This post is built from photos posted in Picasa Web Album Truss Attic, January 2016 .
After five man-weeks, there is this:
65% of attic floor (ceiling insulation) areas are durably plywood-covered, yet insulation is R42 on average. This insulation will endure.
The 1991 home has a living room cathedral ceiling not at risk of rot, as are cathedrals foolishly integral with a roof. Half of the energy savings due to my weatherization are at this now-benign cathedral. Most of those savings are at R30 walls, previously with flimsy R13, not in contact with drywall.
How does one form hard-covered R30 walls? Details are in my comprehensive photo album: Milwaukie_1991_Attic .
Average attic floor insulation was R10, that commonly found in distressed loose-fill that was intended R19. The very large attic was a useless and dangerous place.
Attic walls had no useful insulation. The trampled 3" of loose-fill on traveled floor areas was R7.
Over time, plans become detailed, real and accurate.
Build safe, strong flooring supports, 10" here, and 21" for a raised central deck.
At long last, find a cure of can light energy bleeds over the living room cathedral. Four remodel cans were pushed up 24 years ago during house construction, with no attempt to reset insulation, and with poor delivered light.
The first can light replacement by Nicor DLS4 on a RACO 175 junction box, is less perfect than three that will follow.
Replace three can lights in a batch.The drywall rings are cut from divots found in the attic, carefully aligned with ceiling texture. Two drywall screws deliver perfect patch alignment, each time.
Six Nicor DLS4 now, really pretty, I thought then. My next customer will complain of glare, and I will stop offering them. This customer still likes them.
Work in the attic includes 100% sealing of wall headers , now foolishly disdained by most other contractors.
Most of my work is in preparation, lights below, attic lights, supports for raised flooring/ covering of insulation, and more. Placement of new insulation is dramatic and satisfying, but is a small part of the job.
Top layer placement for each periphery half, takes three hours.
I knew I would hard-cover most attic floor insulation, and the attic walls, by practiced methods. Here in the living room cathedral ceiling, I settled on something new. Hostility of this tight space, with bristling nails overhead, and real risk of falling into the living room, must be overcome. Create safe access by encapsulating an R25 base on 5" cuts of plywood. Cover all with crossing R19, exposed.
With the base-layer covering, it is possible to safely reach all the way to the soffit blocks in placing added insulation layers here, too.
Noisemaker bath fans were belatedly replaced with nice Panasaonic FV08 VQ5. Shuffling the plywood flooring and insulation is painless.
This attic is not a trash heap. A small salting of hole saw debris is collected.
I don't use a Sawzall. Grind off the coarse shingles with turn of deck screw points. Then make the simple, clean 4 1/8" wood cut. I have pulled a minimum of nails in three rows, parting tar-down with a blunt Wonderbar. The spun-aluminum penetration adapter plate is confirmed oversize, and will be trimmed.
I have owned spun-aluminum roof penetration adapters, four each in sizes 4", 6" and 8", for a year. At last, I get to employ two 4" adapters, with new ducting of two new bath fans. Trim off some excess material. Take one third of that 3-tab shingle, split it at the tar line, and insert a piece over each installed adapter, to gain one more layer of assured roof water-tightness.
This is the watertight hole of a roof penetration adapter. Nails are reset. Unexpected shingle damage is tarred. Needing now only a rain cover, roof integrity as good as new.
Use no tar," it says. I come close to complying. I must learn to use screw heads that self-seal.
New Panasonic FV08 VQ5 fans and needed straight initial ducting, are protected under insulation and decking. Use shortest run, nestled with truss elements, in the reach to the roof penetration. Added 2x4 under the roof sheathing compensate for the roof cut, and anchor screws of the vent cap.
Hard covering is important everywhere, It matters especially at tight passages.
Admire the Calvert attic ladder, installed with innovation. The "broad top step" importantly adds to safety. Every new safety invention becomes mandatory.
