Cold sheathing: condensation & mold?
![](https://secure.gravatar.com/avatar/a60da5df5f8ca9470998cdc3fcae274e?s=52&d=https%3A%2F%2Fimages.greenbuildingadvisor.com%2Fapp%2Favatars%2F200x200%2Fb.png&r=g)
Hello,
I have read most of the other threads on this topic–and would like to complement Martin and others who have provided great, detailed answers with patience!
My situation is similar but slightly different and I would appreciate some assistance:
- Home in the hills of Seattle, WA (elevation 800 ft, so colder winters than typical for Seattle metro area)
- Built in 1998. From outside-in the building envelope is natural wood siding (cedar/doug fir?), tar paper, 5/8″ OSB, 2×6 wood studs, R19 fiberglass. Solid slab foundation.
- The photo attached is from an approx 400 sqft room on the first floor, a half-basement by the technical definition. In this room, approx 1/3 of the wall space is a retaining wall, 1/3 interior walls, and 1/3 exterior walls.
- The builder left the room unfinished, so on the exterior walls and the retaining wall sides, R19 bats with the reflective Kraft paper were installed (although sloppily stapled). The ceiling is left unfinished as well, which has been useful over the years for access to wiring, running a new A/C line, etc..
- About 5 years ago, we installed a Honeywell humidifier to the main furnace (it was also replaced with a 95% efficient unit that vents to the outside) to increase ambient air during the winter when the blown air would cause dryness….to reduce colds/flu during winter months with small kids at home. We have traditionally set it to 50%, with a measured R.H. of around 46-47% in any given room. I have since turned the unit down to 35%, although we’d like to operate it around 40% if possible going-forward.
- In this HVAC room is installed this furnace and two gas hot water heaters, and the usual two 8-10% diameter conduits to the outside to provide fresh air venting for the hot water heaters (and former furnace)
We planned to finish this space to add value to the house by installing drywall and painting it, leaving the ceiling to be installed later likely using a drop-tile system so we can still access the overhead wires, etc.
Preparing the room, i removed the silver kraft paper so the drywall could attach to the studs with a better seal and also started to install a new power outlet in one of the bays when i noticed the dampness behind the R19 batts. I checked with a moisture meter, and 80% of the bay was 66% or higher. It seemed the moisture and mold were heaviest towards the bottom of each bay, with less moisture and mold at the top of the bays and near the edges where the insulation was less dense. I removed all insulation and every bay had moisture and mold (about 350 sqft).
Interestingly, after a panicked weekend a few days ago thinking we had an exterior leak and much research on the internet and from GreenBuildingAdvisor, came to the consensus that it’s a Cold Sheathing condensation problem. Played with a Dew Point Calculator and when I re-measured interior surface temperature of the OSB behind a (reinstalled) R19 batt, was around 43 to 46-degrees F, below the dew point for our 68 degree, 47% R.H. interior space. Things i’ve never known and learned!
So, my questions are these:
- Why dont i see moisture levels or mold in the 2×10 dimensional lumber blocks between the floor joists (see in the photo) since they too were also behind the same R19 batts?
- Should i really believe that the other exterior walls in the house that are finished dont have moisture or mold in them? Yes, we have at least two coats of high-quality latex paint of them (Devine is the brand), but i’m sure they must be somewhat leaky given what i have seen with some of the other construction over the years at our home.
Our remediation plan underway is to kill the mold (done, Microban Disinfectant Spray Plus), dry out all materials with heat and dehumidifers (underway), then re-install the R19 batts, add drywall (probably will use the mold-resistant version), mud&tape, then apply a PVA primer followed by two coats of latex paint. One contractor suggested we also put down a layer of 4mm plastic on the studs before the drywall to minimize re-introducing moisture from the mud/tape and paint to the bays given its December–but that seems suboptimal from what i read online.
And, our plan is to keep the ceiling as-is, meaning those 2×10 blocks between floor joists will still have R19 but be otherwise exposed to the warm interior air. Seems like these spaces should have problems exhibiting the same symptoms as the other bays, but i’m not sure why not…
Appreciate your detailed reply–this is a great, fun educational exercise for me and for the contractors i’m working with, although with a moderation of concern given the mold and moisture accumulation.
