Yeah, well... (sorry!)

I did just receive a sample 25psi EPS panel. It is really tough. This will go in the new pole barn for sure. Thank you and Rob for turning my head.

Miight as well hijack the whole thread and ask how best to insulate an existing pole barn with radiant slab of course.

Uh, what?

To make those deration numbers work you basically have to presume that all heat hits the wall and conducts laterally to the nearest stud and out with perfection, using only the R value of the outer inch thickness of the insulation. I do not believe this is accurate. If it was, every house in america with R30 walls would have undersized heat sources right now and frankly, I have a lot of homes like that out there which are performing at or better than expectation.

I do believe and understand that cross sectional averages are not enough by themselves as well, but in the "cavity spray foam" example I think it's closer to the truth than the R15 performance would be.

no doubt breaking thermal bridges is a good idea though.

Posted By BadgerBoilerMN on 04 Mar 2013 10:16 AM
I stand corrected. Just specifying some closed-cell foam over open-cell, which is in fact about twice the R-value. As for EPS vs. XPS, you have had me thinking for some time now. From the numbers it still EPS appears to be about 25% short of the equivalent inch of XPS--5.72 vs. 4.17 -- ( the ultra-fine point of off-gassing aside) and still a little structurally flimsy for my taste, but if the price is right, I will spec it on my next project (provided our staple guns will hold the PEX down long enough to pour the slab over it.

http://arlingtonpassivehouse.wordpress.com/tag/eps-v-s-xps/

Sorry but even I am not crazy enough to try spray foam under a slab ;-}

Posted By ICFHybrid on 05 Mar 2013 09:00 AM
You don't have to worry about the mesh getting trampled if you don't use it in the first place. Did you have pricing for the concrete without the fiber additive?

The thicker a slab is, the better it holds together. When you lay concrete, you are creating a composite. The composite isn't very good if the reinforcing parts aren't fully contained within the matrix.

You seem to be contradicting yourself, first you say leave out the mesh, then you say your are creating a composite including reinforcement. If I leave out the mesh and go with a thicker slab I'll end up with PEX stapled to the insulation and +4" of concrete on top - that's not good radiant design. Do you have a workable slab thickness and reinforcement design?

Thanks Dana for the clear "steer away from" polyiso under a slab.

Yours posts are well written and very informative. Do you have anything to add regarding my radiant slab thickness and reinforcement design?

You seem to be contradicting yourself, first you say leave out the mesh, then you say your are creating a composite including reinforcement

The following quote from your post indicates the use of rebar.
"(Probably overkill but going with mesh on 10M (3/8") rebar on 24" centers, AND fiber reinforcement." If you took out the mesh, it would presumably leave you with the rebar.
4" slab thickness. I like the #3 bar on 16" or 18* centers, where you make the bar spacing a multiple of your planned tubing spacing. For example, if you are running the tube on 9" centers, the 18" rebar spacing would be good.

When you put the bars on chairs, you can tie your tubing to the bars and the tubing ends up higher in the slab, where it belongs.

Posted By ICFHybrid on 15 Mar 2013 01:22 AM
The following quote from your post indicates the use of rebar.
"(Probably overkill but going with mesh on 10M (3/8") rebar on 24" centers, AND fiber reinforcement." If you took out the mesh, it would presumably leave you with the rebar.
4" slab thickness. I like the #3 bar on 16" or 18* centers, where you make the bar spacing a multiple of your planned tubing spacing. For example, if you are running the tube on 9" centers, the 18" rebar spacing would be good.

When you put the bars on chairs, you can tie your tubing to the bars and the tubing ends up higher in the slab, where it belongs.

Well that is interesting. Current thinking is #3 @ 24" with mesh ~ $440 / 1000 sq.ft.

I had run the numbers just to see what #3 @ 12" without mesh would cost, ~ $510 / 1000 sq.ft. - not a whole heck of a lot of difference. One issue for me is the mesh has to be small enough to come in through a doorway so I'd end up with lots of edges that would best be wired together. All bar might be a better solution.

My tubing spacing is 12", how does rebar @ a multiple of that work? If 24" does the PEX just span between and with enough ties doesn't get pushed down much between bars?

John

Posted By BadgerBoilerMN on 14 Mar 2013 07:47 PM
I did just receive a sample 25psi EPS panel. It is really tough. This will go in the new pole barn for sure. Thank you and Rob for turning my head.

Miight as well hijack the whole thread and ask how best to insulate an existing pole barn with radiant slab of course.

Is the EPS borate treated?

Where did you get the sample piece from?

how does rebar @ a multiple of that work? If 24" does the PEX just span between

Sure, but I almost always ended up with rebar at 16" - 18". Must have been a sweet spot on the engineering tables.

Mark your ICF walls going in both directions and on both ends with the bar spacing. Lowest guy on the totem pole can usually cut and drag bar once the edges of the slab are marked. If you cut carefully, you can poke the end of the bar an inch or so into the ICF foam on each side and that helps set the location and depth. About now, you can see that 16" is nice because it lines up with every other form web.....

