Here is an article just posted on tankless water heaters.

I want smart green, that makes financial sense for efficiency, not just green for green sake.

What is the experience with tankless w/ everyone here?




Consumer Reports: Tankless Water Heaters Efficient But Not Necessarily Economical
Posted By: Jamie  on 01/06/2009

Heating water accounts for up to 30 percent of the average home's energy budget. Some makers of gas-fired tankless water heaters claim their products can cut your energy costs up to half over regular storage heaters. So is it time to switch?
----------
Source: Consumer Reports

Photo courtesy of Vacationtime on Flickr.

Probably not. Gas tankless water heaters, which use high-powered burners to quickly heat water as it runs through a heat exchanger, were 22 percent more energy efficient on average than the gas-fired storage-tank models in our tests. That translates into a savings of around $70 to $80 per year, based on 2008 national energy costs. But because they cost much more than storage water heaters, it can take up to 22 years to break even—longer than the 20-year life of many models. Moreover, our online poll of 1,200 readers revealed wide variations in installation costs, energy savings, and satisfaction.

Basing the numbers on an EF test is a bit silly- it really depends on how you use water, and how MUCH!

IIRC, the EF test is 6 equal draws of 10.7 gallons taken one hour apart, which is EXACTLY how you use water, right?  (Riiiiggghht... thought so! :-) )

In reality, if you use less than 60 gallons of water/day, your tank HW heater will perform much LOWER than it did in an EF test.

Similarly, if the bulk of your water is a gallon or less at a time (say, primarily for hand-washing, and you only wash clothes or take baths once/week), your on-demand won't meet it's EF numbers, may even perform lower than a decent tank.

There's a ton of information (prolly more than you ever wanted to know :-) ) on the subject here:

http://www.energystarpartners.net/ia/Water_heaters/Documents/WHPAGette_Final.pdf

For the true tankless perfomance, take a look at Fig 5.1 (p. 48 of the .pdf, or 42 of the printed document).  At 2 gallons/draw it's no better than a tank's EF performance, but it'll clean the tank's clock at anything over 5 gallons/draw.

So, it kinda depends- if you're on the go, living alone (or a water-sparing couple), often take weekends away etc, you'll likely meet or beat your tankless EF numbers by a few percent, but won't even come CLOSE to your tank's (lower) EF numbers.  If you're a mostly stay at home family of 5 you'll likely meet or exceed the EF numbers for either, and the Consumer Reports analysis is (or could be) correct. (But check your actual fuel costs- they vary considerably with location.  Here in MA we pay 2.5-3x the rate for natural gas as they do in UT.)

But the real reason to go tankless is to save your marriage- you never hear the screech from your spouse about the cold shower (did they factor in the cost of a divorce in the financial model for the tank? :-) )

But...

...there's another option:

If your space heating system is done with a mid-efficiency (83%+) or higher hydronic (pumped hot-water) boiler, an indirect-fired HW heater running off the boiler will give you similar performance for less money. Done right, it should be capable of keeping up with continuous-demand of back-to-back showers (even with guests), just as a tankless would.

[edited to add] Having just read the full text of the article, the tests THEY performed was a heavy-use scenario, (heavier than the EF test), which will increase the relative performance of the tank, minimizing the performance savings delta. Low volume users will see a much bigger performance gain with a  tankless system.  But a reasonably designed combined heating + indirect-fired HW system will still be the better option for most, since it increases the duty-cycle (and thereby the net efficiency) of the heating system's burner, with much lower standby losses than a standard tank.

Finally a rational thought. Thank you Dana. I think maybe you are more than a casual blogger.

Most tankless water heater have a thermal efficiency vary close to the average water heater or boiler (low). If you want high efficiency the gas appliance must condense e.g. Navien and all the Mod-Con boilers.

Well done Dana!

"very close" and "low" efficiency a matter of perspective, eh? 

The raw combustion efficiency of a tank heater is limited to about 78%, and the tankless's raw combustion efficiency is about 85%.  But the odds of getting the tank to actually run better than 70% as it's true installed-efficiency  is very low due to the higher standby losses (they score in the mid- 60s in an EF test), whereas the odds of getting a tankless to run at 80-82% on average in real life is pretty good ('ceptin' for chronic obsessive-compulsive hand-washers, that is.) 

That's about 15-17% better fuel economy for the tankless vs. a tank.