-Brian
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
![](https://www.greenbuildingadvisor.com/app/themes/greenbuildingadvisor/dist/img/modules/detail-library/thumb1.png)
![](https://www.greenbuildingadvisor.com/app/themes/greenbuildingadvisor/dist/img/modules/detail-library/thumb2.png)
Replies
Brian,
First of all, you're right about one point: many U.S. homes have hidden mold on the interior surface of the OSB wall sheathing.
Warm sheathing will always be dryer than cold sheathing, which is why the best solution to this problem is to install an adequate thickness of rigid foam or mineral wool on the exterior side of the wall sheathing. For more information on this problem, see How Risky Is Cold OSB Wall Sheathing?
In your case, the problem has been exacerbated by two factors:
1. You have been running a humidifier -- which is always a mistake. If the interior air in your home is too dry during the winter, that's almost always a sign that your house has air leakage issues. The solution is to plug the air leaks (using a method called blower-door-directed air sealing), not operating a humidifier. Operating a humidifier is extremely risky. For more information, see Air-Sealing a Basement and Air Sealing an Attic.
2. You had no drywall on the interior side of your studs, so the warm, humid, indoor air had unimpeded access to the cold sheathing. At the same time, the foil facing on the fiberglass batts limited the ability of the wet sheathing to dry inward. (Your 2x10 rim joists aren't as moldy because it's easier for the rim joists to dry inward.)
Once you have lowered the humidity of the room, and once you have convinced all of the members of your family of the importance of destroying your humidifier with a sledgehammer, you should probably consider insulating these problematic stud bays with closed-cell spray polyurethane foam. The foam will provide an effective air barrier to prevent the interior air from contacting the cold OSB.
Wintertime dew point averages at ~800' of elevation in the Puget Sound region are about 36-37F, which corresponds to a healthy & comfortable 30% RH @ 70F. Pull up a dew point graph for say, Enumclaw as a representative example:
http://weatherspark.com/#!dashboard;a=WA/Enumclaw
The only way you would ever need to add humidity is if you were ventilating at ridiculous rates during the middle of a cold snap. Normal human activities like bathing/cooking/breathing will add humidity to the house, bringing average indoor humidity above that of the outdoor level. Anything over 50% RH indoors increases health issues for the human occupants- dust mite populations can get established, and skin fungus infection risks increase). Under 30% and airborne virus transmission rates increase, but anything in between is healthy for YOU...
But not for your house. If you humidify to 50% RH @ 70F the dew point of that air is about 51F, but the mean outdoor temp at 800' in winter is about 39F (pull up a temperature graph for Enumclaw as a representative example), which means the cold sheathing is going to take on significant amounts of moisture. Rather than adding humidity, you should limit the the interior RH in winter to no more than 35% @ 70F (dew point = 41F) via ventilation &/or dehumidifiers, so that the interior side of the sheathing won't dwell far enough below the dew point of the interior air long enough to take on much moisture. Latex paint is somewhat vapor retardent (3-5 perms), and is sufficient to limit the rate of moisture accumulation to non mold-inducing levels if you keep the interior at 35%RH- even 40% would be fine during the milder winter weather.
So yes, your mold problem WAS created by the humidifier- the reason it wouldn't actually quite hit 50% is because the OSB was soaking it up as fast as it was being delivered, along with whatever modest reductions you were getting from ventilation/infiltration.
The kraft facers on R19 batts (the crummiest fiberglass insulation ever sold) are sufficiently vapor retardent that you needn't sweat the drywall mud-moisture- just be sure to give it plenty of drying time before painting it. If you want to really reduce the mold risk, dry-blow cellulose in mesh (roll it flat to the stud edges before putting up the drywall) instead of using batts- the hygric buffering of the cellulose will safely share the moisture load with the OSB, and the borate fire retardents used in the cellulose are also powerful anti-mold-anti-fungus agents. The air-retardency of cellulose even at 2lb density is about 90% better than that of low-density fiberglass, (at 3lbs+ density it's about 99% more air-retardent) nearly eliminating the convective transfer of moisture from the interior side to the exterior side to boot.
Regarding R19 batts: They only test at R19 at their pre-installed ~6" loft. According to the manufacturers' own compression tables they're no better than R18 when compressed to 5.5" in a studwall, and that's only if they are installed perfectly with no gaps or compressions. But because they are so low-density they suffer a measurable loss of performance to convection within the batt at the temperture extremes. An R19 is essentially a "fluffed" R13 batt- it has the same weight per square foot as an R13. R13s compressed into a 3.5" cavity actually perform at R13 at the ASTM C 518 test temperatures, and are far more air-retardent. Rather than losing performance at the temperature extremes, they gain a modest amount (as does cellulose at any density.)
