Dana1,

How thick should the XPS be for decent thermal bridging? Would it go between trusses and OSB, or between OSB and drywall. It goes between trusses and OSB, would it have to be full sheet? Or could it be cut into strips and I could use something like carpet tape to temporarily hold it before the OSB gets screwed?

Thanks.

V,

I would recommend doing something to reduce the thermal bridging in the ceiling. I have used Thermabloc on the inside as well as on the outside of the studs and joists. They shipped the product in a box. You can put a large number of ¼ “X 1 ½ “X 48 “sticks in a shippable box. They are very light. They have a double sided tape on one side that you peel and stick on to the joist.

You have eliminated most of the thermal bridging with the 1" on the outside. I always use at least R-10 thermal bridging in new designs that stops 90% of the problem. Using the Thermabloc on the walls would ½ the the remaining conduction I would have to weigh the costs to determine whether it is worth the costs in that case. XPS strips are less expensive but take a 1” of space and require longer fasteners because of that fact. I have also used XPS and it worked well though you have to cut and staple them in place which takes a little extra effort.

I changed to a less expensive method in the last couple jobs that I designed. I used 1X2 firing strips across the bottom of the trusses orthogonally 16 inches on center. At the interface between the firing strip and the bottom chord. I placed 1 inch blocks of foam there. This suspended the the ceiling dry wall with an insulator and reduced the potential conductive path by a factor of 10 as well. We then sprayed 2# foam on top of the drywall both gluing it to the chord and sealing it completely. I think this is a better method overall. I am the only person I know that had done this method. It has several advantages; a 90% reduction in conductive path and R-4 insulator in path. That reduces total conduction by 95% plus. That is equivalent to an R-20 thermal block of the conductive path. The FEA analysis of that embodiment shows it provides the most performance for the cost of systems that I have modeled in using standard type framing. It will lower your ceiling 1.75 inches. This is typically not an issue. Your ceiling drywall will be flatter with 16” on center attachment. If you attach the strips 8 inches on center it will eliminate the insulation load issue for the drywall as well while reducing the thermal blocking somewhat less.

My window comment was not directed at your window but simply trying to communicate that you should fix the least energy efficient things first and that there is a hierarchy to follow in getting the biggest bang for the buck.

Brian

Thanks Brian. I got a quote on price for the thermoblok strips. About $1300 to add that to my roof, but that may be worth the investment since it there would be very little additional labor other than to peel n stick!

But until we get our appraisal done and can get an actual loan amount, I can't be certain that I'll be able to include it in my project. I have about 1200 linear feet of truss chords to cover, so not sure how equivalent that would be with using your method for cost. My problem is that with not having a double bottom plate, and adding a dry sandwich radiant, I am already taking 1.5" at best from floor height. Another 1.75 inches off from the ceiling and I am definitely wasting drywall with cutting off 4" along each sheet to fit! The nice thing about the thermablok is the thinnes of the material!

I'll post again once I have made some progress.

Let's really look at the ALLEDGED thermal bridging in an r35 loose cellulose insulation sitting on a ceiling supported by trusses.  Let's use r1/" for wood and r3.2/" for cellulose.  35/3.2= 11" .  Let's first calculate the r value at he truss, 3.5" is wood (bottom chord) and the diagonal braces take up less than 1/8 of the area and have an average angle of about 45 degrees so the path length is (11-3.5)*1.4=6.1  so the r of 1/8 the area is 3.5+6.1=9.6 and 7/8 of the area is 3.5 +(11-3.5)*3.2= r27.2.  (7/27.2 +1/9.6)/8=u=.045, r=1/u=22.2   but the trusses are 1.5" of every 24 with the remainder being the r35 we started with.  so       1/(1.5/22.2+22.5/35)/24  =r33.78  The simplest, & by far the cheapest, way to address this small bit of thermal bridging effect is to to add another 1/2" of insulation depth!!!!!  Needless to say the effect of the thermal bridge gets smaller with greater insulation depth & at r60 it's less than 1/4" of depth.  Now if the ceiling were supported by 2x12s on 24" centers we'd have a real issue.as the effective  r would plummet from 35 to 29.7. or even worse 2x 12s 16"OC which takes it down to r29 equivalent.

