[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: RE: GSBN:Re: Question re R-values

Joyce, et al...

I think the discussion is getting down to what is most useful. When I
talk about the ORNL testing, I usually say that the bale wall we tested
in that lab had an R-value roughly double that of the 2x6 stud wall with
perfectly installed R-19 fiberglass batt insulation (+/- R-30 for the
bale wall to +/- R-14 for the fiberglass insulated wall. That's excellent

But I typically also talk about the thermal performance of wall
assemblies versus the R-value, because R-value alone is only part of the
story. Thermal mass and its location and relationship to the insulation
has a lot to do with the overall performance of an assembly. Different
assemblies with exactly the same R-values can perform quite differently
depending on the configuration of the wall.

My typical example is a concrete block wall furred out on the inside and
insulated with fiberglass batts and drywall, versus the same concrete
block wall with an equivalent amount of insulation on the exterior. Same
R-value but one has mass interior to the insulation and the other leaves
the mass out in the weather. The performance is much better with the
externally insulated mass.

Taking that a step further and including radiant comfort - your body's
sensitivity to radiant heat, bale walls' performance is actually
significantly better than the R-value alone would indicate because the
interior plaster represents a lot of thermal mass pretty well isolated
from exterior temperature changes - so the plaster tends to stay at
ambient air temperature in the room. Thus your body is not trying to heat
a cold wall or be heated by a hot wall and so the you can be comfortable
with cooler air temps in the winter and warmer interior air in the
summer. This can make a big difference in heating and cooling loads and

David Eisenberg

>I would be careful about quoting the Sandia Labs' report.  It had some
>serious limitations.  Also, be careful with interpretations of ORNL's
>tests.  ORNL has develoed a term they call "effective R-value."  While it
>does do a better job of accounting for extra lumber in the wall and the
>lost insulation value of spaces left unisulated, it unfortunately also
>mixes infiltration with thermal conductance - what the R-value is supposed
>to measure.
>However, you are right that a simple comparison of fiberglass' R-19 to
>straw bales' R-30 (or R-33) is not appropriate either.  Recent research
>sponsored by the Energy Commission indicated that an "R-19 wall" as it is
>generally constructed, only gives about R-12.  If extra care is taken to
>make sure that insulation surfaces are flush to siding and sheetrock
>surfaces, that insulation surrounds all wiring, that it fits tight around
>switch and outlet boxes, and that it completely fills all bays, then the
>"R-19" insulation will give you about R-14.  Whoopee!
>On the whole, I agree with your approach of avoiding the simplistic answer.
>However, I have also found that some people are more than a little turned
>off by a technical answer to what they thought was a simple question.

"Not everything that counts can be counted,
and not everything that can be counted counts."
- sign over Albert Einstein's desk at Princeton