If a batt solution is the only option, go with R23 rock wool, which is nearly as air-retardent as 2lb cellulose (if perfectly installed.) Alternatively install unfaced R13s and compress the reinstalled R19 into place, at which point you'll have ~R20-R21 performance at 5.5", and sufficient air retardency to maintain that performance, limiting the rate convection. But bear in mind that there is no moisture buffering with either rock wool or fiberglass- cellulose really is the preferred product here.
A flash-foam of an inch closed cell as an air-seal an vapor retarder on the OSB would be sufficiently protective to then compress the re-used R19s in place, but it's expensive- a buck a square foot or more per inch of depth. Any more than an inch would limit the rate of drying toward the interior, which can be important in the rainy foggy-dew PNW if your roof overhangs aren't deep enough to keep from rain-wetting the cedar siding.
Thanks for the detailed replies. I spoke with the architect we have used on a separate project and he agrees with most all of the content in this thread. One recommendation he had was to remove the kraft backing (if we keep the R-19 batts, i'm open to replacing) since that paper if not properly secured under the drywall can create vertical channels for the air to flow and route to. Thoughts?
As for the cellulose insulation option, my naive understanding would be that it would both compress more over time and is also of organic matter and thus could pose more of a potential mold-thriving ground than other options? (not appreciating how effective borate fire retardants are in this insulation option). The recommendation the architect suggested if we dont just keep the (unfaced) R-19 batts is to move up to new rolls of R-21 unfaced batts which are now the local code, and fit into the 5.5" bays. Or go with R-25 (?) batts and compress them a bit with the drywall to fill out the stud bay cavities. I have good experience with the sound-damping Roxul product for interior walls, so also wondering how that might work for their insulating product?
It seems the blowin-in foam would be best, and we have approx 315 sqft of wall space, but i'm not sure how easy it would be to get an insulation contractor to do such a small job without the costs being exorbitant given our budget.
As for the dehumidifier, it's now unplugged (not sledgehammered yet, sorry Martin--photos coming soon if this all works out!). Also, our plan as of last night is to replace the two hot water heaters with ones that dont need the two 8" high/low outside air vents coming into this room (with a tankless system or different tank system), and instead then route those two 8" vents into the HVAC furnace return air plenum to create more of a positively-pressurized house (per the great documents you have on the site).
As for the floor-joist rims (that's what i call them--it's the 2x10 dimensional boards that 'cap' the end of the floor joists that adjoin the OSB sheathing on the outside)...i'm still a little perplexed why they are not exhibiting signs of moisture or mold given they too were behind just the R-19 kraft bats (small pieces, of course, 16"x10" approx for each 'bay')? With only an R-value of ~1.8, they clearly arent doing much. Is it that they are providing a much tighter seal to the plywood for the floor above and to the top-plate of the 2x6 stud walls and thus provide at least a decent air barrier to the OSB behind them? And, that a a soft wood, they are absorbing some of the vapor coming from the inside of the room before that vapor gets to the OSB? Or, that we likely have moisture and mold on the OSB behind the dimensional wood and just wont ever see it?
I'm presuming to address these floor-joist bays, if we cant otherwise do the 1" of spray foam, then a good caulk bead should be sufficient, and then just put back the kraft R-19 pieces?
Lastly, what would you recommend for a waiting period from the time the drywall is finished (6-day process to get to textured surface) until painting? I expected we'd do this fairly soon thereafter. The drywall installer had planned to apply the PVA primer before spray-texturing and using one of the variants of hot-mud for quicker drying and use of less water--any issues with this approach?
Thanks again!
Brian,
If you connect two 8-inch-diameter outside air ducts to your forced-air distribution system, your house will be wildly overventilated (and dry). If you want to install this type of ventilation system -- called a central-fan-integrated supply ventilation system -- you need to include a FanCycler control, and you need to include motorized dampers and a commissioning process that verifies the air flow. For more information, see Designing a Good Ventilation System.
If your rim joists were insulated with R-1.8 insulation, then the insulation wasn't doing much. That means that the rim joists were warm -- and therefore didn't absorb moisture or condensation. For more information on rim joists, see Insulating rim joists.