Liebler,

The total R-value approximation is not an accurate method for calculating conduction in a thermal bridging member. In the case you are looking at the stud is a relative conductor compared to the cellulose. The stud is directly connected to the top and bottom of the delta T from hot roof to drywall. The energy conducted down the member is not reduced by the insulation surrounding it. The fact is that the insulation around the conductor increases the thermal energy left to conduct slightly. The approximation you are using is for distinct different layers of material each as a separate layer where one material does not thermally bridge the other materials.
In a to code home with a standard truss type roof one of our national labs reported that 25% of the energy lost from the standard home is through the roof structure and further states that 60% of that energy is through conduction. This energy loss was reduced by 40% by adding an R-2 surface insulation material on the studs only in their test. Real world, real numbers, this was not an approximation it was a hot box test experiment. Thermodynamics is significantly more complicated than the simple approximation equations most use to “calculate” the answers.

Additional cellulose insulation will only address the 40% side of the problem.

Can you get a link to that study?

Zen, 
  Please give me a source reference.  The data you are citing must have some assumptions or conditions that are unusual to say the least.   I am willing to learn but I have a BSEE and sure understand both 2 and 3 dimensional electric curent flow which is perfectly analogous to heat flow.  I stand by my calculations!  The increase of ceiling heat transfer due to the presence of wood is no worse than stated.   For a ceiling with 2x4 wood trusses 24"OC 11 1/2" cellulose gives r35 + with thermal bridging FULLY accounted for!

Zen,
    Your assertion that I'm using the incorrect formulas is incorrect; resistances (r) in "series" add, conductances (u) in 'parallel' add, resistance is the inverse of conductance and conductance is the inverse of resistance!   These calculations are simple resistance/conductance network analysis (freshman EE101).  The 'bridging' is simply paralleled thermal conductances.    BTW your "solution" of adding r 2 to the wood results in raising the whole assembly from r 33.78 to  r 34.05  equivalent to adding 0.08" of cellulose!

You are not applying the equation correctly because you are making bad assumptions.

The boundary conditions of the problem are not being properly addressed by the simple method you used. I also am an engineer, and have spent 30 years working on self-heated building design, analysis and construction. I regularly perform dynamic FEA thermal analysis of systems. Let’s look at the example but go to the absurd. Let’s make the conductor a steel I-beam connected on one end to an infinite heat sink. Does insulating the I-beam increase of decrease the thermal energy delivered?

The same is true of a copper water pipe. Is the temperature of the water reduced because you insulated the pipe more? When you add insulation it does not reduce the energy that is transmitted down the conductor. It retains the energy for conductance. A 2 inch foam gap in the path will stop most of the conductance. Insulating around the member will not. Also remember that the attic temp in a standard home is lower than the contact temperature of the roof material. The roof contact temperature can exceed 150 F while the attic temperature does not exceed 90. That means that one delta T is 75 F while the other is 15 F. The thermal energy transmitted down the solid member with no thermal bridge is not significantly affected by additional insulation around it. The energy though the insulated portion is reduced, but that makes the left over losses proportionally greater in the conductor which did not change significantly as you added insulation. If you insulate around the truss to 30 inches at R-113 you have eliminated most of the conductance through the insulation. At that point almost all the residual losses are through the truss. Small insulation layers at either end or a break with an insulator in the joist are the only ways to stop the left over heat transfer.

