House energy retrofit project 25

    Another geek-fest ...

I decided that for a change this year I'd actually get myself to the Building Energy conference in early March, partially to see if I could hold my own to meet and interact sensibly with some of the more serious building-science nerds on their level. My head was stuffed with all this new knowledge, a bit of practical experience, and some early stats on winter energy use, so why not go do my own tiny little part to promote the green-building industry in general. Of course I also wanted to sit at the feet of the Great Masters if they'd let me; what's any topical conference without at least a little squeeing over luminaries in the field?

Overall the conference was pretty good and had many entertaining seminars *and* an extensive trade-show floor. It was definitely too much to try and take in all at once, especially when NESEA had organized it into something like eight concurrent tracks. I did get to meet some of the guys whose papers and case-studies and videos I had been devouring over the past year, and didn't hesitate to tell them that and thank them for the great online education. Joe Lstiburek unfortunately wasn't there, undoubtedly busy with some other gig, but I met a few of the other folks from BSI. Oh, and let's not forget to mention hangin' with Martin Holladay and trying to stump him with silly duct questions.

But with all that information overload, I actually didn't learn that much that was truly new to me. Almost everything people were there to discuss is out on the net, and it was a nice affirmation that I'd absorbed a good quantity of it already. That's not to sound snooty -- there's just so much that the choir can preach to the choir and the next big challenge that the community is working on is bringing all this more to the mainstream. The fact that the IECC and derivative building codes are rapidly being updated to add many more efficiency requirements is one visible result, and it sure ain't over yet.

[Click any image for a larger version.]

Different way of interlacing roof-to-wall foam My GC had a booth at the show, and they'd brought along this mockup of a roof-to-wall junction showing a rather different way to interface the double polyiso layers. Obviously this hadn't been done on my retrofit, where both roof layers come down as a single entity and lap over both wall layers at once. In contrast, this looked like a nice detail that could reduce the possibility of a small air channel all the way between inside and outside at the angle change, which if present could conceivably allow small convective paths to bleed heat even if completely closed off. I already suspected that was going on in my own envelope at the front and back roof-to-wall junctions, which seemed to be perceptibly colder than any other surfaces or corners around the second floor and attic area.

Just lose the exterior OSB in favor of plywood, please ...

    ... and a bit of backlash

I also spotted a couple of guys from the subcontracting group who did my main retrofit, greeted them, and mentioned that I'd made a lot of progress on my writeup by then. They got curious, and by the second conference day they had read most of the parts that involved them and they were *not* at all happy with me. While I had successfully expressed how the overall job went together in a pretty awesome way, I also pulled no punches in noting some of the downsides and that hit them pretty hard, to the point that they tried to convince me to take the *entire thing* off the internet.

We managed to resolve this in a non-antagonistic fashion. Details of the discussion and process are here for those curious, especially those who may have read a revision prior to, say, early March 2013.

Daikin unit partially buried The remainder of March brought a couple more healthy snowfalls, which weren't any sort of problem in themselves but in the following days as sun warmed the roof, the snowpack would slowly glaciate over the gutter and break off in fairly heavy chunks as I'd already seen. The thudding as some of them landed on the basement bulkhead was pretty impressive, and as more of the pack came down it tended to build up a densely-packed mound. Over the Daikin it hit the outer edge of the hutch roof and thus not the condenser itself, but as the pile on the ground grew it wound up sort of bouncing back and started to bury the unit anyway. Remember, this unit is up on the pad *and* 4 x 6 blocks, so that's a lot of snow.
So after a couple of the storms I still had to pop out once in a while and huff this wet cement farther away from the house, which during the day carried some risk of more avalanching from the roof onto *me* as well. It was clearly time to investigate a solution for this -- likely some kind of snow retainers as commonly seen on other metal roofs, which hold back the pack and let it melt off normally, instead of letting it cascade overboard and possibly damage whatever's underneath including people.

    More data from the field

Cheap-ass AC installs Cheap-ass AC installs
March also brought a weekend jaunt off to a local-ish SF convention, where I was amused by some of the shoddy minisplit installs tucked into obscure nooks around the hotel. This must be testimony that their existing central system was inadequate, or something, because these were relatively recent additions serving individual offices and the like. It's still fascinating that all these minisplit-oriented units from completely different manufacturers are built in a virtually identical fashion, even down to which side the compressor is located on and how the feet are constructed. And they rarely seem to actually get fastened down to their base pads -- almost like they shouldn't have even bothered with the pads here, but the fact that they're plastic at least keeps the metal baseplate off the corrosive concrete. Then again, my unit back home would also be floating loose if I hadn't insisted on some token screws and washers installed through those holes they give you for the purpose.
At least here I didn't see the bozo mistake made on one of my colleague's installations, where both the liquid and suction lines had been wrapped in the *same piece of pipe insulation*. Completely missing the point of why they get insulated in the first place, and highly ironic where that had been part of a buildout with heavy emphasis on energy efficiency and the HVAC guys were supposed to know what they were doing.

