Hello. I have done some research on exterior wall setup/insulation, and HVAC in preparation for our new home construction. I think I somewhat understand the science, but opinions do vary a bit, and I wanted to run a few options by everyone to see if there are any gross problems or if one may be significantly better than another/most cost effective. Although, I am of the belief that I would rather sacrifice a few high-grade interior finishes to get better insulation. A few logistics. We are building outside of Toledo, OH (zone 5). Two story with vented attic, full basement, with attached garage. Approx 3300 sqft for 1st/2nd floors combined, 1500 sqft in basement, 846 sqft garage slab. We may finish basement at time of build or wait a year or so depending on cost/budget. It's out in the country with mostly propane heat in winter and electric A/C in summer. Natural gas in not currently available, but there is a pipeline about 1-2 miles south. Not sure if the company would run a line up just for me or if I would have to get neighbors to convert. Also considering geothermal (it's somewhat popular here, and we are building on 8 acres and have the space). Not sure what the best return is on HVAC vs. insulation- I lean toward insulation. Anyway, here is what I am thinking so far. We are doing 2x6 studs with structural sheathing (currently OSB, but may ask builder for cost difference of plywood). Option 1: Latex paint, drywall, smart vapor retarder (Certainteed Membrain or IntelloPlus), stud cavity (debating b/w Roxul R22 batts, R21 fiberglass batts, or sprayed FG (Spider)), OSB, housewrap, 1" XPS (R5), furring strips for ventilation and cladding attachment (likely fiber-cement siding). Option 2: same as option 1 with no vapor retarder and using 1.5" XPS (R7.5) instead of 1". This gets me at the code limit for exterior sheathing not needing a vapor retarder, but concerned about outward drying potential when needed. Option 3: same as option 1 with no vapor retarder and using 2" Roxul comfortboard IS (R8)". This gets me at the code limit for exterior sheathing not needing a vapor retarder, but with better vapor permeance for outward drying than 1.5" of XPS when needed. I am a bit concerned with window/door install for options 2 or 3 Basement Slab: 4" crushed stone, 2-3" EPS, 6 mil poly vapor barrier, 4" cement Basement walls: waterproofing layer on exterior then: Option 1: R10 foam on interior (not sure what type is best). Then furring strips/drywall or 2x4 wall with R13/R15 Roxul or FG batt, drywall Option 2: 1" XPS (R5) on interior, with 2x4 wall with R13/R15 Roxul or FG batt insulation, drywall Option 3: 2x4 wall with R15 batt Roxul insulation, drywall Sorry for the long thread. Thanks for reading. Any help would be great.

Try using a different web-browser so that the paragraph breaks show up. Without them it's pretty hard to read.

In option 1 the R5 exterior insulation isn't sufficient for dew point control in your climate, which is why you need the MemBrain.

In option 2 you have the bare minimum exterior R, but that R7.5 will degrade to about R6.5 in only 40-50 years, at which point it's marginal. If you're building for the long term you need more exterior R, or a different approach.

The "whole-wall R" of option 1 is about R19-R20 (depending on whether you opt for 16" o.c. or 24" o.c. stud spacing. Option 2 comes in at about R21-R22.

Option 3 comes in at about R22, but is a more resilient assembly.

You'd hit the same thermal performance as #3 with unfaced 2" EPS for less money and still have slightly more exterior drying capacity than in option 1.

For the foundation insulation...

#1 as-specified comes in at about R23- it willl work, but it's a bit overkill for a basement. You'd be better off applying some of the foam budget to fattening out the insulating sheathing for the first & second floors.

#2 comes in at about R15 after thermal bridging which is reasonable, but for the same money you could install R6 EPS or polyiso instead, without the environmental hit of the HFC134a blowing agent used for XPS. In your climate zone there is sufficient foam-R to not need an interior vapor retarder with that stackup- even R4 would be fine, but R6 is even better than fine- it has margin.

#3 is a mold disaster waiting to happen, since wintertime moisture will condense/frost on the surface of the above-grade concrete, and the stud edges below grade are somewhat susceptible to moisture wicking up from the footing (unless you provide a foolproof capillary break between the footing & foundation wall, and your waterproofing layer is perfect.)