Yes most of the heat flows down the path of least resistance but some flows in the other path as well!  when you "insulated" your example I beam you replaced the previous surroundings with insulation, if the previous surroundings had been a copper pipe with the same cross sectional area as the i beam the total heat transfer through the parallel paths would certainly be reduced, to LESS THAN 1/2 because copper's conductivity is greater (lower r).  You are almost correct adding insulation doesn't reduce the heat flow through the bridge ,very much (see next paragraph regarding path length),  but it reduces the greater heat flow through the far greater area which has a lower conductivity  A needle conducts less heat than an I beam the ratio is the area conductivity product. Adding a second I beam will certainly increase the heat flow, or add a second I beam with 1/2 the area of the first will the heat flow be increased? You are failing to realize that ALL area is conducting in the ceiling example, some areas are more conductive than others we are not adding insulation around a pipe we are dealing with parallel paths with a conductivity per unit area for each path!  Twice the area with 1/2 the conductivity will conduct EXACTLY the same amount of heat!  The wood (thermal shunt) represents 6.25% of the area at the bottom of the chord and 0.8% of the area at the top. Yet the wood to insulation conductivity ratio is only 3.5 Your " fix" of adding r2 to the wood which is already r9.5  changes the conductivity of the much smaller area (6.25%0)  to 82% of what it had been.

As the insulation depth gets greater the path length for the wood grows as well so YES the heat transferred through the wood is REDUCED.  Since  beyond 3.5" the path through the wood is up the angled members the path length grows at about 1.4 times the depth change.  In adding the 19" to go from r35 to r113 the r of the wood paths has had between r19 and r26 added to it since it started at between r11 and r14  the heat transfer through the wood  is between 0.56 & 0.57 of what it was with 11" depth.  while the heat flow over all is 0.31 of what it was.  So yes, adding insulation has forced a larger fraction of the heat transfer to be through the wood.
 

Long ago, when I was a college student, an engineer was defined as: One who uses a slide rule to multiply 2 times 2 and comes up with 3.976 but says we'll use 4 because it sounds more probable.  The point being, engineering  involves using flawed tools to approximate reality then rounding to useful results.  In analyzing the effect of the solid wood in the cellulose insulated ceiling I used a 2 dimensional analysis to APPROXIMATE a 3 dimensional world.  Zen is correct, I changed the boundary conditions.  To eliminate heat flows in the third dimension.   I, in effect, made  the 'lateral' conductivities of the upper and lower boundaries infinite.  This makes all temperatures on either surface equal.  This does introduce some error!  Zen's mention of boundary temperatures made me appreciate this source of error.  While, without, a rigorous 3 dimensional analysis, we can't quantify the error,  we can establish it's direction!  We can, establish that my method, over estimates conductance and under estimates resistance.  Simply the surface resistivity is always greater than zero, as assumed, so all heat flow paths will have resistance added to them.   ie,  The actual r value will  ALWAYS be GEATER than I've predicted!

Umm, can I squeeze into this engineer pee'ing match and request the white paper from Zehnboss please?

Is one of you two saying that the chord members are acting as a conductor and supplying heat from the roof deck down into the home?

I agree where is the study?

The comment that in a code house 40% of cooling and heating loads in through the attic is not correct. I had an energy audit when building and the baseline house had no where near 40% loss in the attic - it was more like 15% of heating and 5% of cooling. Maybe an older house with r-13 batts between joists had 40% loss there.

In my specific example, R30 is code and upgrading to R40 resulted in savings of $50 a year. A great ROI but not the kind of numbers you are quoting. In the end, the energy audit overestimated heating costs by a lot so I suspect the upgrade to R40 really only saves $30 a year but still a decent ROI.

Sure - their assumptions are wrong and so are yours.

I've been away from this forum for a long time and as I come back, I see a lot of extreme viewpoints and things not supported by reality. Sure, code should be stricter. Sure, we should be forced to build with 50 year ROIs. But to say that a code house is necessarily an energy hog is crazy - an old house is an energy hog. An SUV is an energy hog. The fact that people commute with pick up trucks - that is the real energy hog.

My conventional house with basic air sealing and just above code attic insulation. 5300 SQFT. $400 to heat and $600 to cool a year. Central NC. My biggest issue is that the backyard/view is East. The house has 20+ windows facing that way and the new trees aren't there yet. And that is a far bigger issue than thermal bridging in the attic.

SUV, 15k miles a year, $3500 a year.