Access inventions are shared freely with any manufacturer:
https://sites.google.com/a/r5portals.com/www/home
At my r5portals site, please see current, very important advocacy for good access in new building codes.
https://sites.google.com/a/r5portals.com/www/better-building-codes-for-access-portals
If you have a vote on the IECC developments for 2018 , please show interest in this.
With an interfering bath fan duct gone, add stability to the truss array. This 16"x96" piece of strong 3/4" CDX plywood, with four 2" deck screws to each truss, strongly resists rotation of trusses to collapse in a tornado. It is a small cost in offset of cheapness in 1/2" OSB roof sheathing, but could you do this without flooring and lights?
Google it:
the science of sealing attic wall headers
M writing is detectable near-top. At this blog, pick Label: Sealing Wall Headers , here or at the right.
I think the process of heat transfer at attic walls, via paths at wall headers in an attic, can be imagined with this captioned photo:
Always in my work, wall headers are 100% sealed with flexible grout, as they are revealed. Gaps as here, are sometimes very large. Blackness of insulation pushed to the right speaks of the heat transfer process, now blocked by the sealing. Energy in convection pulls air down in some wall areas, and up in others, attempting to equalize attic and wall interior temperature. Dirt of outside air is deposited where passage is down, in the constant churn of air. The process has little to do with air infiltration, but rather is throttled action of attic floor pits.
The process in any home shifts with room use and weather outside. No one has attempted rigorous analysis or testing, or should. Yet, we need to estimate the benefits of sealing in defense of the fair cost of sealing, $200 to $500 in any home.
I estimate the savings as that of an inefficient attic floor pit, 10% bare area equivalent , of conditioned space ceiling area. The home of the photo above has 1366 sf heated space, and I charged a meager $200 for the 100% sealing with my flexible grout.
Estimated savings are $2.4*0.1*1366/3 = $109 per year. Two year simple payback. Bases of the math are: heating with a gas furnace 88% efficient, 4400 65° HDD, natural gas at real cost $2 per therm. I wish for math to show that annual savings are at least 20% of the $200 cost. That is achieved, however you do the math, and whatever you think is fair cost of energy.
Recovery of invoiced cost could never be achieved if the home owner had hired the top HPwES contractor in Portland, Oregon:
This bid is mean. It is knowingly based on lies. It is criminal. Allowed and encouraged criminal conduct kills public trust and kills weatherization activity.
Failure to do 100% sealing of attic floors must be seen as inexcusable. Excuses are offered by powers in Portland, Oregon, that consequences are not measurable by a blower door , and methods tried are rejected by nearly all contractors Final Evaluation Report - New Homes Air Sealing Pilot . In a state of confusion, lacking conviction through scientific understanding, air sealing does not happen. None is required.
There is no longer an Energy Trust incentive for sealing air leaks in single-family homes in Oregon.
And yet, in Oregon, the phony blower door purveyors of Home Performance with Energy Star , have large rewards from public funds, that are denied to a true believer in home sealing. This in effect is a conscious and very great harm against me, where my superior work is acknowledged, as I foreswear blower door madness. It is an action of Gresham's Law , even though the rewarded competition charge ruinously in their fraud. Usually, with counterfeit, good is ruined because it is less profitable to the public.
At April, 2016, this post is in revision. Please be patient as I fix the presentation, until it is no longer based on the notion this is a 900 lumens downlight.
I have been installing these lights since mid-January, 2016, where they came in to Lowe's stores at about the first of the year with no fanfare. Appearance match to Sylvania lights they displaced is obvious, and I will not be the only person who thought they must be 900 lumens light, not the box-labeled 700 lumens.
The following table is true, taking momentary watt readings with a Kill-A-Watt meter.
I mustaccept that Utilitech 0752125 is 700 lumens. Maybe a bit more. It is strange then that it could draw up to fifteen watts, while the older-technology Sylvania 72089, also 700 lumens, draws ten or eleven watts. Surely the Utilitech is an energy waster somehow.