The next thing you notice is that if you space the tubing at 10-2/3" it will line up with every other bar, so you need to make a field adjustment to your 12" spacing plan, by holding the tubing off places that you don't necessarily need the heat so that you will end up with the same amount of footage in the room as was planned.

And why do we need rebar again?

Milliions of feet of PEX and PB stapled to millions of feet of XPS with nothing but 6-6/10-10 wire on top "where it belongs" and nary a crack. Must be really, really lucky. Now we add fiber, which makes the wire obsolete, but for the protection it offers the tube from wheelbarrow traffic. It does make planking stand up a little stiffer also.

Posted By NRT.Rob on 14 Mar 2013 07:58 PM
Uh, what?

To make those deration numbers work you basically have to presume that all heat hits the wall and conducts laterally to the nearest stud and out with perfection, using only the R value of the outer inch thickness of the insulation. I do not believe this is accurate. If it was, every house in america with R30 walls would have undersized heat sources right now and frankly, I have a lot of homes like that out there which are performing at or better than expectation.

I do believe and understand that cross sectional averages are not enough by themselves as well, but in the "cavity spray foam" example I think it's closer to the truth than the R15 performance would be.

no doubt breaking thermal bridges is a good idea though.

The BSC guys simulate the 3-D heat transfer aspects of studwall construction, and the studwall assemblies listed in Table 3 on page 13 (.pdf pagination) come well within R2 (usually within R1) of dumbed down 2-D simple-math models, using a single U-factor for the framing, and a single U-factor for the cavity insulation. They modeled 2x6 16"o.c. 25% framing fraction with R19 full cavity fill at R13.7, rather than R12.8 that a dumbed-down 2-D model comes up with (assigning R1 for the combined R of the siding/sheathing/gypsum), which is only about a 7% error.

The ORNL whole-wall numbers use models validated on REAL ASSEMBLIES tested in their large scale climate simulator. The ORNL model yields a R25 for a clear-wall value on 2x6 16" o.c. studwall with polyurethane foam, but R17 for a whole-wall number, but mind you, that's for 5.5", not 5" of foam, which is a more realistic depth average. (Ain't nobody overfilling then trimming closed cell foam.  OK, so with 5" of foam might actually be something like R16 rather than R15, but it's nothing remotely like R30, (or even R20.)

The bigger the R/inch difference between the cavity fill and the ~R1/inch framing, the bigger the fractional reduction you get.

High-R foam penetrated 1/4-1/2" by a higher thermal conductivity is presenting %~50%  more surface area to the warmer side than in a full cavity fill- it's a heat sink, but it takes a 3d THERM model to get how much of a hit that really is.  Whatever it is, it's slightly worse than a 2-D model of the same assembly with only 1.5" of width exposed to the gypsum, but not a huge amount worse.

If there's a failure in accuracy in the models to explain better than anticipated performance, consider that the U-factors in heat load tools are constant, not modeled for the materials, and the WAG usually thrown in for infiltration/ventilation rates, is usually just a WAG, and even when blower-door tested, the "ACH-natural" models based on ACH/50 numbers are a JOKE, since the location(s) of the actual leak(s) can render the estimate wrong by more than 50% (in either direction!). 

Then there is the extremely non-linear performance of low-E windows: Window performance improves dramatically with ever-higher delta-Ts (say, at the 99% outside design temp) since the radiated fraction of the heat transfer (which is most of it, with low-E tight windows) is a function of the differences in the fourth-powers of their absolute temperatures, and the emissivity & reflectivity of the low-E materials at the blackbody spectrum of the radiation temperature(s).  Assigning a single U-factor is making linear approximation of heat transfer per degree-delta-  picking the slope at a somewhat arbitrary (if agreed upon by convention) point of that extremely non-linear curve.  (Google "Stefan-Boltzmann law", or if you need the thermo-101 refresher, search that name on this site. ) 

The R-value of  foam (or fiberglass, for that matter) itself is also not constant across temp & delta-T which adds it's own nonlinear twist to the model, but it's small compared to the performance shifts of low-E coatings within relevant delta-Ts.

Bottom line, heat load tools use only linear-approximations and 2-D models, and use only single U-factors for any of it, don't attempt to model the non-zero thermal mass effects of timber-framed buildings, etc. While it's good enough to design a heating system around, it's not a high precision model.  Real world performance usually does outperform what Manual-J methods would suggest, for any number of reasons.

But, whatever- I'm standing by the, "Dude, that R30 is only R15!" whole-wall estimate of closed cell foam in a 16" o.c. 2x6  25% framing fraction wall.  That's only ~12% under the R-value the folks at Oak Ridge come up with at a full cavity-fill of close cell foam, and the real world never sees a truly full cavity fill of closed cell foam.

Posted By Lbear on 16 Mar 2013 01:33 AM

Posted By BadgerBoilerMN on 14 Mar 2013 07:47 PM
I did just receive a sample 25psi EPS panel. It is really tough. This will go in the new pole barn for sure. Thank you and Rob for turning my head.

Miight as well hijack the whole thread and ask how best to insulate an existing pole barn with radiant slab of course.

Is the EPS borate treated?

Where did you get the sample piece from?