I'd call that a real difference, a pretty good step-function, especially since:

A  condensing Navien's steady-state efficiency is about 98%, but probably only gets a true 95% when used solely for hot water heating (those hand-washin' short-cycles just KILL, even on condensing units, even for non-compulsive washers . :-(  )

That's also about a 16% improvement in fuel economy over a pretty-good non-condensing tankless- another pretty-good step, a real difference.

The condensing tankless gets a whopping 36% better fuel economy than a better-than average tank heater too(!).  But at what cost?  The base component cost for condensing vs. standard tankless HW heaters is a ~$1K adder to slip into your net-present-value analysis.  It may be worth it in some markets, but I dunno... depends on what NG prices do over the next decade or two.  Sometimes the middle road really IS the right path, sometimes not, but the middle road here really is somewhere in the middle, not lumped in at the lower end of the scale.

In the Consumer Reports article they deftly dance around the longevity & service issues too- they didn't add in a full -replacement of the tank heater, simulate only an 11 year use (by excessively hardening the water- which is also a dubious method) yet say it "can take up to 22 years to break even—longer than the 20-year life of some model".   In my accounting book I'd have to double the purchace & installation costs of the tank, yet apply only a 1.25 multiplier to the tankless.  Tanks RARELY give 20+ years of best-efficiency service before out & out replacement replacement is required whereas tankless burners are easily servicable (even heat-exchanger replacements), and the bulk of the "extra" installation costs are presumed electrictal outlets for the powered venting & controls and presumed gas line capacity upgrade.  A full replacement of the tankless after 20-25 years of service won't be NEARLY the numbers quoted, since electrical power, gas lines & stainless flues etc. would already be in place- it'll be about the same as the tank. In many instances the gas lines might already be big enough to deliver, needing no upgrade.

It's true that while it's easier to service, the tankless will likely NEED to be serviced at least a couple of times in 20 years (I've got one that's 15 years old and going strong without any service beyond brushing out the heat exchanger once, but it's less complex than the Noritz & Takagis they tested).  But then again, is swapping anodes every years and annually draining sludge from your tank really less work than swapping filters & occaisionally checking for lime scale on the tankless, cleaning as-necessary?  It's a crap shoot, sez me, but in most situations the efficiency of the tankless will be better.  Whether it ends up being totally dollars & cents cost effective is something you'd need to analyze, but I've yet to hear from anybody who went back to a tank after living with a tankless.

If you're considering it ever, PLEASE do a better job of the analysis than the Consumer Reports folks, eh?  They adequately explain neither their test procedures nor their financial analysis, projected fuel prices, etc..  I found the article pathetically devoid of essential detail, and high on "we know better, and thus recommend..." attitude.  (Can't say I'd recommend buying their mag, eh? But I s'pose I'm just coppin' and attitude myself. ;-) )

And to repeat:  If you already have a decent hydronic boiler as your primary heating source, adding an indirect fired HW heater as a priority zone will almost ALWAYS be your best most cost-effective bet over either a self-standing tank or tankless, since it'll have all of the efficiency & longevity of the latter (sans-quirks like the cold-water-sandwich effect) at a lower installed cost, while somewhat improving the overall efficiency of the combined system.

There's no discussion of electric tankless water heaters in the article. How does their efficiency compare?

Probably kickin' a dead horse here, but I tend to believe people who actually measure stuff.

This is much more succinct & graphical comparative test summary from PG & E than the WHPAGette_Final.pdf document, but demonstrates clearly just how much of a difference actual use patterns effect operating efficiency (and how truly sucky standard-efficiency tanks perform relative to their EF numbers at low & moderate levels of more realistic use than an EF test):

http://www.aceee.org/conf/08whforum/presentations/1a_davis.pdf

Even one of the CONDENSING tank systems (heater #4) couldn't muster 70% under moderate use(!), and in the mid 60s under low use. The only tankless that was tested was a condensing version (heater #6- they didn't specify whose that was), and it too suffered slightly under low use, but not nearly as severely as standard efficiency tanks or even the crummier condensing tank. The only one that beat all of it's standardized tests under all conditions was a (definitely not cheap!) stratified condensing tank (heater #5).

I kinda wish they'd tested at least on of the standard-efficiency tankless units in this comparison, but it is what it is. The low mass of a tankless keeps it from falling off a cliff the way tanks do under low use conditions, but the low-use profile they used has a larger fraction of the total test volume in draws under 2 gallons (where tankless performance gets killed)- I suspect a standard efficiency tankless would have a similar degradation to the condensing tankless- subtracting 5-8% from the condensing version's performance is probably reasonable, meaning that in heavy use it'll be in the low 80s, moderate use around 75-78%, and low use around 70%- still better than the low-tech tanks under their best-case heaviest use.

