? I don't have my window quotes in front of me, but for the whole house to upgrade the window and door glass to reflective seems well worth it if that's the real expected performance.

The highly reflective window coatings DO work, but can really darken the place, and aren't always the best solution. Cooling loads soar when there is a lot of west-facing glass, since they're difficult to shade from the exterior without obscuring the view, and the sun angle is low for high-gain during the hotter part of the day.


My gut agrees with docjenser's that 44KBTU/hr is a pretty high heat load at 23F for a house that size, but may be a function of a higher-than average window/floor area. A 2x6 wall w/ cheap R19s has only about a quarter the heat loss per square foot of a pretty-good U0.28 double-pane. If your calculation tool isn't taking drapes/shades into account, that can also exaggerate window losses beyond in-situ reality. Similarly, site factors like mid-day and afternoon shading from trees/buildings/hills on windows can make huge differences in the actual vs. calculated cooling loads when not properly accounted for.

It's often cheaper & better to reduce both the heating and cooling loads by reduced glazing area. You don't have to live in the dark, and daylighting is useful, but even modest reductions in window area can make measurable differences in both heating & cooling loads. Reduced west-facing glass in particular can be key for peeling off the peak cooling loads.

but for the whole house to upgrade the window and door glass to reflective

Or perhaps just the east and west sides.

Posted By jonr on 28 Apr 2013 11:48 PM

I'm looking into warmboard, but that seems like it could be a bit cost prohibitive.

I would also look at RHT Floor Panel System and any of the systems that embed tubes in plaster on the walls or ceiling.

The warmboard is a good quality product but it is difficult to install with the joist bays open. I like the Raupanel, which performs the best, but the sheer amount of protruded aluminum makes it spendy as well. A cheaper alternative are the aluminum transfer plates from Radiantmax.

Posted By Dana1 on 29 Apr 2013 11:21 AM
It's often cheaper & better to reduce both the heating and cooling loads by reduced glazing area. You don't have to live in the dark, and daylighting is useful, but even modest reductions in window area can make measurable differences in both heating & cooling loads. Reduced west-facing glass in particular can be key for peeling off the peak cooling loads.

I like big windows, they are part of the quality of life in a house. But Dana is right, even the best windows have an R factor between 3-4. But also keep in mind that solar gain can be your friend and reduce your heatload in the same fashion it increases your heatgain.

At this point the windows are not negotiable, as the wife is very set on the big open floor plan with lots of natural light, hence the reason I selected high efficiency windows (if momma ain't happy, ain't nobody happy!). I dodged the skylights, but had to keep the windows as-planned.

I'll play with the coating on the east and west sides and see what it does. The ones we looked at had very little light loss but high radiant reflectivity, so as not to darken the house too much.

I figured all the windows with no outside shading unless they are under a porch or porch roof. During parts of the day we have good natural light filtration through the trees, but I didn't want to depend on that, as the trees are only ever a storm or bug infestation away from coming down.

Big windows properly scaled to the thermal mass for passive solar heating can dramatically reduce the heating energy use, but has a much smaller effect on the peak heating loads. The 99% outside conditions tend to occur before dawn after a long cold night, when it's perhaps more appropriate to talk about "lunar gain" than "solar gain", eh? :-)

Big windows with highly reflective exterior coatings often reduces the daylighting benefit beyond all reason. Yes, it kills the gain, but it also dims the view. Scaling back on unnecessary square footage using glazing with nicer & brighter visible spectrum is usually a happier compromise. Rationalize every square foot of glass, and minimize where you can, eg: Bedrooms aren't usually where you're looking the big sweeping views, but you still need enough egress area to meet code. Casement & awning windows offer more egress area per square foot than single & double hungs or sliders, and they typically have better air-tightness to boot. Similarly, swinging patio doors/French doors offer more access area per square foot of glazing than sliding doors. Peeling even 15% of the total glazed area can have a very real effect on the BTU/hr bottom line, usually without much impact on the quality-of-life issues. Where you fieel it's worth going big, consider the SHGC and U-factors carefully relative to which side of the house it's on along with the shade factors.

Glazing solar gain may be your friend in Buffalo, but most assuredly not in Atlanta.

If the spousal unit insists (against Dana's advice) on acres of glass, my advice would be to crank down the SHGC on east and west exposed glazing without regard for concomitant loss of visual light transmittance...the glazing acreage will overwhelm the tint.

We have two rooms (master and dining) whose high value exposure is west to a waterway. Each has triple 3050 Andersen 400 series double pane with sun tint, driving the SHGC into the 0.2x range, but also VT below 50%. Despite heavy tree shading, no worries - still plenty of light into those rooms. I wish I'd spec'd the same tint on the south and east sides. Even with plenty of trees in those directions, solar gain is noticeably higher and light levels such that curtains and shades are (too) often deployed.