Martin,
What i had intended to ask is why would our rim joists not have any signs of moisture or mold even through they sit right in front of the OSB sheathing and behind the same R-19 batts? (note: there is no R-1.8 insulation of the rim joists, just the outside house envelope with the OSB...the R-1.8 is the estimated value of the side of the rim joists).
So--how could the surface of the OSB be hitting temps lower than the dew point but not the rim joists right above them?
On the FanCycler control, we have that now connected to a 7" duct; we have programmed the control to open it as much as the programming will allow (i think 50% of a 24hr period). However, the concern that is not addressed in that document is coordinated timing: there is no guarantee that when systems that evacuate internal air (ie, bathroom fan during a shower, whole-house fan, cooktop, ...) the external damper via the FanCycler will be open, thus we cant ensure the house will be consistently positively pressurized. So, the proposal is to always be pulling some amount of outside air into the central system (at the expense of increased energy consumption, understood).
Peeling the kraft facers will do nothing good for the performance, and would only make it harder to install. To install it correctly you'd normally press it in at the corners and edges to make sure it's not leaving any gaps in those corners, then tug it back lightly to where the facer is just proud of the stud-edges, then compress it in place with the wallboard. Any electrical boxes should be pre-sealed with can-foam on the back side, and the batt needs to be trimmed to near perfection to fit around it. Wiring that penetrates the framing needs to be foam-sealed at every framing penetration, and the batt needs to be split, filling completety around the wiring with a minimum of gap. If you're not flash-foaming the OSB, seal the framing to the studs and stud-plate with a bead of acoustic-sealant caulk (which never gets brittle or hard), and the seams between any doubled-up plates. (This is a bit less critical if going with blown cellulose, since the pressure of the installation drives fiber into any leakage points. But it's still best practice to air seal the cavity, since you have the easy access now.)
Cellulose a 3lbs density or higher in a PNW climate will never settle. Settling in cellulose is a function of the depth of seasonal moisture cycling- the deeper the cycling, the higher the density. Your climate is SO temperate the indoor and outdoor humidity averages operate in such a narrow zone (compared to climates with cold dry winters and warm humid summers) that that settling is unlikely even at 2lbs density, but a minimal-dense-pack of 3lbs is more air retardent than 2lb cellulose.
The cellulose itself can take on significant moisture without reaching mold potential. The borates would be considered only "backup" from a mold point of view, but necessary for fire resistance. These products have a long an successful history of being protective of the structural timbers in many climates far more severe than yours. In a deep energy retrofit was involved with on a 3 story balloon framed house in central MA (US climate zone 5A), and when the walls were gutted the 30+ year old ~2lb density cellulose was still intact- fully filling the cavities, and self-supporting, and the framing timbers & sheathing were mold free, despite having no interior side vapor retarder. (I've seen sloppier lower-density installations that settled several inches in the first 10 years though.) In the retrofit the perfectly good cellulose got removed in order to better air-seal it, but it was definitely doing the job, and doing it quite well, despite being at a less than ideal density for the climate.
For only 315 square feet you could buy a 200 board-foot foam-kit and do your own flash-foam- keeping it a bit shy of an inch. One vendor among many: http://www.tigerfoam.com/products.php A full fill of open cell foam installed by a would run about a grand, and is to vapor-open to be protective. ( Even a half inch of closed cell 4-5x more vapor retardent than 5.5" of open cell.) Open cell foam would be better than R19 batts alone, but not as protective as cellulose, or a flash'n'fill of closed cell foam and fiber of your choice.
The surface of OSB takes on moisture much more rapidly than sawn timbers due to the much higher fraction of exposed end-grain. From the side grain a 1" plank of doug fir is about 1 perm, but if you cut a 1" slab across the grain it's more than an order of magnitude higher. (That's why logs will check and split at the cut ends if you don't seal it with something to limit the drying rate.) With OSB it's somewhat random grain orientation, but the exposed end grain on the surface chips of OSB are a sponge compared to the sidegrain of a sawed plank. (It's the binders used in OSB that reduces it's permeance from one side to the other to about a perm, or a bit less.) That's why you often get mold on the OSB, and none on the rim joist. Since the face plys of plywood are always side-grain, it takes a bit more for mold to get going on plywood than it does on OSB too.
It's possible to have mold on the OSB on the other side of the rim joist, but if you caulk the seams with acoustic sealant (or flash-foam with closed cell) to prevent air-transported moisture & mold spores you will have mitigated any unseen problem that may or may not be there.