The tank heaters DO fall off an efficiency cliff! None of the non-condensing standard-efficiency tanks (heaters 1-3) tested actually met their EF numbers in moderate-use profile, although they beat their numbers in heavy use profiles. Like I wrote in my original post, if you're an on-the-go 1-2 person family that often takes weekends away (or only bathes once/week) the tankless is a far superior option. But the stay at home family doing 2 loads of laundry every night after the kids go to bed will see an efficiency gain with a tankless, just nothing like multiplier the on-the-go single/couple experiences.

From a consumed fuel at the powerplant to hot water entering your pipes point of view, it's in the low 30% range. 

From a where the power lines hook up to your meter to hot water entering your pipes point of view it's in the high-90s- slightly ahead of EnergyStar versions of electric tank heaters (mid-90s).

From a general greenliness point of view, only Japaneses heat-pump type hot water heaters come close to beating the highest-efficiency tankless or condensing tanks ( or indirect fired tank on a high-efficiency boiler.)

This maybe a little off the topic.  We really need help to find a tankless water heater and you all seems very knowlegable (I have to admit that I don't completely follow the technical comments).

Our weekend cabin is under construction.  We need to decide on the water heater now so we can start electrical work.  It has 2 bathrooms (two showers and one bath), no dishwasher, no washer/dryer (maybe in the future), two panels (320 amp total) and no gas service (and no space for propane tank, either).  It will be mostly used on weekends, for a family of three (maybe occasional guests).  We have been looking for an electrical tankles water heater for two years and heard enough conflicting stories.  There are the models on our short list now:

Bosch powerStar AE-125
Stiebel Eltron Tempra 29

Any other mfr/model for our situation?  Most people we talked to don't like electrical units because of the power comsumption.  We believe that we have enough amp to support it.  Our concerns are which model works for us (both max output and min flow required to triger the heater), how reliable they are, and the service in case we need.  We will install it ourselves (and do all of electrical/plumbing work except panel).  We've heard that it might take longer for hot water to start and a re-circ device might be needed.  All of the plumbing fixtures are within 15 feet of the heater and we wonder if it helps. 

We would like to hear from you, anything would help. 

Thanks!

I spec. the Marathon super-insulated tanks and use off-peak electricity where it's available. Marathon water heaters require virtually no maintenance and DIY friendly. Efficiency is excellent considering coal-fired power plants power most of my installations.

Tankless are more viable in new construction applications.

The added and considerable cost of conversion (almost always including moving, gas, vent and adding electrical) commonly costs three times as much as a straight change-out.

These additional labor and material costs can't be ignored. As for maintenance; the owners of conventional gas water heaters do not maintain there water heaters, period. They may replace them every ten years on average, but they don't call me until the shower runs cold.

In climate such as yours and mine here in Minnesota (9 months of heating season) the standby loss argument is weak.

I like tankless technology (installed a Bosch in 1983) but the condensing models make the cost of labor and material more logical, cost effective and GREEN.

Still, the Mod-Con boiler with indirect water heater rules.

For a weekend cabin, a decent sized well insulated electric tank is reasonable- just turn it off at the breaker whenever you leave. To go with an electric on-demand you need a much higher amperage service than it would otherwise need (you say you have enough, but when the breaker trips while your wife is in the shower, your daughter is drying her hair an you're reheating the coffee in the microwave while making toast ain't nobody gonna be happy.)

It'll be more efficient in real-use to have a small sub-5 gallon under-the-sink kitchen tank (eg. Bosch Ariston GL series- size it for your dishwasher needs- even the 2.5 gallon version will probably work) and another shower/bath/laundry sized tank (40-50 gallons) elsewhere, rather than a central on-demand electric with a circulation pump, even if it's only calling from 10' away. (If you DO go with a recirculation scheme, avoid those that run continuously- it'll out & out RUIN the overall efficiency of the water heating system.) Electric tanks have inherently lower standby losses compared to gas-burners since the insulation can be continous, and there are no flues. But the more hot water that's left in the piping between and the point of use there is, the less efficient it is as a system. Whatever you use for a water heater, insulating the full lengths of the plumbing runs makes a difference.

You don't say how you're heating the cabin, but if there's any way to combine the space heating & hot water systems it'll usually be more efficient overall.