New vs old insulation inside a wall Armed with the IR imager, before we ran out of cold weather I went around to visit said colleague and some other friends for some informal energy audits. On this particular house the residents had recently redone the kitchen which included better insulation in its part of the exterior wall. The difference between that and the non-upgraded mostly-hollow wall was fairly profound, but obviously the windows were a larger loss point.

Note the heat profile on these windows, too -- apparently most transmissive at center-glass. Newer but fairly generic units -- I don't recall if they were single or double pane but certainly didn't contain any exotic gas mix.

Different window thermal profile Back home, I was still trying to determine any contribution from the Reflectix shades -- where the leftmost window here was left unshaded for a while, and the rest had their barriers in place. Again, a slight difference around the sashes, but the main thing to notice here is the very different overall thermal profile of the Serious windows here versus the ones on the friends' house.

Oddly, the tree seems warmer than the house but that might be that same low emissivity characteristic of the siding vs. that of oak bark.

Rear of house thermal shot There was still my larger and more annoying loss issue to figure out, visible in a whole-house shot from the back in the daytime -- the basement wall, persistently glowing far more brightly in the IR spectrum than anything else. Clearly, the meager R-3 or R-4 of the old white styrofoam wasn't cutting it against winter chill outside. And the brightest spot along there, about center in the image, happens to be right where the top of the water heater is located inside. Funny, that. I considered if it might be worth wrapping a big fiberglass blankie around the piping on top...

So I really wanted to do something about this. Some sort of additional insulation was likely needed, and while outside the foundation wall is the theoretically ideal place to put it there are real-world issues with insects, frost, and excavation that make it rather impractical for an existing wall. But I still had all this leftover two-inch polyiso that I kept the builders from pitching into the dumpster, some of which was already loosely puzzled into a layer over the slab and the rest of which that people hadn't taken away was piled up out back. I didn't want to just apply that for fear that an interior vapor-impermeable layer would trap moisture and do bad things to the wall behind it. I thought about using the 2" polyiso but stripping all the foil-face off first to make it a little more permeable, which seemed like a lot of extra work for a dubious theory. And then at the conference, I had the "epitome conversation" with one of the guys from Building Science.

Basements are tricky, and popular wisdom had come to be that since concrete in a foundation wall cannot dry to the outside because the soil is always at 100% humidity, it has to dry to the inside. In examining numerous research results and futher consideration, Dr. Joe [the industry's go-to guy for such wisdom] had recently reversed some of his thinking on this matter. Let's face it: concrete is fundamentally rocks, like all the other rocks along with the dirt that surrounds the foundation. Whether or not it's faced with dampproofing, those carefully structured rocks we set our buildings down on are going to basically stay moist. We work hard designing our envelopes to keep excess moisture on the outside, so why give those basement rocks any special consideration and continue a largely futile effort to dry something that's always being fed water from the other side? The conclusion was that as long as bulk water ingress isn't part of the picture, it's okay for concrete to be wet -- it likes to be damp, in fact, and entombing all of that behind an impermeable seal would be the right answer as long as said means of entombment did not include any moisture-sensitive materials.

Which does *not* include the wood framing with fiberglass batts and paper-faced gypsum board that so many builders of "finished basements" have had backfire on them over the years and turn into instant mold farms. Closed-cell sprayfoam, directly-applied rigid foam with sealed joints, or combinations thereof had returned the most favorable results.

It was already clear that despite being porous and thus quite vapor-permeable, the ancient styrofoam was holding back some pretty serious funkyness on the walls anyway. I had been going around patching up some of the missing areas and taping up its seams in an early halfhearted effort at air blocking, and sometimes when I removed a bit of the old stuff to replace an area with a more intact piece I got a pretty strong whiff of spores or mycotoxins or whatever the wall was throwing at me. While mold can't feed on the concrete itself there was something on the cinderblocks that clearly could support it, perhaps just accumulated dust, so I realized up front that trying to take the styrofoam *off* first would likely expose an existing mycological disaster to the rest of the house. No, the answer was to simply bury all that and add another well-sealed layer on the inside, letting the block wall be as moist as all the other rocks out there but not in a way that could condense interior air. Preventing evaporation at the block surface would likely stop any more efflorescence, too.

    Foam, foam, on the wall

Test insulation layer on basement wall Would insulating the wall make any thermal difference, though? I set up a quick test, pinning foam panels up against part of the wall and left it that way for a few days. Since the concrete is so heat-conductive I knew it would take a large area to create any visible difference on the other side; this wasn't a situation where a one-square-foot test patch would do.

Any difference? Maybe I was just kidding myself, or maybe I could actually see a *very* subtle gradient on the exterior around where the edge of my test patch was, somewhere around the dotted line. But with the internal webbed block structure and convective air columns comprising a broad thermal bridge to the inside, I really had no idea what to expect to see here.

Any difference? On the side wall it might have been a little sharper except that the edge of my patch also stopped at the edge of the window, so what I was seeing could have easily been thrown off by effects from the cutout or the window frame itself.
One possibility I'd discussed with my basement-wall mentor at the conference was that the *outside* soil could also be contributing warmth from its perpetual trying-to-stay-55-degrees deeper down, i.e. sending some amount of heat into the wall from below which would then emerge from the above-grade area. That aside, I knew the walls were fairly cold all the way down because of the air columns inside, and it was becoming clear that effectively insulating this was going to pretty much be an all-or-nothing effort to make any difference. But blocking the inbound humidity diffusion seemed like a worthwhile goal too.

