Anyone have any thoughts on this one?

BSI-064: Bobby Darin and Thermal Performance

By Joseph Lstiburek
Created: 2012/10/08

This all started pretty innocuously. I just wanted a client to have a warm floor. How complicated could that be? Well some seemingly innocuous code language made things irritating but not difficult and I sucked it up. But then the LEED fascists made things difficult and unworkable and finally I said enough is enough. It is time for the old guy to strike back (OK, I am not striking back, I am whining. But I have it on good authority that whining is a form of therapy. Apparently lots of folks I know are in this process of therapy).

When you have a bedroom over garage in a cold climate builders typically get complaints of cold floors. What a surprise.  It is easy to construct a warm floor. The best way to do it is shown below. Continuous rigid insulation is used across the underside of the floor framing. Notice the air space above the cavity insulation. That airspace makes the floor warmer than if the cavity was completely filled. Huh? Yes, the airspace makes for a warmer floor. Filling the airspace completely reduces heat flow and saves energy but does not make the floor warmer. Old guys taught us this.

So what is the problem?

The code is the first problem. The code states that I must install the insulation in direct contact with the underside of the floor. Fortunately, most building officials are not morons. In fact many of them are very smart. You can actually talk to them. Good luck with LEED raters. Judgment is not allowed with LEED. I have no trouble with the code folks as they have judgment - so doing it “right” is not a code violation. Different story with LEED. LEED says you have to follow ENERGY STAR. EPA ENERGY STAR says that you must “align the thermal barrier with the air barrier.”

Views: 752

Reply to This

Replies to This Discussion

Well, I am curious about the detail in Figure 8

I wonder how well the detail would work during the cooling season in a humid climate?

Is the "tub" upside down during the cooling season?

and what happens if the sides of the "tub" are not-so-perfect?

Where do you see the problem? Condensation on the underside of the podium slab? At least the air gap would allow easier evaporative drying rather than keeping the moisture in the batts if they were tight to the concrete.

What I don't understand is why it would cost $100,000 more to insulate the underside of the slab, and why not use spray foam with intumescent coating. It must be a very big building.


I was "wondering" if  the orientation of the "tub" might make a performance difference depending on the season.

The "tub" analogy reminded me of John Straube's Cylinders.(BSD-014)

In Straube's example...if he had not inverted the cylinder before bringing it back indoors... the cold air would have fallen down and out instead of being "contained".

I see what you're getting at - that the cooled air under the slab would "fall" through the suspended ceiling and not be contained.

I think Joe's response, which he mentioned in the article, is that there is no significant stack effect in shallow horizontal floor and ceiling spaces, and all the delta-P is across the slab-as-air-barrier.

I do suspect, however, that there could be potential for condensation on the underside of the slab, depending on how cool that surface becomes with the thermal lag of the concrete and what appears to be relatively little insulation below the floor.

Nice Robert, I think you uncovered a potential source of some confusion.  

 there is no significant stack effect in shallow horizontal floor and ceiling spaces

Stack effect is not simply the height of the structure.  It requires a connected air path from bottom to top.  The weight of stack requires the whole column to build upon itself.  If you have a 10 story building, but penetrations between floors are completely sealed you have broken the column.  You basically have 10 1 story boxes when it comes to stack pressure, not 1 10 story column.  

What Joe appears to be saying is losses through porous insulation in unsealed cavities via increased convection are maybe not terribly significant.  Just significant enough to improve comfort when designed thoughtfully.  

The big loss is from air flowing in and out of that space, carrying the btu's completely away to someplace else.  Solve that leakage, and you can get creative in your placement of insulation.  Make the floor comfortable and you keep humans away from thermostats.  

I can see how easy it might be to have this misunderstand of stack, make some attribution errors, and therefore see this suggested design as a complete departure from best practices.  I guess this would not be something for building science 101 as it could lead to a fair amount of confusion. 

It seems to me that the attic bathtub concept would not work so-well during the heating season...
In the same way that the crawlspace bathtub concept would not work so-well during the cooling season.

Of course, that's the picture of virtually all insulated attics. The delta-P across 12" of batts in your scenario is about 0.25 Pa. minus the resistance of the batts. The significant issues are air sealing at the ceiling (that's why we call it a "sealing"), and windwash through the batts if the sides of the "bathtub" are not well-sealed.

Dr. Joe's designs are always "perfect", so there's never any problems. ;-)

The idea being the delta p caused by 65-30 delta t would be enough to overcome the resistance of the batts?  