And I don't know, but I always figured that the thermal bridging in the attic was pretty much taken care of by blown insulation over the joists. Sure - it really diminishes your R value at that point but the easiest thing to do would upgrade the blown insulation. I feel like I have 20 inches with 12 inch joists. So yeah, at the joists, I only have 8 inches. It isn't like having nothing like in a conventionally built wall...

Posted By David_Cary on 25 May 2012 05:42 AM
I agree where is the study?

The comment that in a code house 40% of cooling and heating loads in through the attic is not correct. I had an energy audit when building and the baseline house had no where near 40% loss in the attic - it was more like 15% of heating and 5% of cooling. Maybe an older house with r-13 batts between joists had 40% loss there.

Even in that older house it's hard to come up with 40% heat loss via the roof/attic, unless one is counting on extreme air-leakage through the ceilings.  Now that the IRC includes ACH/50 leakage limits, there's almost no way to come up with 40% of the total loss via the attic assuming everything is code min. (Maybe with a code-min attic/roof and high-performance envelope in every other regard you could get there.)   With any kind of reasonable air-sealing and insulation efforts  even on older houses it's under 20% of the total for most. 

It doesn't many iterations of Manual-J type heat load calcs (crude a model as that is) on multiple homes to conclude that in code-min houses the windows will always account for more of the heat loss/gain than roofs.   But from an envelope upgrade cost-effectiveness point of view it's usually cheaper in $/BTU-peak to improve the attic insulation performance than the windows as first-level shot at reducing loads.

There will be exceptions to prove the rule, especially regarding cooling loads. All the attic insulation in the world won't make up for the heat gains of even a modest amount of west facing window area with code-max U & SHGF factors in a cooling dominated climate.

Great discussions. Thanks!
I'd love the idea of stopping all thermal conduction in my attic, but unless the scratch tickets come through, I doubt I'll go that extra step. In the walls i tried to do a little something with the iso on the exterior.

I now have a question on the issue of using OSB to support cellulose in the attic. One drywaller is concerned that the OSB will create bumps and waves in the ceiling that will be really noticeable. My internal reaction is an exasperated give me a break! But he is the expert in drywall, and if it will lead to waves and bumps, i'd like to know if there is a proper way to do the OSB to avoid that.
Assuming my trusses are perfectly level on the same plane and all 2x4's used on the bottom chords are perfect then having no OSB will mean a perfectly smooth ceiling. I put a 6 ft straight edge against said perfect trusses and they are not perfectly level. So why would an extra layer of OSB add more bumps and waves? I went through the home shows and most ceilings look ok, and I'm sure their trusses are just as level as mine. Though they did have quite a bit of texture. Maybe that's how you hide that.

Anyone have experience with OSB creating issues and bumps and waviness? Sorry if I sound frustrated but I'm tired of getting these weird responses to my simple request for a bid to do R100 in the ceiling. In the end I'll be happy if I can find someone to give me R60, and most want to talk me into fiberglass.
I had assumed the OSB sheets would go parallel to longest trusses (I have hip roofs so around corners trusses go both ways) then drywall would be joins offset from OSB. one question I do have is would the drywall orientation go perpendicular to trusses as well and screwed to trusses or just OSB since it might be difficult to make sure they hit the trusses without being able to actually see them behind the OSB, or even if they go by the screws on the OSB. Or doesn't really matter?

Other concern from insulator is trouble venting the attic with an R100 insulation. I've asked for an explanation since I would think that as long as the vent baffles come up higher than the insulation, there should be ventilation in the attic? We have continuous vented soffit all around, and ridge vents. More intake than outtake, but I read somewhere that it's better to have more venting area on the soffit than the more ridge venting and not enough soffit. Otherwise I'd need to cover up a lot of the eaves vents! Somewhere I ran the numbers to estimate how much venting at the ridge compared to the attic space, but can't find where I saved that.