Typical slab corner thermal There was also the slab to think about, but connected to a steadier mid-fifties seasonal ground temperature underneath the house instead of ambient air it probably wouldn't leach nearly as much heat out of the premises. Especially if the basement air was allowed to stratify and be colder at the bottom. It was already clear that the walls were coupling to the edge of the slab via conduction or interior air convection or both, cooling the edge, but things were much more moderate toward the middle of the slab. As something of a compromise, I figured I'd include insulation a foot or so in on the slab itself and build up the wall from there.

Prototyping foam corner I prototyped up a rough idea of how a corner might look and took a few measurements. The air just inboard of this simple setup became substantially warmer than the air at the raw corner.

It occurred to me that the experiment of laying foam on the *slab* may have hindered instead of helped, preventing the warmer floor from compensating a bit for all the cold coming off the walls. Time to reverse that situation.

Thermal 'shadow' behind foam board While slinging and stacking the many odd pieces of foam around I noticed an interesting effect: when I leaned a piece of the polyiso against the EPS-covered wall, a thermal "shadow" would form behind it. Remember that this is heat, not light, and it wouldn't appear instantly but rather took 20 or 30 seconds to develop. The "image" would take about the same time to disappear once the polyiso was moved away. My best guess is that some heat was radiating from the items inside the basement [including your humble author's corporeal being] and the floor structure above, helping to warm the inside surface of the styrofoam, and when that path was blocked the conduction into the colder wall would take over and chill the area down. Whether this theory was correct or not it was more evidence that the styrofoam was pretty wimpy insulation.

Candidate adhesives Glue tests
The next question was how to attach the polyiso. I knew about the Hilti IDP plastic insulation fasteners, but didn't really want to get into banging a bunch of holes into the block. The styrofoam had remained pretty firmly up for years with only whatever old brown glue the previous owner had used -- possibly Liquid Nails or the like, and could probably continue doing fine with holding a combined layer to the concrete. Since the polyiso is very light and would be stacked from a ground layer, any adhesive used would have very little long-term stress on it.

Still, it was worth testing various candidates for cured adhesive strength: silicone goop, Liquid Nails, and the "fifty year" caulk the builders had used all over the retrofit. I let test patches between the foil-face, EPS, and concrete all cure for a few days and gave 'em the pull test. The Dynaflex caulk came out clearly on top, forming bonds that were stronger than the EPS and the foil-face itself. Liquid Nails basically fell right off the foil, and the old silicone stuff was completely useless on any of the joints and only left a little slimy wet patch on the concrete floor that I had to wipe up.

Polyiso flame test There's also all this talk of ignition barriers needed over various kinds of foam insulation on the interior -- intumescent paint, drywall, or whatever. Code calls for such things over materials that already have pretty low smoke and flame-spread ratings, and various tests I'd done showed most of the foam I was working with to be largely self-extinguishing. Since the foil-face polyiso would become my new interior surface, I did another torch-test on it. It took a fairly insistent flame to penetrate the foil, after which the foam burned a surprisingly small amount and quickly put itself out leaving a largely intact cellular structure that simply *would not burn* anymore. We're talking space-shuttle tiles here.
Maybe it wasn't a full-on ASTM "corner test", but I torched it bright-yellow-hot without seeing any further combustion and concluded that the foil-face would be even safer than the styrofoam that I already wasn't particularly worried about, without messing around with any additional coatings for the moment. Moisture-sealing or not, there was no effing way I was going to put up any *drywall* down here. In researching what compatible intumescents were available at the time I had an amusing conversation with the engineering guys at TPR^2, whose BlazeLok and Fireshell products weren't specifically geared or tested for application to foil-face but they said they were definitely working on the problem. And really, the foil also serves as a radiant barrier and I could see no good reason to cover it and destroy that capability as my heat issues in question seemed to have a lot to do with radiant transmission.

The local building inspector had already been through the basement a while back with all the ratty old white styrofoam in full view, and never said a word about it so for the moment I was likely off the hook about any of this. Maybe I'd worry about it if I went to sell the house.

Trimming bottom-styrofoam seal Okay, so I'd decided to go ahead with this and see how far the cut-offs I had on hand would get me. First step was to complete the first-layer air seal against the walls by finishing the remaining seam-taping, and filling in the gaps at the bottom. I foamed all the way along with low-e Great Stuff, and cut it back flush after it finished curing. This was step one in closing off the remaining high-to-low convective air paths behind the styrofoam, which were probably plentiful since its contact with the block was anything but uniform. [Not to mention mouse-munched in a few spots...]

While the Tyvek tape didn't adhere as tightly to the EPS as I might have liked, it was better than nothing to close up the often substantial gaps between the sheets of the stuff. I wound up simply caulking some of the narrower gaps near the top where it now disappeared under the new sprayfoam around the sill.