Some inefficiency is in an adverse profile. There is no progress in setting the light engine deeper in the luminaire. The number of diodes is reduced from about 26, to 9, while diode size is cut in hlaf. Surely diode luminance is up by an order of magnitude, at a time that luminance as blinding glare, is recognized as a product defect.
Installers will be happy with versatile features for retrofit in most can lights, now to be pulled up simply with springs engaging can harps. Can attachment clips identical to those of Sylvania packaging are also provided. This is the best packaging I have seen, for can retrofits. It is best packaging too, for clipping to a deep ceiling junction box.
A new concept for clipping to a ceiling junction box may be this light's best feature. Keys guide the luminaire to its point of spring-detent engagement. The junction box must be slightly recessed in the ceiling. The box here is RACO 175 in a mating mockup, without power leads.
With about 1/8" recess of the junction box, expect perfect mating with the ceiling. Know the luninaire is easily pried from its secure spring engagement. One should not need a pry tool.
Here is Sylvania 72089, 700 lumens, 10 watts, at left, and 1020 lumens 13 watt Nicor DLS56 at right, at full power through a Cooper DAL06P dimmer. I accept that the brightness difference is by ratio 1020/700, 46% brighter at right.
Here is Sylvania 72089, 700 lumens, 10 watts, at left, and Utilitech #0752125 at right, at full power through a Cooper DAL06P dimmer. The Utilitech is brighter, but by perhaps only 10%.
And at 4/6/2016 here is the same comparison, with Utilitech 0752125, date code 0216. This time, Utilitech 0752125 is at left. Sylvania 72089 is at right. The lights are much more nearly, equal. I am inclined to think product design changed at the 0216 date code, but that is not true although there are visible product changes in addition to advance of date code; the date code 0216 lens is more blue and shiny, less pretty.
Make a direct comparison, Utilitech 0752125 date code 0216 at left, date code 0915 at right. The date code 0216 light seems less bright by a negligible small amount.
Here is the 1020 lumens Nicor DLS56 at left and Utilitech #0752125 at right, at full power from a Cooper DAL06P dimmer. The Utilitech #0752125 is less bright by about the expected amount for 900 lumens rating. Both are pretty 3000°K.
Brightness Comparisons to "65 watt" flood bulbs:
Here and for all photos of this post, find original postings with captions in a Picasa web album .
Six bulbs in sequence are set at stage right in my comparison test stand, with offered advantage. The bulbs are closer to the screen by distance ratio 36"/40", and are not diminished by can light recess. In photos below, proceed counter clockwise from left. Utilitech #0752125 stage left. Comparisons are at full power except as noted.
Where reflector flood lights are taken as comparable to LEDs with 120° beam angle, expect Utilitech at about 900 lumens, to be brighter than floods bout 650 lumens, by about 40% . The comparison is difficult where color temperatures are not matched.
Stage right: GE Reveal Incandescent 65 watts actual.
Stage right: 65 watts incandescent, no manufacturer marking.
Stage right: FEIT Conserv Energy CFL flood, 15 watts.
Stage right, Commercial Electric (Home Depot) 16 watt CFL flood.
Stage right, 670 lumens 65 watts actual, GE incandescent flood. Again an incandescent flood gives decent light, though short-lived and expensive. Is it dimmable?
Incandescent floods are NOT dimmable. This comparison is at about 30% power.
Try one more light at stage right, a Philips LED PAR30 spot, 750 lumens, 13 watts, 3000°K.
Dim to about 30%, since the Philips LED spot is labeled "dimmable." I think a spot light should just be made smaller if less brightness is wanted. A spot bulb is not for practical general illumination. This LED spot was quite expensive and is inordinately heavy. Surely it would not survive a drop, while the Utilitech #0752125 will.
Where a 13 watt Utilitech #0752125 replaces a 65 watt incandescent flood, savings are by more than the 80% of ratio 13/65. If you dim to matched brightness, by 40%, the savings go up. Make a fuller cost comparison including cost of replacements in a matched 50,000 hour service period:
Relate these numbers to savings claimed on the Utilitech #0752125 package:
13 watts * 3 hr/day * 365 days * $0.11 per KW. Estimated Energy Cost $1.57 per year. This simple math is correct.