The basement-R of #2 is about right for moderately priced energy, but if you're heating with propane you might want to step it up to #1, and go higher-R on the upstairs walls as well.

Geothermal prices (and contractor competence) varies- a LOT. Done right it may eventually pay for itself against high-priced propane, but there is some design risk there. If you are in a high-priced geothermal market (like I am), the geo money is usually better spent on getting to R30+ on the wall-R, and other envelope improvements to the point at which point the heating loads can be handled by ductless mini-splits, which will have comparable efficiency to middle-of-the-road geothermal designs, but with none of the design risk, and at a much lower upfront cost.

In any new construction it's worth orienting & siting the roof pitches for optimizing solar PV. At the rate the installed cost of PV is falling it will be cheaper on a lifecycle basis than residential retail electricity everywhere in the US, even without subsidy by 2025. (It's already cheaper than my retail electricty rates in high-priced MA). By 2030 it's widely expected to become the cheapest source of energy of any type, but if you oriented the roof the wrong way or built the house with large shading factors you won't be able to take advantage of that.

Interesting about the rates in MA; NH politicians still think solar is decades away.

Two points: whatever else you do, as Dana said, build so you can heat with electricity, which long before the mortgage is paid off, you'll be able to supply with homegrown fuel.
Second: on your inside basement walls install 2" polyiso; you can glue it to the wall or fasten it with special plastic fasteners & a hammer drill (actually pretty fast & easy). I've done numerous methods; this is by far the easiest and works the best. Also, be sure to separate the concrete floor from the footing and wall with 2" strips of EPS. Easy to do this thermal break now; impossible later.

Dana and Bob thanks for your info. I posted that from my iPhone, so I am guessing that messed up the paragraph formatting-sorry. Quick f/u though for Dana.

You mentioned as an alternative to my option original 3 for the walls you mentioned 2" EPS unfaced. I thought the labeled R of EPS was 3.5/in, therefore 2" wouldn't be enough to exclude the vapor retarder (need R7.5 minimum). Can you get EPS with a higher R value/in? Also, what type of EPS is best?

I am hesitant to go over 2" of exterior insulation as I don't know how routine it is for the builders in our area, and I don't want to make the window/door install overly complex. I have read here and elsewhere that up to about 2" of exterior insulation doesn't alter the window/door install much vs. routine install against wood sheathing.

I'll look more into the geo cost (i doubt it's as high as MA costs based on the relative number of people doing it), but will discuss with builder/HVAC about some form of electric powered heat.

Thanks.

Posted By Bob I on 10 Mar 2015 08:58 PM
Interesting about the rates in MA; NH politicians still think solar is decades away.

Two points: whatever else you do, as Dana said, build so you can heat with electricity, which long before the mortgage is paid off, you'll be able to supply with homegrown fuel.
Second: on your inside basement walls install 2" polyiso; you can glue it to the wall or fasten it with special plastic fasteners & a hammer drill (actually pretty fast & easy). I've done numerous methods; this is by far the easiest and works the best. Also, be sure to separate the concrete floor from the footing and wall with 2" strips of EPS. Easy to do this thermal break now; impossible later.

The levelized cost of rooftop PV @ 3.50/watt installed price (the current US average) on a 20 year lifecycle basis using a 3% discount rate and a 15% capacity factor is 19-20cents/kwh, and that includes an inverter replacement at 15 years.  If you expect a 12% capacity factor (conservative) it's about 24-25 cents.  (It gets a better if you assume a 30 year lifecycle.)  This winter's residential retail utility rates in MA are running about 25 cents. It'll probably drop to bout 20 cents in the summer when the next rate adjustments get implemented.

In Germany & Australia using the same panels, racking systems, and inverters currently runs $2-2.50/watt.  (It's not because German or Australian labor is cheaper or more productive than in the US- the bulk of the difference is in "soft costs", of marketing & sales, and the cumbersome multi-layered permitting & inspection processes in most US locations, where the utilities & regulators are dragging their feet.)  At $2.50/watt the 20 year levelized cost per kwh drops to about 15 cents, which is lower than the MA state average has been for a few years now.