Most Type-IX EPS is usually rated 25psi, but I don't know of any Type-IX EPS treated with borates (it may exist though.)  MN isn't exactly termite country, and if the ants in the 'hood haven't been eating 1.5lb XPS they won't be eating 2lb (Type-IX) EPS either.

This is true. No termites around here. They don't like the snow...

dana, your ORNL calculator includes 14% losses through a theoretical join to a slab on grade and another 10.5% for the roof connection, and I have no idea what "partitions" have to do with anything (another 24% of wall area) unless they are assuming you don't insulate the wall at the connection to the outside wall? That seems a little odd.

that's no 25% framing factor at any rate. the ORNL calculator is basically assuming your wall is half wood. Maybe that's typical construction... I do have a dim view of typical construction... but it does seem a bit... conservative.

Hey now. That is a good question. I am still trying to figure out how 2.5"x16=25%?

Posted By NRT.Rob on 19 Mar 2013 11:37 AM
dana, your ORNL calculator includes 14% losses through a theoretical join to a slab on grade and another 10.5% for the roof connection, and I have no idea what "partitions" have to do with anything (another 24% of wall area) unless they are assuming you don't insulate the wall at the connection to the outside wall? That seems a little odd.

that's no 25% framing factor at any rate. the ORNL calculator is basically assuming your wall is half wood. Maybe that's typical construction... I do have a dim view of typical construction... but it does seem a bit... conservative.

And yet the ORNL model retains remarkable correlation in whole-wall R with the 25% framing fraction 2x6 R19 wall modeled by Building Science using THERM (R13.65 whole-wall per ORNL, R13.7 per BSC's THERM modeling of a somewhat different structure). It's in the range, and it's in the range of simple 2-D models as well.

Adding an inch of exterior EPS to the R19 wall in the ORNL case delivers R16.1 whole wall, which is comparable to the R17.3 full-fill polyurethane. (But R16 is still more likely at something actually buildable, with 5" rather than 5.5" of cc foam.)  BSC did not model adding an inch of foam to the 16" o.c. 25% framing fraction wall, but did model adding foam to the 16% framing fraction OVE wall.

A 25% framing fraction is typical, and would even be LOW in seismic zone areas that also required fire-blocking at mid-cavity (many location in CA have 30% framing fractions as typical.)  The 16% framing fraction BSC uses for the advanced framing/OVE simulations is actually quite difficult to achieve, even with 24" o.c. stud spacing.  But even at 20% framing fraction (usually deliverable at 24"' o.c. stud spacing without bending over backwards) it doesn't add more than about R1 to the simple-math model.  Assuming 5.5" of R6/inch foam & R1/inch wood, with an R1 allowance for the siding gypsum & sheathing at a 25% framing fraction a 2-D simple model delivers R15.67, dropping that to a 20% framing fraction delivers R17.5.  Even using a barely-buildable 16% framing fraction only delivers R19.33, all at a center-cavity R of  R33.   A more careful modeling using THERM or the ORNL methods might bump that another ~10% but not much more. Call it R18-R19 whole-wall best-case, R20+ only in your dreams.

There's no way around the thermal conductivity of the wood short of insulating over the framing.  Framing fractions cannot be infinitesimally small in real houses- the kind with windows & doors. With closed cell cavity fill, no exterior foam, R20 would be a best-case unbuildable whole-wall level.  But with 2" of exterior iso, even derating to R5.5/inch for zone-6 MN mid-winter temps and a cellulose cavity fill you'd be north of true R20 whole-wall performance, and  it would have sufficient exterior R to skip the interior vapor retarder, resulting in a more resilient assembly that can actually dry, all for less money up front than 5" of 2lb polyurethane foam cavity fill. It would have less than 1/2000 the global warming potential to boot.

Posted By BadgerBoilerMN on 19 Mar 2013 12:04 PM
Hey now. That is a good question. I am still trying to figure out how 2.5"x16=25%?

You have no narrow cavities, no stud plates under the studs, no window & headers (maybe even doubled), no top plates (maybe even doubled), only single studs at door framing- how did your house ever meet code? 

And it's a division, not a multiplication, and the studs are only 1.5" in most houses..

1.5/16= 9.4%, which is referred to as the "clear wall" R.  In a tall windowless doorless long industrial building wall it's possible to get to about a 10-12% framing fractions, but not in houses.

I'm polluting my own mind on this... we've always done headers and such by expanding our "window" areas to cover the entire rough opening and trim and areas between close windows where additional support is likely. that glosses over SOME of this which is probably why we don't come up short more often... it's unlikely a really conservative ACH in our particular case as we are not particularly conservative there.

but, even if we assume windows and doors are taken care of, I still don't get the % of wall area they are attributing to partitions, floor, and roof. I get those may have doubled top plates and such, but as a percentage of wall height these numbers are not clicking for me at all. even 3" of double top plate is at most 5% of an 8 foot wall. there must be some factor for the actual top surface or something in play here. I hate to nitpick but that plus the slab connection (assuming uninsulated slab) account for a significant amount of deration here.

Pick apart the BSC THERM modeled assemblies, which are explained in greater detail.

Bottom line, I trust these folk's attention to detail and field-measured reality on framing fractions.