Exterior foam may have it's effective r value seriously degraded by metallic fasteners passing through it!   I just did a quick calculation with #10 steel screws 6" apart into studs 16" oc.  The results surprised me, r5 insulation is degraded to effectively r2.55.   I then tried the same screws with 12" fastener spacing, this reduced r5 to r3.4.  Then I tried r10 with 12" fastner spacing & r10 becomes r 5.1.  I used 1/650 for the r value of steel, this is in the middle of the conductivities I found.

Hmmm, at first glance I get some different results, but somebody please double check me as it's late and I'm doing the calcs in Excel rather than Mathcad and probably buggered up the units. Perhaps you didn't take into account the relative areas of the two materials? Thermal resistance (Rvalue/X) = 1/Rvalue = x/(k*A), where k is thermal conductivity, A is area,and x is the thickness.

For steel, k = 43 W/mK
For EPS, k = 0.035 W/mK

Assume 1 ft2 of EPS
Assume 1 fasteners per ft2 (it's probably less than this based on the stud spacing) each with a diameter of 1/8 in. (not sure what dims a #10 screw has)
This works out to an area for the fasteners of 0.001 ft2

Assume 1 inch for both for the thickness in the direction of heat transfer (through wall).

Thermal resistance for the EPS (x/k*A) = 7.8 K/W
Thermal resistance for the screw = ~78 K/W

Therefore, for the same delta-T across the wall, the EPS will transfer ~10x the amount of heat as the fasteners

Thanks for bringing this up, as it's worth evaluating. Someone let me know if these numbers look right.

Let me see if i can explain the differences it the 2 estimates.

#10 screw has a major diameter of 0.19"  so area of screw is 2.31 times the 1/8 Sonslidr 74 used.
Which would change the thermal resistance of the screw from 78 to 33.76

1 fastener every 6" into studs 16" OC works out to 1 fastener per 2/3 square ft.
Which would change the thermal resistance of the EPS from 7.8 to 11.7

Using Sonslidr 74's conductivities would give steel an r value of about 1/246/" 

If I redo my average r value for the first case with the corrected (?) r for the steel the r5 with fasteners on 6" centers is equal to r 3.67 without fasteners.  Still way more than I'd have guessed.

It looks like the two different methods agree when we start with the same data.

I think it is rather startling that, in a wall with 1" of EPS, 1/4 of the heat transfer would be through screws every 6" into studs 16"OC.

I would consider 1) using glue to reduce the number of screws and 2) using thinner, longer, high strength screws. Perhaps ones with large threads (for grip) and then a narrow shank and then a large head. I would use the minor diameter for calculations (about 1/2 the loss).

Interesting discussion! Just to mess with your minds, what happens when we take the backer material into account. Assume glass/cellulose insulation behind the foam and wood stud behind the screw as well as foam around the screw. As the metal is highly conductive it transfers the cold quickly but it can only loss its heat/cold to the stud or the surrounding foam. Does this mean the actual heat loss/gain for the screw would be the same as wood (assuming wood is lower than foam.) Would the screw be at its stated values if it was penetrating a steel stud?

Say you are attaching furring strips that also hold the foam to the studs. Countersinking the screw head into the furring strip and filling in the hole would make a big difference.

A builder in the Portland/Seattle area does just this. They are call Hammer and Hand.

Posted By jonr on 26 May 2012 01:47 PM
I would consider 1) using glue to reduce the number of screws and 2) using thinner, longer, high strength screws. Perhaps ones with large threads (for grip) and then a narrow shank and then a large head.

Fewer screws, yes, but you would want to minimize the area at the end of the screw, for less heat transfer there. Perhaps a very narrow head screw could be used, but with a plastic or other minimally-conductive washer under it for mechanical load distribution.

Stainless steel screws or nails conduct significantly less heat than steel. But all of this analysis assumes that the fastener is exposed on both ends and there is no R value from an air film. Neither is true in actual use. This makes a huge difference and changes it to "nothing to worry about".

Now if we are talking a bolt through a steel SIP...