PV has had a 40 year "learning curve" that reduces cost by 22% every time the installed base doubles.  The most recent 5 year learning curve has been north of 30%.  The doubling period is now about every 2.3 years, (and has been accelerating.) If we assume that the world price is $2.50/watt for PV is about right (it's actually closer to $2), and that the US can eventually hit the world price (likely, if the Aussies & Germans can/have), that means in the summer of 2017 the world price will be about $1.95/watt, and before 2020 it will be $1.52/watt, using the conservative 40 year learning curve, and no further acceleration of the doubling rate (which is super-conservative, since when PV becomes cheaper than the retail rate the doubling rate will probably soar, as it did in Germany.)  At $1.50 (the likely 2020 cost) the LCOE of PV becomes about 9 cents/kwh. 

What do you think the installed cost of PV will be in 2025?  Those who track these industry trends closely think it'll be essentially nothing by 2030- too cheap to matter- cheaper than $10/bbl oil.  The international investment bankers do too, even those bankers deep in the oil patch.

So, WTF do politicians in NH really know, eh?   They probably know a lot more about it than the analysts at Citi-Group, Bloomberg, Kepler Cheuvreux, USB, Sanford Bernstein et al who make a living/killing investing in energy infrastructure, right?

The bankers' assumptions are usually WAY too conservative in fast moving markets like this.  Kepler Cheuvreux is looking at $3/watt by 2030 in their transportation sector analysis against high-cost oil sources, which is laughable given that utility-scale solar in 2015 costs about half that, and rooftop solar will be half that before 2025.  There is rooftop solar being installed in the US for under three bucks a watt right now.



Many utility companies are hyper-aware of this, and what it does to their business models, which is driving them to take self-protective measures (sometimes with undue influence & subterfuge with regulators & legislators), but this has reached the point of inevitability.  Countries/states that take measures to slow this down or stall are only setting themselves up for economic competitive disadvantage down the road.

Posted By mmdoh on 10 Mar 2015 10:28 PM
Dana and Bob thanks for your info. I posted that from my iPhone, so I am guessing that messed up the paragraph formatting-sorry. Quick f/u though for Dana.

You mentioned as an alternative to my option original 3 for the walls you mentioned 2" EPS unfaced. I thought the labeled R of EPS was 3.5/in, therefore 2" wouldn't be enough to exclude the vapor retarder (need R7.5 minimum). Can you get EPS with a higher R value/in? Also, what type of EPS is best?

I am hesitant to go over 2" of exterior insulation as I don't know how routine it is for the builders in our area, and I don't want to make the window/door install overly complex. I have read here and elsewhere that up to about 2" of exterior insulation doesn't alter the window/door install much vs. routine install against wood sheathing.

I'll look more into the geo cost (i doubt it's as high as MA costs based on the relative number of people doing it), but will discuss with builder/HVAC about some form of electric powered heat.

Thanks.

Type-I EPS (the fragile 1lb per cubic foot density stuff)  is about R3.9/inch, Type-II EPS (the more rugged 1.5lbs density, the stuff they make insulated concrete forms out of) runs R4.2/inch. 

Even though you can get there with Type-I, it's easily damaged in handling unless it comes with foil or plastic facers, both of which would lower the vapor permeance too much.

If you want to stick with 2", fiber-faced 2lb roofing polyiso runs about R5.5/inch, and typically runs 0.5-0.9 perms.  In my neighborhood you can even buy used-once roofing foam from multiple vendors in good shape for about 1/4 the cost of virgin stock, making it cheaper than high density batts. One of those has national distribution with several storage & transfer facilities, but I'm not sure how close their closest depot is to you would be (but they'd be able to tell you.)  With the foam reclaimers it's good to order months ahead of time  (provided you have a dry place to stack it), since the inventory is all over the place, varies pretty dramatically from week to week, month to month.  Figure on a 10% scrap rate beyond your gross wall area (including the window & door areas)- not EVERY sheet will be perfect or salvageable, but you'll have some left over if you up-size by 10%.