Wouldn't it have a modulating/self-balancing effect?  

In other words; if the point where delta p did NOT overcome the resistance was 30 delta t, wouldn't you expect that to be where equilibrium settled?

Early studies of (very) loose-fill fiberglass in attics found enough vertical convection within the insulation to significantly undermine the insulating value of the fiberglass, but the same was not found with batts (except for poorly-installed batts with gaps), and as little as 2"-3" of cellulose on top of loose fill fiberglass was enough to virtually stop internal convection.

So it's really not much of an issue with modern materials and techniques in horizontal applications.


during heating season....

Isn't the 2"-3" of cellulose that you mention acting like a continuous cap that retards the cold air from spilling into the 3-d network of voids(and displacing the warm air)?

and isn't your example a "better" air barrier than a lay-in suspended ceiling ?

No, the cellulose cap prevents warm air from rising up ;-)

Oak Ridge National Lab ignored the silly debate about whether cold air falls or warm air rises and simply ascribed the loss of R-value to "convection" (which is both processes occurring simultaneously).

Whether loose-fill cellulose is a better air retarder than lay-in suspended ceiling tiles is anyone's guess, but fiberglass batts did not suffer from the R-value decline.

And what their testing showed is that the stack-effect convection that is not normally a problem in horizontal applications with low "head" can become so with very high delta-T.

Fixing Fiberglass Convection Problems with Loose-Fill Cellulose

Energy Design Update Volume 13, No.5; May 1993


Researchers at Oak Ridge National Laboratory are exploring ways to prevent or fix the problem of air convection in low-density loose-fill attic insulation.

Previous testing at Oak Ridge showed that low-density (0.5 pounds per cubic foot [lb/ft3]) loose-fill fiberglass suffers up to 50% loss of R-value at very cold attic temperature (-18°F) due to air convection within the insulation. However, laying fiberglass batts over the loose fill effectively stopped the convection problem. What if the low-density material was covered with higher-density loose fill instead of fiberglass batts?

To answer that question, Oak Ridge scientists tried blowing both cellulose and 0.7 lb/ft3 fiberglass on top of low-density R-30 loose-fill fiberglass. The results were mixed: cellulose worked, fiberglass didn't.

When approximately two inches of cellulose (R-8) were added over the R-30 low-density fiberglass, the measured overall R-value remained at R-38 down to -18°F, indicating that the cellulose covering effectively prevented convection.

But the higher-density fiberglass apparently failed to stop convection at low temperatures. When an R-8 layer was added over the low-density material, it added R-8 to the overall R-value, but failed to restore the lost R-value due to convection.

Thismakes so much sense to me. At some point delta t creates enough delta p to "break" or overcome the natural tendency of things to move toward equilibrium/mix and instead create/move toward stratification.



  • Add Videos
  • View All


Latest Activity

Shaun posted a discussion

Commercial duct testing

Hello all,I am a BPI BA, HERS rater and QC inspector. I cant find a ton of information on this…See More
6 minutes ago
Dennis Heidner replied to Dan Antonioli's discussion Net Zero Energy Hot Water
"Dan, Solar PV and solar H2O work in northern climates...They can still provide the desired goal in…"
13 hours ago
Joseph Frederick added a discussion to the group Energy Auditing Equipment for Sale, Trade or to Purchase

Blower door Equipment for sale

Minneapolis Blower Door model 700, DG with cruise control. Hoses and manuals, door frame. Used once…See More
16 hours ago
Dan Antonioli replied to Dan Antonioli's discussion Net Zero Energy Hot Water
"For my particular application the heat pump heater would be integral to the tank and thus be…"
19 hours ago
Dennis Heidner replied to Dan Antonioli's discussion Net Zero Energy Hot Water
"What climate zone are you in?  Where would the hot water heater be set?  Northern…"
20 hours ago
Larry Weingarten replied to Dan Antonioli's discussion Net Zero Energy Hot Water
"Hello:  Looking for solid numbers,you might want to search "Martin Holladay solar thermal…"
21 hours ago
Profile IconGabriel Lieb, Beth Neri and Aaron Beck joined Home Energy Pros
21 hours ago
Kim Tanner posted a discussion

Game of Homes

Have you seen the new online game by EnergyLogic, Game of Homes? If not you should check it out!…See More
23 hours ago

© 2016   Created by Lawrence Berkeley National Laboratory.   Powered by

Badges  |  Report an Issue  |  Terms of Service