There is no practical difference between Roxul Safe'n'Sound and it's thermal insulation products- only the dimensions are slightly different. But like fiberglass batts the tuck & tug to ensure full fill with a minimum of voids is critical in thermal applications, and not so much in acoustic installations. Acoustic batting is designed to NOT be pressing on the surfaces, which would give it a somewhat higher mechanical coupling, whereas thermal batting is designed to be a compression fit. Thermal Roxul designed for 2x6 framing is R23 (and pretty good stuff IMHO.) R21 fiberglass is far denser than R19s and far denser than the near- criminally-labeled R22s that only perform at R19 when compressed to 5.5"from their 6.5" manufactured & tested loft. (Yes, R21 fiberglass outperforms R22 fiberglass when installed in 2x6 wall- go figure!)
While mold problems on band joists are rare with just cut-up R19 stuffed in there, the true performance of the R19 is nowhere near the labeled value since it's nearly impossible to air-seal the facer at the joist-bay ends, and convected moisture is guaranteed to be getting in there.
Ventilating a house with the furnace is a common but pretty lousy way to go about it, since you only get the ventilation when it's running, and the ventilation rates go up when you least want it (when the outdoor air is colder and drier.) And as Martin points out, couple of 8" holes would likely yield GIANORMOUS ventilation rates when running.
I'm not clear why you would need two water heaters, unless you have some monster-spa tub to fill(?). If it's for showering capacity, you can get the same or better capacity at half the energy use by installing the biggest drainwater heat exchanger that fits downstream of the shower(s). If there are local subsidies for them the smallest 48 gallon condensing tank type hot water heaters out there have nearly 2x the BTU output of a standard 50 gallon tank and much shorter recovery periods. In combination with drainwater heat exchangers you can support two simultaeous showers 24/365 if you wanted to, for the same or lower installed price of a typical gas fired tankless (and much higher net efficiency.)
Dana,
Great answers, thank you! So in our situation, if you were to choose between the cellulose batts and the Thermal Roxul for 2x6 framing @ R-23 for these exterior walls, which would you choose?
For the band joists, would you recommend anything different other than the acoustical caulk and putting the cut-up R-19 back into place? (ie, replace those with the same cellulose pieces or Thermal Roxul, spray with PVA primer, etc...)
We have used acoustical caulk in another part of our house and i recall it having a strong odor/off-gas. Do you know if this lessons over time while still maintaining is flexibility?
On the hot water tanks, the home is approx 4,500 sqft, 4bd, 4.5 bath. With relatives visiting, we have in the past run out of hot water. I will certainly look into the drainwater heat exchanger + condensing tank--do you have a link for those I can review? I'm not sure we have access to the drain pipe for just the shower(s) and/or how that grey water mixes with the sewer lines so i'm not sure that would be something we could retrofit but i'd like to consider it.
On the two 8" outdoor air returns, i hear the feedback and acknowledged. That said, should the damper be barometrically controlled vs electronic timer to adjust for the random-hours in which inside air is vented out of the house?
Ah the things one learns and can do differently with home N+1!
Brian,
Q. "For the band joists, would you recommend anything different other than the acoustical caulk and putting the cut-up R-19 back into place?"
A. Yes. As I did in my last response, I'd like to refer you to my article on the topic: Insulating rim joists.
You don't want fiberglass insulation at this location; you want rigid foam or closed-cell spray foam.
Re: FanCycler
Martin mentioned a commissioning process for the FanCycler that involves verifying airflow. It is unlikely that "open it as much as the programming will allow" is the right setting.
The FanCycler is intended to assure that the house achieves a specified amount of ventilation. The ideal number is something that needs to be tabulated/calculated to match the house, and represents a compromise between many factors. There are several devices that implement the same principles, but the ideal FanCycler type apparatus aims to:
Opportunistically open the outside damper for a certain period of time per hour while the air handler is naturally running (i.e. during heating or cooling) in order to let in a specific amount of air.
Compel the air handler to run for a certain for a certain period of time each hour if it won't do so naturally (i.e. no heating or cooling calls), so that it may provide the prescribed degree of ventilation and also provide some mixing of the indoors air that otherwise wouldn't happen
As you can see, It's intended more as a means to meet a specified target level off ventilation, rather than provide constant positive pressure to the house. The fact that it does induce a small degree of positive pressure is a side-effect of how it happens to operate, rather than a design feature. So opening the supply vents or adjusting the programming to force lots of outside air is kind of a bad thing. It will lead to over-ventilation, and won't really do anything useful or consistent as pressurizing the house goes. I'd say it's best to just calculate the "right" level of ventilation for your house, and program the FanCycler to aim for that.