"So, WTF do politicians in NH really know, eh?"
whatever ALEC tells them to "know"

there was an op-ed in the Concord paper this morning by a UNH Chemical Engineering professor emeritus who says that coal and nuclear are "here to stay", that they are "safe, reliable & affordable". And "renewable energy sources ...like wind & solar...are of little value when the weather isn't cooperating." Welcome to the alternative universe!

Dana: how are you pricing electricity - total bill/KW, or cost of KW only portion of the bill/KW?

mmdoh - sorry to hijack your thread, but it is this type of press that confuses us all and points us away from what is really happening to what the FF companies WANT to happen.

We've used type 1 EPS (either two or three layers of 2") under several floors with no issues, covered by heavy reinforced poly. I thought we would get a lot of breakage, but it doesn't happen.

Thanks. Also the garage is attached, and I am not conditioning it. What is the best strategy for insulating that (i.e. garage walls that separate outside air vs. garage walls that touch/separate conditioned interior rooms).

"interior" garage walls are identical to exterior house walls except possibly that the interior finish is fire rated drywall rather than sheathing. Same insulation and air sealing. Consider it an exterior space as it will be when the doors are open. Also consider installing a small exhaust fan tied to the door so that when you drive in you have a fan exhausting fumes and off-gassing that occurs up to an hour or more after the car is turned off.

Posted By Bob I on 11 Mar 2015 12:18 PM
"So, WTF do politicians in NH really know, eh?"
whatever ALEC tells them to "know"

there was an op-ed in the Concord paper this morning by a UNH Chemical Engineering professor emeritus who says that coal and nuclear are "here to stay", that they are "safe, reliable & affordable". And "renewable energy sources ...like wind & solar...are of little value when the weather isn't cooperating." Welcome to the alternative universe!

Dana: how are you pricing electricity - total bill/KW, or cost of KW only portion of the bill/KW?

mmdoh - sorry to hijack your thread, but it is this type of press that confuses us all and points us away from what is really happening to what the FF companies WANT to happen.

We've used type 1 EPS (either two or three layers of 2") under several floors with no issues, covered by heavy reinforced poly. I thought we would get a lot of breakage, but it doesn't happen.

I'm pricing the full retail cost of grid-sourced solar, not energy-only.  At some penetration level of distributed PV (say, 20-25% of all power on the grid) the value of the distributed solar falls low enough that "demand charges" would have to be applied to cover the grid costs, but until then there's a good argument for the simplicity of net-metering.   State regulators in some states seem to be more in the mode of protecting the interests of the utility monopoly over the interests of the ratepayers when (prematurely) attaching grid fees to PV owners, but I don't expect that to happen broadly in New England.  At current penetration rates of PV in New England the PV owner is providing more value to the other ratepayers in lowered peak power rates and deferred grid upgrade capital expenditures than the revenue stream they take away- they are in effect subsidizing the other ratepayers, though the argument from the utilities is that there is a cross subsidy in the other direction.  It eventually WOULD be a cross subsidy of the non-PV ratepayers to the PV owners, but we're nowhere near that yet.

Yes, coal and nuclear are here to stay, but that doesn't mean their capacity factors is bound to remain anywhere near the same.   The mid-day peaks in total grid demand will become mid-day sags, with a morning peak and an evening peak. That's not too different from the current shape of ISO-NE hourly demand curves except that the peaks will be lower, and the mid-day demand will be lower than the night time baseload, which is the opposite of what it is now.  This is already happening in CA & HI, more pronounced on some days than others, resulting in what has been dubbed the "duck curve" in CA, but the "Nessie Curve" in HI:


^^^ if it looks like a duck..... .......................................................................... or maybe an aquatic serpent thingy^^^


Solar detractors like to talk about how this is creating a big stability problem for the grid operators, but there are many cost effective ways of managing excess power and peak loads (which is why Germany's and Italy's grids are more reliable than the US, despite MASSIVE PV penetration).  But what it's really creating is a big financial problem for the baseload generators, especially less flexible and slow ramping baseload generators such as nukes & simple-cycle coal.  The CA-ISO grid operator has a mandate for a gigawatt of grid storage to be built by 2020. Originally criticized as overkill or over expenditure, when bids have gone out they have generally been over-subscribed and quite competitive- far more competitive than building gas-fired peakers to manage the ramp issues.  Companies in the electric car charging biz can also provide both massive power sinking and peak load management services to the grid for less cost than traditional utility-owned methods too.