Thermal profile under removed foam Under a recently-lifted larger piece of polyiso, the thermal profile of the slab floor pretty clearly showed the cold area taking kind of a diagonal across the corner, which makes perfect sense given that corners present more surface area to the cold.

Foam corner going in, lots of supplies So I figured my "footing" on the slab should do the same at the corners, to better follow where the cooler areas would lie. My footing width wasn't necessarily a full foot wide everywhere, but I was using whatever long pieces seemed to conveniently fit in.

In addition to coming home from HD with most of a "contractor pack" of Dynaflex caulk, I found a good source for rolls of the Weathermate tape to seam-seal the polyiso once it was up. I got the less expensive two-inch instead of the three-inch, as this assembly wouldn't be exposed to weather.

Special notching around floor features Some adaptations were necessary for parts of the floor that were uneven, notably around the old oil-tank leg stumps. In most cases I could just gently pre-bend the foam to crinkle the foil on one side to follow the floor contour, or even dimple it a bit with a rubber mallet and accomodate larger bumps and ripples.

The footings went down adhered near the lower edge of the styrofoam and the floor, forming another layer of sealing with mostly continuous beads. I left some small gaps in this just in case bulk water ever did come in through the wall, to allow a few seepage paths rather than building pressure behind the whole assembly and blowing it off the blocks. Even though the basement hadn't seen any real leaks for close to 20 years other than the dribble from the concrete-cutting cleanup, I'd rather mop up a little water than have to rebuild the insulation layers.

Bracing foam up to let caulk cure The attachment process was very slow, as I'd have to cut and fit a couple of pieces and then lean heavyish things against them while the caulk took an initial cure to the self-holding point. But I wasn't in any hurry; over the next several days I could easily slap up a little more foam and brace it and then go off to do other things for a few hours in between.

First two walls done In the next three or four weeks, the crazy-quilt nature of how all this got patched together emerged and eventually covered most of the north and east walls [which I figured made the most sense to start with, likely being the coldest ones]. Taping the inward right-angle between the wall and the footing involved some trickery with a straightedge to lay it in tight, but it all got sealed up.

A developing study in silver and blue, and I had plenty more to go. The shelf set was a bit of an impediment as it was pretty firmly set in place, but it represented a substantial wall area that also needed to be covered.

Insulation in between shelves It was easier to trim pieces to fit neatly into each shelf bay rather than uproot the whole thing. The wood wasn't a total thermal bridge since the styrofoam passed between it and the block, so adding the polyiso in the remaining space would be fine. That all got foamed and taped up too, and that shelf wasn't goin' anywhere afterward.

Custom-cut foam piece to go around duct Much later in the process I had reached where the ventilation exhaust duct goes through the rear wall. This had a pretty tight seal around it already but may have developed a couple of tiny air paths, so to make sure I added more caulk and tape before burying all that under a piece of XPS with an exactly-placed custom cutout for the duct. This was also the first time I'd taken this junction of the exhaust line apart since it was assembled, so here we finally see how the Cape backdraft damper was installed.

At this point I had used up quite a bit of the leftover polyiso, and a chunk of the same ratty cast-off blue XPS from the lumberyard that had gone onto the basement door happened to be the most appropriate-size piece on hand. It's all good; the Dynaflex and tape stick to it just as well and it's close to the same R-value. Visual ugliness was *not* a relevant factor here; the idea was to neatly use up a lot of remaining materials stock in a productive way.

An interesting mix of foam technologies The results in general were an interesting mix of technologies -- polyiso, XPS in various grades, tape, and topped all the way along with a creamy layer of Great Stuff to tie the upper edges to the sill sprayfoam and complete the second-layer air seal. A long piece of one-inch XPS brought full coverage all the way onto the doorframe but left just enough room for the door to swing open against the stopper protecting the duct.

A new bead of caulk went around the pipe on the inside, and the outer duct sheath was later brought within about a half-inch of the new wall surface and tied down there -- lesson having been learned about exposed fiberglass around cold pipes. Any condensation forming in this gap could just dribble harmlessly down the wall if needed.

As it turned out, the leftovers of polyiso and other random stuff I still had kicking around was enough to bring my odd patchwork all the way around three-and-some out of four basement walls -- much more than I expected. It was quite the puzzle sometimes figuring out which pieces to put together to minimize waste and cutting, but it worked out rather neatly and stopped just short of "infrastructure row" where the water-heater and HRV and laundry stuff is clustered. That was good enough for the moment and I had several more general fixups in that area to do anyway before getting more foam to complete the run.


So now I was committed to having stopped whatever moisture and infiltration air path may have existed through the cinderblocks, which would likely help keep the basement drier in general. Some could still permeate the slab, but probably not nearly as much as the block and with the new overhangs at the roof and built-out walls, the dirt by the foundation was generally staying far drier than it ever had before so going forward there would be less water around the walls at all. The only remaining small concern was a drying path for the sill and beam-pockets, but everyone I talked to and showed my little diagram [below] agreed that the generous above-grade span of unfinished foundation wall would provide plenty of diffusion area. There are still plenty of uncertainties in the field where insulating basement walls is concerned, especially once a wall starts getting colder in the winter. The recent Report 1005 from Building Science gets into this around page 50, and concludes that "more information is needed". Well, I was probably well on the way to having a rough candidate data set.