Aaron,
Thanks for your comments; I agree. Your comments led me to re-read Brian's last post, and I concluded that Brian (once again) ignored my recommendation to read the linked article.
Brian,
Q. "On the two 8-inch outdoor air returns, i hear the feedback and acknowledged. That said, should the damper be barometrically controlled vs electronic timer to adjust for the random-hours in which inside air is vented out of the house?"
A. Read the article that I linked to the last time you asked the question (Designing a Good Ventilation System).
The article explains:
"Central-fan-integrated supply ventilation systems ... can only be used in homes with forced-air heating or cooling systems. The systems include three important components:
A duct that introduces outdoor air to the furnace’s return-air plenum;
A motorized damper in the fresh air duct;
An AirCycler control to monitor the run-time of the furnace blower and to control the motorized damper.
"The AirCycler control (also known as a FanCycler) prevents both underventilation and overventilation. When the AirCycler notices that the furnace fan hasn’t operated for a long time, the control turns on the fan to prevent underventilation. When the control notices that the fan has been operating continuously for a long time, the control closes the motorized damper to prevent overventilation.
"During the swing seasons — spring and fall — the furnace blower will need to operate for ventilation purposes. In most climates, about 15% of the annual blower run time for such systems will be devoted to ventilation only. If the system is properly commissioned, the furnace will supply a 7% outside air fraction during ventilation mode.
"The big downside to central-fan-integrated supply ventilation is that the installer needs to understand how to design and commission the system. HVAC contractors capable of this task are rare. Unless the designer of a central-fan-integrated ventilation system takes great care when specifying the furnace and programming blower operation, such a system can have unreasonably high operating costs.
"A well-designed central-fan-integrated supply ventilation system needs a furnace with an energy-efficient ECM blower. Such furnaces cost between $1,000 and $1,500 more than conventional furnaces. If you end up using a furnace with a conventional blower motor — that is, one that draws 700 to 800 watts — the ventilation system will incur a big energy penalty. (For purposes of comparison, a Panasonic exhaust fan draws 11.3 watts, and most HRVs draw 100 watts or less)."
In re drainwater heat exchangers, EFI is the US distributor for one of the major vendors, and has the best pricing:
http://test.efi.org/sites/default/files/power_pipe.pdf
You can order them through Home Depot or direct from the manufacturer, at a more retail-type price.
Since it is a fully double-walled separation between the drain and potable, they meet code even for black-water drain mixes. Anywhere downstream of the showers works, but it has to be mounted very close to perfectly vertical to get the full performance out of it.
Natural Resources Canada maintains an apples-to-apples model-by-model third party tested list here:
http://oee.nrcan.gc.ca/residential/personal/retrofit/13302
Basically the longest and fattest one that fits is the right one, since both diameter and length increase the heat transfer surface area. It's sometimes easier to have the output of the heat exchanger feed all of the cold supply of the house rather than just separating it out for just the water heater & shower. The output isn't much above room temp even in summer, so don't worry too much about tepid cold water at other taps while someone is in the shower.
The AO Smith Vertex AO Smith GPHE-50 is the cheapest condensing tank out there that I'm aware of ($1.5-1.8K internet/street price) and the smaller burner one has a 76KBTU/hr burner, which is enough burner to sustain one shower forever even without the drainwater heat exchanger. As long as you don't have any big soaker tubs that needs more than 50 gallons in a gulp, it'll be hard to run out of water with one of those and a drainwater heat exchanger, and it'll fill a tub faster than a 199KBTU/hr tankless. With 2x the burner output of a standard 50 gallon tank, the recovery time is cut in half too.
For more money there is a 100K burner version (called the Vertex-100)- not sure it it's worth it in your application. For a lot more money you can get an all stainless condensing Polaris, which comes in a number of different sizes (burner sizes start at 100KBTU/hr tank sizes start around 35 gallons), and would last 20+ years in most homes (to maybe a dozen years for the Vertex).
http://www.aosmith.com/About/Detail.aspx?id=188
http://www.aosmith.com/WorkArea/linkit.aspx?LinkIdentifier=id&ItemID=216
http://www.amazon.com/Polaris-Heater-34-100-2NV-Natural-Residential/dp/B002ZXSNOK