When solar eats into the capacity factor of baseload generators it disrupts the financial model on which those capital assets were based. That cuts into generating company profits, but under rate decoupling most utility companies in the region have divested from most of their generators, which are now the assets of private companies. That's called economic competition, something that captalist economies are based on, but something that utility companies have largely been protected from. Vertically integrated utilities with lots of generating assets may be at risk. Whereas they had been pretty much guaranteed a profit on those capital expenditures in a non-competitive market, that won't be the case when widely distributed privately owned PV goes on the grid. So sure, there will be coal & nukes for a long time to come, but that time may be shorter than many people think.  Competition from midwestern wind has already retired nukes that were still fully licensed for operation but no longer economic to run at the low capacity factors they were relegated to in the face of extremely cheap wind power that ate up all of the nuke's peak-demand hours profits. Combined cycle gas plants ramp fast enough to manage the variations of wind power input in places like IA and ND, but the new power on the grid lowered the base load to where nukes had to power-dump during windy days to keep running, and the price of peak-demand power was severely curtailed with the new capacity.  Coal & nuclear power plants have to remain profitable to stay open, and competition from distributed generation makes that ever more difficult. Higher efficiency faster ramping combined cycle coal may still legs at least until carbon taxes get applied, or until buck-a-watt utility scale  or commmercial scale PV arrives (maybe as early as 2025), at which point they're toast.

There is significant risk of generating asset stranding with the rapid rise of distributed PV, for both the utilities (publicly owned or investor owned) and for privately owned power generators, which is why Barclays Bank downgraded the bond ratings of the utility sector broadly last year.  But if the regulators are doing their job they will be working for the ratepayers, not the utilities or generating companies.  There is a great need for regulatory reform to smooth out the transition period- NY has taken this head-on and the dust has yet to settle.  There was an attempt to deal with grid costs for PV in MA last year with a legislative proposal for a "minimum billing" to go into effect for grid-tied PV, but it didn't come up for vote before the session ran out.  It may come up again this year, but many are looking to see what happens in NY- there may be a better way than the minimum bill.   The way the min-bill was supposed to work was that in any month that a PV owner produced more than they used they were billed some fixed minimum charge, but the net excess power came off the total annual bill.  At total net-zero house would still be paying something for the grid, but only paying grid charges when their monthly output was more than they used.  (Some other states have tacked a fixed fee based on the total peak rated power of the array, independent of how much of that power was used on site,  which is effectively a tax on PV.)  The minimum bill has the virtue of simplicity, but it's not clear how the amount of minimum bill would be fairly calculated and adjusted over time.

The crazy one to watch is what happens to Georgia Power, which is a vertically integrated statewide monopoly. They have been allowed by their regulators to rate-base their new nuclear plants in advance of them coming online. As with most projects that size they're currently months/years behind schedule and billions of dollars over budget.  By the time the full cost of those projects get fully baked into the rate-cake some observers are speculating that it will become financially rational for those with access to both sun & capital to simply defect from the grid, since the cost trends on both PV & storage are still fairly steep declines.  It won't take very many grid-defectors to start the landslide, at which point they may well fail financially. 