Wall cross-section diagram

Said diagram came from some playing around with GIMP in advance of the conference, so I'd have something to show people while asking my questions. I sort of followed the informal DER case-study conventions where existing structure is shown in light grey and new construction uses more indicative colors and fill patterns. Obviously I could never apply anything impermeable on the *outside* of those blocks to seal them, as that would channel any rising ground moisture straight up to the wooden parts and likely be their death knell. The firm butted contact between the sill piece and the joists and studs does provide a tenuous path for some sill moisture to escape toward the interior, but I wouldn't want to rely on that as the only means. And yes, that lowest piece of sheathing does have a big aluminum heatsink to the outdoors and is going to run colder than the rest -- that's one minor downside of this configuration, but the other side is mostly encapsulated by the sprayfoam and I didn't notice any condensation in that area during cold times.

Early spring shoots coming up Late March brought a few warmer days and some first signs of spring: some early shoots popping up through the decidedly frost-heaved soil in the yard. That didn't stop me from trying to squeeze in a few more heat-loss tests, as the nights were still fairly cold. With the supply ducts now better insulated against radiant loss I wanted to see if the toaster would perform any better, among other things. I ran a few nights alternating between running on the heat pump alone and the toaster alone and comparing energy used against local degree-days, and confirmed that the heat pump was delivering a COP of right around 3 at 32F ambient.

I also fiddled around with high fan vs. low fan, which didn't seem to make a whole lot of difference but it wasn't anything like a rigorous test. With the ducts rendered less lossy I thought that low fan might create less static pressure through the system and cause less air transfer between upstairs and basement, and possibly reduce the amount of warm air washing against the exterior walls opposite all the supply registers. Fine points, probably not relevant since rough calculation indicated that (CFM) * (dT) = a constant regardless of high or low fan anyway.

Old vs new wall section mockup Old vs new wall mockup
Inspired by some of the mockups I'd seen at the conference and having noodled up some decent drawings of my own, I wanted to do my own wall-section model to show the old and new assemblies together. I had saved a few bits of the old material, the siding and tarpaper and ratty Kimsul insulation ... and had some appropriate-thickness boards to simulate the plank sheathing. Once a pseudo stud was attached to a suitable baseplate, the rest fell together pretty quickly including a way to show how the Headlok screws tie the strapping into the original framing through all the layers, how foam seams get staggered and taped up, and how the "buttons" temporarily hold the foam on and that it's common for those mid-field screws to poke through the sheathing. A couple of extra parts clarified window and door buck construction, and of course I had to toss in one of the cut-off rafter tails with its visible history of water issues, to help explain the benefits of the "chainsaw retrofit".
I took this with me when I went out to visit a DER open-house out in Worcester, where it was well-received and I met up with a few more industry luminaries. One of them posted later about our discussions on heat pumps; the mention is fairly far down that page so search for "goodman" to find it. That particular retrofit was a largish three-family with three separate Mitsu outdoor units, all hung on the house on wall brackets about 4 feet off the ground. An interesting approach to a basement-bulkhead door there was to build sort of a small insulated "tunnel" off the exterior wall to hold a second door about two feet inboard of the foundation line, leaving the old ratty outside door in place.

Thermal gradient up back wall Some warmer days arrived rather suddenly, and with the daytime sun still fairly low it was hitting the foundation wall in back through the leafless trees and making a pretty clear thermal gradient up the still-exposed styrofoam on the inside. Okay, so exterior temperature changes were showing up on the inside. I already knew that would happen, just not exactly how much, but once the foam job was completed that was going to largely cease.

Intake duct isn't too dirty I could finally take apart the HRV inlet duct and start reworking it to be more condensation-resistant. While the system hadn't been in operation all that long, the inside of the duct still looked fairly clean and there wasn't any evidence that stray outside water had gotten any farther in than the designed-in weep hole at the PVC elbow.

Extension of inlet duct I coupled together two elbows to make a nice broad sweep that would be easier to slide the replacement insulation jacket over, and to position the flex drop right over the HRV inlet. Once spun to the right angles, the elbow "gores" were locked in place with aluminum tape. The insulation could wait until later, as summer temperature deltas wouldn't make this short bit of duct a significant heat source in the basement. The wood all around the inlet pipe got sanded and protected under three coats of urethane, so any condensation from interior air would be far less likely to affect it.

    Follow the money

Early to mid April saw the wrapup of heating season, and I went back over my data-collection for a review of the overall energy needs. Glossing over a lot of detail, the indoor meter showed 1963 kilowatt-hours used for HVAC alone, e.g. pretty much the moral equivalent of how many gallons of oil I would have burnt with the old furnace. Averaging the on and off peak electrical rates to 11.3 cents per kwh that worked out to $222 spent on heating from November to April, over 5041 heating degree-days. It was probably even less than that after programming the system to revert to a low setback over the peak metering period and let the house just flywheel across it until 7pm, but I'm ignoring specifics on that for this year.