By the way Dana1, on another post you said that with 25% thermal bridging in walls an R32 is more like an R22. Here normal build is R19 in walls - how much is Rvalue then if houses here are not OVE framed and probably have the 25% bridging? Just curious.

Thanks.

Boontucky-girl,

OSB has never caused a problem in application for me. It also gives an opportunity to shim or plane any trusses that are not level, increasing overall flatness of the ceiling in problem areas if they exist. OSB opposite long length of trusses increases strength in that direction equalizing the strength in both directions. Drywall should be perpendicular to OSB. You are supposed to leave an 1/8th inch gap between the sheets for expansion issues. The OSB also allows for additional nailing in any location and stops missed screws. Have your installer explain where the additional waviness is supposed to come from, because I have not seen it.

The issue you are running into is that most contractors simply do not want to do anything that they do not commonly do. They commonly do to code. Code is the worst possible construction and the lowest cost that the contractor can legally get away with. Bragging about meeting code is like being impressed with a person because they are not sent to prison on a regular basis.

Venting requires 1.5 to 4 inches of air gap over the insulation. If you are filling as much as possible netting the air gap bottom is the safest thing to do to keep it clear.

Brian

I may be wrong, but I would run the OSB perpendicular to the trusses and then run the drywall perpendicular to the OSB.  This method should result in the least amount of cutting or at least allow cutting the OSB across the 4' wide butt ends instead of the 8' side.  This method also allows each individual OSB board to be attached to more trusses which should help level it out and also reinforce the roof structure.  Drywall can be attached to the OSB thus resulting in less waste.  Do allow for moisture and temperature expansion by leaving space between the OSB boards.  Using a taut string and gauge block on the bottom of the truss is an easy way to judge if the trusses are out enough to require shimming.

Thanks Brian. When you say 1.5-4" of ventilation over the insulation, you mean between insulation and roof sheathing, or the clear space in the vent baffling? What do you mean by netting the air gap bottom?

Yes, I'm tired of feeling like I'm up a creek without a paddle. I firmly believe a good job is worth paying for, but I'll have to sacrifice a few things to fit the budget I will have.
I would agree that OSB perpendicular to trusses should be the way to go. Most trusses run one way, but as they get closer to the ends of the house, the hip trusses turn perpendicular and hang from girder trusses. Would the OSB sheets be perpependicular to these hip trusses as well, or continue the same orientation as most of the house?

Thanks Alton. So you're basically saying with the OSB, attaching the drywall is not as critical since it can be screwed to the OSB compared to a normal roof with only drywall.

Thanks for the tips and info, i greatly appreciate it.

Boontucky-girl

If you have a 2X4 as the top of the truss I would leave the gap between the them open. As you go up from the soffit to the top of the roof most codes require 1.5 to 2 inches clear minimum clear air gap as the minimum restriction for air flow from bottom to top. This limits the insulation fill to leave that minimum gap spacing. Vent baffling is supposed to maintain that minimum air from the vents but are often made of non-breathable materials. The insulation needs to have a path for moisture to escape above the insulation in the vented area.

The air gap needs to be clear of materials. Low cost netting or perforated radiant barrier or other materials can be used to stop insulation fill from getting into the air gap.

The netting of sheeting is easy to apply with a stapler and a razor knife. Great job for the handyman you found if it is an issue for the installer. It is probably only two days labor to have him sheet and fill the area with insulation as well. Usually Home Depot will let you use the machine for free if you get the insulation from them. Insulation goes on sale regularly for 20 to 30% off. If you house is already built there are often utility, State, Federal tax rebates as well.

OSB perpendicular to trusses is stronger and therefore better. This is beyond code and therefore is up to you as to what is easier for you to do. Either will increase the strength of the home. The home will twist less in the wind after extra sheeting. Note to make sure gap between sheets is maintained. If no gap the edges will sometimes squeak when the home moves in a big wind storm.

Dry wall is not structural. It is stronger in the buildup if each layer is perpendicular and that the seams should never line up. Overlaps are always best. You can also have additional screwing of the drywall to keep it flatter than 24 on centers can accomplish.

Brian