In Germany one out of every five new PV installations is being sold with local battery storage (an artifact of the new requirements for the percentage of PV output that has to be used on-site), which is driving the cost of storage down.  Tesla (the electric car company) is building a new mega-factory in NV with Panasonic to make both car batteries and grid-storage batteries, initially to supply SolarCity, the largest US residential solar installer, that holds about 1/3 of the total market.  (SolarCity's board chairman is Elon Musk, founder of Tesla.  The CEO is Lyndon Rive, Musk's cousin.)   Since winning the right to install behind-the-meter storage in CA SolarCity's installations of battery-backed PV has taken off in that market.   They are also making a huge play in the Australian market, which has some of the least PV friendly net metering environments imaginable, but standard grid rates high enough that already market penetrations of PV have hit north of 20%.  The fact that PV output in Australia is only compensated at a fixed wholesale energy price (and in some instances, not compensated at all), even retrofit behind the meter storage on existing PV systems is viable there.  One reason the grid rates are so high in Australia is that the coal-fired generating capacity was allowed to keep climbing even as demand flattened and fell, and those generation assets operating at paltry capacity factors are rate-based, cooked into the retail cost of electricity (which is north of 30 US cents/kwh.)  Worse, operators of older coal plants are still being paid for their (unneeded) capacity by the grid operators, which is also rate based, which makes it more expensive for the owners of said plants to just shut them down.  Hopefully this isn't the scenario the professor emeritus envisions when stating that coal & nuclear are "here to stay" !

So with mega-battery factories going up in NV, Germany and China, how long will it be before cheap PV and cheap storage make the higher rates required to pay for Georgia Power's capital expenditures more expensive than grid-defection? How long before that makes New England grid-rates more expensive?  This IS coming, and utilities & their regulators need to stay ahead of it, or there will be a bleeding edge of grid defection that can snowball.  Utilities in Germany have failed on this basis, and there is no structural reason why it can't happen here.

Thanks. Is it worth air sealing with caulk/can spray foam or putting any batts in the stud cavity of the walls that divide the exterior air from the garage or just leave it empty?

  It eventually WOULD be a cross subsidy of the non-PV ratepayers to the PV owners, but we're nowhere near that yet.

Agreed, but people are making buy decisions about solar today and may be assuming that their net metering deal will continue for 20 years. So the issue needs to be discussed/made clear now.

--------

batts in the stud cavity of the walls that divide the exterior air from the garage

In a cold climate with salt on the roads, I prefer to keep the garage cold (in the Winter) and well ventilated. Ie, not insulated.

It appears that folks in Colorado is finally starting to have that conversation too, just like the folks in NY, CA, MA, HI, MN...

In most states net metering is grandfathered for 20 years, and can't be withdrawn at some later date. That doesn't mean that new PV gets the same deal as older PV, but the new deal is the new deal, there's financial certainty to it.

In a handful of states utilities have convinced the regulators that it's OK to pick away at the deal, or start adding connection fees to PV owners after the fact having the same effect as a bait & switch.  That's extremely short sighted and poor governance on the part of those regulators, but it happens. In that sort of regulatory climate when grid-defection becomes financially rational some will be ready to bolt!

Grid-defection kits are being advertized in Hawaii right now, and HI is actually pretty PV-friendly!  It's beginning to happen in Australia too, where the current administration is anti distributed generation, and very supportive of the coal mining industry (who curiously, are installing significant amounts of PV at mines to cut down their fuel costs for the diesel gensets. )  The states discussing it rationally well ahead of the arrival of the impending PV tsunami are more likely to get a smoother ride out of the transition than those who are stiff-arming the solar installer to prop up the status-quo monopolies.

http://www.greenbuildingadvisor.com/blogs/dept/musings/how-insulate-basement-wall

mmdoh,

As noted in discussions elsewhere, I have found this article a good review. Hopefully it will answer many of your basement insulation questions. Toward the bottom of the article are comments concerning basement stud wall fill.

I'm not meaning to hijack this thread with the solar PV discussion, but want to underscore just how rapidly the stuff is becoming cheaper.  This report just showed up in the Greentechmedia blog today ,with this graphic showing the average costs of different scale systems in the US as a whole, by quarter in 2014:



At the end of Q4 2014 the all-in pre-subisidy cost of residential rooftop systems averaged $3.48/watt.

Commercial rooftop systems were averaging $2.25/watt.

Utility scale solar farms were running $1.55/watt.