Assuming a *very* modest average COP of 2.5 at the heat pump, it took on the order of 16.75 million BTUs [plus whatever came from plug and people loads] to keep the place at a nominal 65F most of the time. Burning oil to make that in the old furnace, based on my efficiency figures, would have consumed a *minimum* of 186 gallons or about $650 at the cheapest per-gallon rate I saw advertised that winter. All this is very rough and ignores the fact that extra electricity was consumed over times I was test-running on the three-kilowatt "toaster" -- but that probably balances my occasional use of the bedroom space-heater with the main system set way back and not running. Nonetheless, I could already see significant heating savings, not to mention the priceless peace of mind not worrying anymore about flakey old combustion appliances wheezing away in the basement. And keeping the entire pre-retrofit house that warm with oil would have cost *far* more, even with the storm windows and trailer-park plastic over the insides.

      Did you catch that?   $428 or better calculated savings over the first heating season from the heat pump alone.

Better, the all-told combination of heat pump and retrofit had likely reduced the annual heating energy *cost* by an order of magnitude from the pre-retrofit state of things. In the case of this and indeed most existing-building retrofits, it's unfortunately still not worth talking about "payback time" which is on the order of multiple decades, but all of these lessons learned plowed into *new* construction can mean huge energy reductions for relatively little incremental cost. Another less-tangible but monster benefit was the fact that this had been the first winter in a *long* time where I didn't get knocked down at least once during it by some sort of two-day flu/fever bug. I attribute that directly to radically improved interior air quality.

The additional energy factors were numerous and made exact calculations far more difficult, but could be somewhat ballparked from the difference between the main meter and the HVAC meter. In this case that's about 6.8 million more BTUs worth of input, but consider that some of that heat energy went right back outside the envelope again carried by things like drain water and a few times I used the block heater on the car. All told the 23-ish million total BTUs I can account for over the HDD65 history for the period shows about 190 BTU / hour / degree-F total envelope loss, in comparison to the 300 I'd come up with for the old envelope. But ... only a one-third loss reduction from all that retrofit work?? That couldn't possibly be right.

Two major unknowns were likely throwing things off. One was the basement loss, a huge indeterminate nonlinearity which would be at least partially mitigated before the next winter. The other was that likelihood, again, that I'd underestimated the total furnace output. An old service card I found from the last time "professional" hands had touched it showed test results at 73% efficiency with the big ol' "wall of flame" nozzle still in it, higher than I was willing to allow it given what I was able to measure but which would have put its production at around 75,000 btu/hr in that configuration. I also never considered that the warmed chimney might be radiating a bit more coziness into the structure, even if some of that was escaping up the chase into the cold attic. Oh well, there was no way to re-measure the old loss figure now even though I went back through some of the old data and tried to look up official degree-day history from then and "fit the points to the curve" -- that kept landing in the neighborhood of the low to mid 300s anyway. And with the old thermal boundary pretty much at the second-floor ceiling instead of the roofline, the old setup had a slightly smaller total surface area for net R-value figuring.

So clearly more study was in order, but at least additional enclosure improvements were under way and I had a better idea about how and where to collect data. As imprecise as some of it could be, I was totally loving the engineering aspects of all this.

    Glaciers push things

Roof seam ends pried open by snow pack But for now there were other projects to start tackling. Spring was on the way! It was time to get up on the roof to see how things had fared over the winter, and I made a rather disturbing discovery there. Look closely, use the big-pic. The snow-pack had caught several of the little seam-end tabs [which I hadn't re-crimped flatter back here, only on the front] and actually *pried them open* in the process of slowly sliding down past the ribs. Argh. Maybe the roofer's "improved" idea of seam-end details wasn't so great after all.

Here, by the way, is how I ultimately allayed my minor fear about damaging the gutters any time I went to put a ladder up here -- my cheapass consumer-grade extension ladder with the stabilizer bar just makes it over the top, where with a little adjustment of footing the bar pads can rest safely on the panels and bridge the whole thing away from the gutter.

Seam ends sorta sealed up I bent the tabs back and pinched them flatter this time, and as a little extra touch applied patches of aluminum tape over each tail to try and smooth the surface over the exposed edge -- with the understanding that it might not hold long-term and that I might need a fillet of something like hard mastic there instead. And/or maybe just shorten the tabs so they wouldn't present such a bendable lever-arm to stick out. Whatever. If any of the tabs wound up completely broken off I could just stuff a little fiberglass up the end of the seam, as it's not a critical water-shedding detail at that point.

It had already dawned on me that what I really needed up here was a proper snow retention system after all, so I began looking into where to get one.

S-5 snow guard parts Researching this was fairly easy; the ATAS web site points right over to the S-5! site as the system they most recommend for their standing-seam roofing. I had also visited S-5!'s booth at Building Energy. Obviously I wanted a non-penetrating type of attachment, and the seam clamps that go with their "ColorGard" system looked perfect. Loading math given the roof angle, panel width, and assuming a couple of feet of heavy wet snow at 20 pcf showed about 250 pounds maximum longitudinal load, well within the rating of the "S-5-U" type clamps if I installed one on every seam. I already knew the roof structure was good for heavy loads in the vertical, from a couple of winters prior when I was up there shoveling like everybody else in the area -- and if anything, the whole structure was stronger now.