$286 lifetime savings vs. a 65 watt incandescent with life of 2000 hours.
My math has $569 - 252 = $317 lifetime cost difference where on average, LED dimming 50% or having half as many LEDs, gives match to dim lighting we used to accept, with 100 watt point-source bulbs. If in fact, doubled brightness is valued and maintained, savings are doubled. There are yet more savings where energy bleeds at hacked-in can lights are replaced with junction boxes.
Numbers like $327, or double that, are real but not easily understood. A basis for the $286 box claim is yet harder to understand where lights are not fully comparable. Savings more than ten times the product cost, sound good. Real savings are much bigger, and we need to see this.
I offer another use of the table, in seeing how far we have come in reducing operating cost of lighting. Divide the $569 Utilitech savings number by the total cost of getting same illumination from 100 watt light bulbs, $78 + 550 = $628. See that operating cost is down by 91%. I think that at our best, we will reduce that cost by 95%, where electricity is grid-supplied. Go to 100% with good LED lights and power self-produced off-grid. Lowe's is part of a really good thing, moving away from LED as dim, fragile, point-source bulbs. Let's celebrate. And, let's admit that in the useful math, we need Brightness Numbers. Fix the packaging error, with a new kind of number.
Controlling Luminance
A fully-revealed LED downlight should not be uncomfortable to look at. This is a design concern that has been neglected, perhaps until now. The reduced lumens output from same-size disk in the current Lowe's shelf offering, has more-pleasant luminance.
Luminance is the phenomenon sensed in looking at, a light. If the glare of individual diodes is knocked down by the obscuring lens, and is fully dispersed on the lens, this table works. In fact, there is some failure on both counts of the dispersion attempt. The individual so-bright diodes are not fully obscured.
The light emission is from a very tiny total surface are of diodes. Just nine, about 3/32" diameter phosphors.
Behind the lens, the diodes are screamingly bright. Is this the best we can do? Can't we get the same illumination from much-larger diode surface area? I imagine there is constant progress in product design. Yet, there are steps backwards with this product. Why should there be diminished efficacy, not still pushing toward or beyond, 100 lumens per watt? Why has the manufacturer in China accepted design that is objectionably noisy on a dimmer? Why can't we choose less-wasteful packaging, as is done with all Nicor LED downlights?
Should we speak of something like "perceived diameter" in the lighting density, luminance table? Here the light engine diameter (edge of circuit board), is 2.5", and might approximate what captures our gaze-at-the-light. This is interestingly similar to that of the good old 100 watt incandescent light bulb. Are we perhaps wrongly, again tying ourselves to past convention? If we do stick with the convention, then luminance of the 2.5" board, however packaged, might be limited to 100 watt bulb B4 illuminance, 450 or 500 received lumens over beam angle up to 180°. Luminaires at greater than B4 should have correspondingly larger perceived diameter, by an area ratio.
Lowe's surely is not "done" with this light. In future offering, might there be other options sharing the so-improved mounting features? I wish for options of 500 lumens and 1000 lumens too, at best-available efficacy, and with luminance never worse than in this offering?
I have gone further in study of this light, in blog post Luminaire Luminance Definitions .
Among 750 lumens LED downlights, initial offerings have been by far, easiest to look at. See that (lumens/ diode area sq in) of about 1000 was achievable. Convert to lux as * 1550 sq in/sq meter, 1.55x106lux. Perhaps have a limit of 1x106 lux at diode surface, and 1x105 lux at array size. Suggest that luminance of Utilitech 0752115 is poor design.
The light engine of a luminaire should be set near the lens, not at all recessed and with minimum loss of efficiency in opacity of the lens. Compared to first-version Sylvania 70732 or Sylvania 72089, Utilitech 0752125 is a big step in wrong directions.
The evolution of luminance in 1000 lumens luminaires is also in the direction of worsening glare.