And it just keeps falling with no end in sight, and no technical barriers to slow that fall.  At the end of the blog piece the researcher who compiled the data was quoted:

"Total installation costs for utility and large commercial systems are now below 2011 module costs, underscoring the magnitude of the impact of falling module prices and incremental balance-of-system reductions."  (emphasis mine).

Module costs in the US are now running about 75 cents/watt.  Is anybody willing to bet that utility scale PV will be more than 75cents/watt (total installed cost) in 2020, or that residential scale won't be under $2/watt by then?

It may be a tiny fraction of the power generating market right now, but as manufacturing has scaled up the costs have fallen along predictable levels, and this industry is currently mushrooming at an accelerating rate, driving the costs down at an accelerating rate. Ten years ago even utility scale PV was costing 2-3x what residential scale PV costs right now.  Buck a watt PV at the residential scale may happen before 2025.

Any new construction that doesn't keep the ever cheaper PV factor in mind could be passing on the opportunity for super-cheap site-sourced power if the roof design is all cut up with hips, valleys & dormers, or the orientation or shading factors preclude photon-farming on the roof top.  While efficiencies are going up, they aren't going up anywhere near as fast as the price is coming down.  Barring a major breakthrough you'll still need substantial acreage of rooftop to hit net-zero-energy even on a fairly well insulated house, but the up front cost of hitting net-zero will be dirt-cheap long before your first set of shingles are toast. 

So, design & orient your house to be ready for it!

Buck a watt PV at the residential scale may happen before 2025...and no technical barriers to slow that fall

If you don't account for the costs of storage or grid disruption. Luckily, thermal storage is fairly easy.

The notion that PV needs to be coupled with storage for a stable grid even at 25% penetration is not supported by real-world experience. At higher penetrations there can be costs associated with PV curtailment or small amounts of local storage, but those costs are still pretty small until the mid-day peak outputs are a significant fraction above the local-grid load. Those limits have been tested in a few neighborhoods on Oahu, and there are multiple solutions available with dramatically different cost structures. A fleet of electric vehicles (even with only 1-way power, grid-to-car) isn't particularly expensive to incentivize at $4/gallon gasoline (a bit tougher at $1.59 gas), and the additional cost of adding smarts to car-chargers is also quite low. On an island grid with expensive gasoline it won't take a huge subsidy to get smart car-chargers & electric cars plugged into the local grid feeders that have excess PV capacity and potential backfeeding issues, if different subsidies were allowed based on local grid needs (as will soon be the case in New York, but as yet uncertain in Hawaii.)

I'm sure there are ways to make high grid penetration PV expensive, and there is a lot of incumbent-generator/utility press spelling out how to do it in the most expensive manner possible, as if that's the only way to do it, but they're simply wrong. The grid operator can avoid most of the capital cost of the solutions if done intelligently, but in a regulatory environment that compensates the utility for their capital upgrades there is no incentive to do it that way. And, people can & will take it into their own hands if the utility or regulators get in the way. In Australia people currently get around the fact that power going from their PV onto the grid is essentially uncompensated by the utility with a cheap thermal solution: Shunting any excess power they may be producing minute-to-minute into a heating element in the hot water heater, a hack with a hardware cost of about $100. That's an ugly and less flexible use of the output, but if all kwh drawn from the grid cost the same, it's the right thing for the PV owner to do.

But the cost of battery storage + PV is already below the residential retail cost of power in Australia, which has piqued the interest of Solar City and others who already have pre-engineered, tested, safety-certified residential scale solutions for grid tied batteries. That party is just getting started, but by the time PV hits 25% penetration in Kansas the cost of storage + PV will likely be below the residential retail rates in most of the US. The rate at which electric vehicles show up on the grid relative to the build-out of distributed PV will affect some of the local-storage economics, but both are coming.

The traditional utility business model will essentially be dead in the next decade or so, but there is as-yet no clear simple business model that has replaced it. It remains to be seen what NextERA will do withe HECO, now that they have bought up the Hawaiian utility, and it remains to be seen how the massive regulatory revision in NY will shake out as PV and local storage takes off in a big way there. But I expect the lights will still be working throughout the transition.