Within a couple of weeks I obtained a nice kit with all the necessary parts.

The only mild concern was the panel attachment itself at the top, and for that I had to go back and look at some of my installation pictures. The panel edge clips are designed to slide and can't be counted as "fixity". Each panel got tagged in with one screw as they went on, and then each Z-bar had two more holding it down for a total of at least three per panel. While the S-5 tech-support folks I talked to implied that they'd normally expect more like five, they were thinking wider panels and more like every other rib or so supporting the snow-guard system. Clamping every panel would distribute the load across all of those screws, and besides, they were all into plywood and not OSB so that was a big plus on holding strength.

For what it's worth, adhesive-based retention systems weren't even on the table for my consideration. Too many detachment failure stories.

Building materials stock I had ordered through Beacon Sales a couple of towns away and went over to pick up the parts a few days later, and while in their yard happened to notice a pile of stuff allocated for another retrofit that my GC's subcontractor was doing. That was pretty amusing to find. Clearly, regardless of whatever I might have ever said about those guys they had plenty of work keeping them busy.

Anti-wasp fiberglass stuffed into holes While actual installation of the snow retainer could wait, I had a more immediate priority. Warm weather was already bringing the wasps around to investigate the generous cavities under the roofline split, just as I predicted. I hadn't designed a workable screen enclosure for these areas yet, but in the interim I could stuff wads of fiberglass into the key places to try and discourage them from starting nests in there.

Did I call that one or what...

Guck in the gutters The seam tabs in front had apparently survived the snow-pack slides, probably due in large part to my early flat crimping as the loads would have been even higher on the steep slope. I did the aluminum-tape job here anyway, just for consistency. As I didn't really care about the mounds of snow that landed in the front yard, this side wouldn't get snow retention for the moment. Implying some extra risk to the gutter, but it had fared okay so far and I'd keep an eye on its load handling.

While up here I pulled one of the gutter-guard pieces to see just how much organic muck had gotten in there. It wasn't actually that bad. Mostly it was worth cleaning off stuff lying on *top* of the gutter screens so water wouldn't bridge across and drip straight over, but that's pretty easy. As spring progressed it turned out to be nominally necessary to remove all those little brown oak pollen "threads" that collect into big clumps as they were starting to choke the screens. Yeah, never clean your gutters again my ass.

    Heading for summer

Wishing-well pipe finally dug out I finally finished digging out the "wishing well" thing, and rolled the piece of pipe toward the back with that years-old intent of turning it into a raised fire pit. I still wasn't sure how that would work out, so it just sat ignored at the edge of the yard for a while. But I could at least level out and landscape the bare spot by transplanting some mossy topsoil "sod" from the back. This entire so-called lawn is all about the moss, not the grass, as the more moss I get the less maintenance the whole thing needs.

Roof heat bridging down screws Thermal bridges like a theatre marquee
A while later came a couple of *really* warm days with sun beating down on the roof, and I was curious how the new structure was handling that. The IR image was really amusing. Each long screw from the outer plywood into the rafters was clearly forming a little thermal bridge all the way in and warming spots in the wood, looking sort of like a theatre marquee. The exposed ends of the mid-field button screws, particularly from the outer layer, were quite warm to the touch.

But even with metal conduction the bridging is all quite small in terms of area, and wasn't actually bringing a lot of heat into the attic. And it was already clear that the roof metal, with its finish SRI of 51 or so, was bouncing a healthy share of the heat impinging on it -- while up top I found that with full sun on the panels the metal quickly got almost too hot to touch, but cooled rapidly whenever a cloud came along. That, similar to what the siding does, does lead to a certain amount of creaking and groaning from expansion on those in-and-out partly cloudy days.

Stink pipe thermal gradient The stink-pipe showed a healthy thermal gradient, but nicely contained in the PVC section. That would extend much farther down if the whole thing was still black cast iron.

Alternative return end-cap with duct takeoff Anyway, it was time to think more about summer HVAC mode. The HRV integration with cooling was ready to go, but I wanted a better return pickup for the basement to complete the summer-only air path down there -- collecting from the already cooler air near the floor instead of high up where the duct itself was. The removable end cap with nice pull handles was a fine start at configurability, but what I needed was an attachable duct piece to redirect the intake anywhere I wanted to put it. The obvious answer would be an alternate endcap with a duct fitting added. This was cut and bent from scratch, using the existing endcap as a template, and a generic six-inch takeoff from HD fit neatly into the hole.

New end cap in place So here's summer mode rigged up, with the alternate low return pickup attached through a piece of flex duct with the other end simply resting open on the floor [that's okay, it's all pre-filter], and the supply register at the other end of the space opened up to loop the whole basement into the HVAC system's drying capability. Winter mode would go back to both items closed off and the basement running considerably cooler than the upstairs, with only a tiny bit of circulation from the HRV pickup.

    A far better ODU screen

Clip blocks for attachment system Even though it was more relevant to the snow problem, I also wanted to get the new screen for the outdoor condenser unit done. The main headscratcher was how I'd attach it without drilling holes in the unit and risk sending drill points into the coil or something. There also wasn't a lot of workspace behind the box. With generous steel flanges available top and bottom, the elegant answer seemed to be some sort of clip system. A little work in galvanized sheet and bits of Azek later, I had a prototype mounting system that would be completely nondestructive to the unit.

Clip block in place Here's how the clips go on, and the screen frame simply screws into the other side of the Azek block. All zinc-plated screws, for best compatibility with galvanized.

New condenser screen The frame was constructed from various old trim scraps, in minimal thickness to avoid interfering with airflow and all urethaned up for weather-resistance. I was pretty happy with the result, even if the piece of half-inch mesh was kind of ratty and bent -- it was free from some friends already in a mostly perfect size, I didn't have to go buy a roll and then only be able to use a small part of it. It sits over an inch away from the coil, with the intention that it would totally prevent that blowing-snow freeze-up problem.

It was now a little more like that shroud structure I'd spotted with the iced-up pipes at the local show hall a year ago, than the practically useless open plastic grille that came with this thing.

    Air box access

One thing that the HVAC installer had never taken into account was how to access the underside of the indoor A-coil for cleaning and inspection. That's the side that will get dirty despite the filter, of course, and may very well need a brushing-down once in a while at minimum! While I could sort of see it from the filter box with a mirror and a light I couldn't actually reach it, and what I really wanted was an access opening in the box underneath the whole air-handler assembly.

The problem was that the wood structure inside the air-box would interfere with a hole cut into either of the two surfaces I had available, limiting the maximum working opening.

Air handler box marked to cut Width wasn't that limited but I could only do 9 inches high between the internal framing. A rectangular solution seemed in order, except that most companies that make duct access-hatch fittings make them square, and I only had about 16 vertical inches of workable sheet metal to mount against. I came home with a 14 x 14 hatch from a local supply house, and via measurements and just plain feeling where the wood was backing the box marked it up for cutting.

Fitting access hatch to body The 14 x 9 net opening I'd have would be enough. I templated it up with the hatch frame and figured if I could get my head and one arm/shoulder in, that was good enough for working on the coil.

Behind me here is the progress at that point on the re-insulation job. At the time it ended just out of frame to the left and just shy of the dryer; that had gotten three walls fully covered and a good start on the fourth before I ran out of retrofit scrap. Unbelievable that they were just going to chuck all that -- over 500 square feet of usable foam, several hundred dollars' worth.

AHU box cut for hatch Uh, I think I've made myself a massive return-air leak here!

The problem with the wood placement is pretty clear from this, but I was already happy with this opening. Especially after filing the cut edges smoother so I wouldn't slice myself to ribbons by sticking body parts through it.

Reinforcing air handler support frame Because of the way the wood structure was put together the air-handler was resting mostly on the horizontal rails and transferring all that load through a handful of cheap-ass drywall screws into the uprights whose tops were pretty much contacting nothing solid. I had no faith in the long-term viability of that, and figured the least I could do was send a couple of good exterior-grade coated screws into each upper joint to reinforce the whole thing a little.

Before drilling for those I jacked the rail up just a tiny fraction, to make sure my screws were preloaded the right way once installed. It did move just a little bit -- not enough to distort any of the ductwork connected to the unit, just enough to make the new holes *not* quite match in a favorable way once the screws were in and the lift was relieved. Did that at all four corners.

And the whole point was for *me* to be relieved that the assembly would be a little less sketchy now.

Duct access hatch done Hatch all installed and sealed up with aluminum tape. Despite a bit of waviness in the air-box metal it seated fairly well once a few additional zip-screws went in, and the tape covered one or two minor fitting flaws.

This is something very rarely seen in residential systems, because most HVAC installers don't think ahead to routine servicing. If they have to cut ductwork open, they usually just slap a flat sheet back over the hole with some tape and call it a day.

A-coil loaded with water Now I could finally see and reach the whole underside of the coil. It had even gotten warm enough to test-run the cooling, and watch it load up with water which is what's going on here. The plastic drain pan has a fairly generous height, and while the coil fins fit fairly tightly down into it I can watch for bottom-coil corrosion here and maybe even vacuum out a little additional water at the end of cooling season. Later on I discovered that because the whole tray with the coil in it gets cold, a little condensation also forms on the *bottom* of the tray and drips off into the bottom of the air box. It's not a lot, but despite Martin's misgivings about pressure-treated lumber in an air plenum I'm glad it's in there.

Routine activities in here would be way more palatable that grubbing around in that sooty old furnace firebox, for sure.

At this point I also removed the "toaster" heater entirely from the top of the unit and set it aside, so it wouldn't get washed with the high-humidity air coming off this all summer. While evaporator coils remove water, the exit air's resulting 100% relative humidity doesn't really drop until it circulates back into the space and warms up, which by the time it's through the blower and distribution box hasn't really happened yet. Why risk faster corrosion of the heating element when it's trivial to just take it out when I don't need it? And my COP-1 watt-hour measurements aside, I arguably don't even need it back in for winters unless the heat pump system craps out or it gets *soopah-wikkid* cold outside.

Next section   (26)

_H*   130706