Dana1,
Thanks for your reply. I'm confused: in previous post, you mentioned that for smaller family, less frequent use, tankless is ideal. I thought the cabin fits this description well. Instead of waiting for a full tank to be heated on a Friday night (after a long drive) and waste a tank of hot water during weekdays, all we need to do is to flip a switch and turn on the faucet. Other reasons we thought would work well here: space saving (for a small cabin) and potential leaking tank. We were gone last month and there was a snow storm. We are dealing with frozen pipes and manifolds.
We have two electrical panels installed, with 200 amp each. We plan to have 120 - 150 amp dedicated to water heater. The heating will be electrical and a wood buring stove. We looked into radiate floor and don't think it will work here (same reason as tank heater). The summer here is very cool and we don't need air-conditioning.
I asked the tankless sale person if smaller point-of-use heater works better. He thinks whole-house uses less amperage in this case. By the way, we'll only have one bath (1 shower) for the first 5-10 years. We are trying to get OC first then apply for phase 2 permit. It will be only the master bath (shower+tub). The tub will be the farthest away from the heater (15 feet) and the main shower is nearest (0 feet). We will take your advise and insulate the pipes.
Since we are doing most of the work, we have time to research products (another way of saying it's very slow). We are almost done with siding and electrical will be next.
Thanks again.

"smaller family, less frequent use, tankless is ideal. I thought the cabin fits this description well.  Instead of waiting for a full tank to be heated on a Friday night (after a long drive) and waste a tank of hot water during weekdays, all we need to do is to flip a switch and turn on the faucet. "

In that post I was talkin' 'bout gas/propane fired apples.  When you move into electrics you're well into citrus (clementines, if not out & out oranges.)  Standby losses on electric tanks are inherently lower than that of gas fired beasts due to the fact that they

A: never standing pilot burning 24/7

B: don't have flues that convect heat away 24/7 (even on forced-draft models there's usually some leakage

C: they can be fully insulated with only small plumbing & electrical penetrations (as opposed to the internal materials & clearance & drafting requirements of a gas-burner)

In practical terms the standby losses of an electric tank can be made almost arbitrarly small with added insulation, but never for gas-burning tanks.  With gas tank heaters, total use volume is everything- use more water, the efficiency goes up, limited only by combustion efficiency. With gas tankless heaters burn-cycle-length above  some minimum (varies a bit by model) is everything- as long as most draws are over the minimum the overall efficiency is maintained.

But with electrics, you've already been reduced to ~30% fuel efficiency (the rest of the energy went up the stack & cooling towers at the  powerplant and heated up wires & transformer cores in the grid).  The difference in efficiency between tankless/tank electric HW heater just isn't that much.

In all HW heating distribution losses need to be minimized, but with electric heaters distribution losses are more likely to exceed the efficiency differences between heater types.

The point about potentially throwing away a tankful of HW is a vaiid one, but one that can be mitigated with a bit of forethought.  Turning off the shower tank right after the last SHOWER, and the kichen tank right after the last dishwasher load may require more thinking, so if you don't want to think about it you'll do fine with a central on-demand with insulated pipes.   Using the woodstove so's you don't have to power up the electric resistance heating will be a bigger factor on your power use.  If you're already using the electric heating system, the distribution losses will just heat the house a bit (although you still may end up throwing away a lot o' BTUs as luke-warm water.)

"We were gone last month and there was a snow storm. We are dealing with frozen pipes and manifolds."

Freeze controls the use the electric heating is probably the right thing to do, unless it's pretty easy to drain the water system before you leave (not likely.)  Mind you, it takes a lot longer for a tank to freeze up than a tankless HW heater, but by then the rest of the water system is already locked up solid.  Hope you didn't lose too much. :-(

"By the way, we'll only have one bath (1 shower) for the first 5-10 years. We are trying to get OC first then apply for phase 2 permit. It will be only the master bath (shower+tub). The tub will be the farthest away from the heater (15 feet) and the main shower is nearest (0 feet). We will take your advise and insulate the pipes. "

Unless you're taking back-to-back baths, the distribution losses to the tub draws are pretty much guaranteed fixed on per-bath basis.  It's the runs to the kitchen and other sinks that have smaller, more frequent intermittent draws where insulated pipes can make a difference in distribution losses.  But as long as the runs are inside of conditioned space (the insulation and pressure boundary of the building), those losses become just part of the space-heating.  Just try to minimize the "lukewarm water down the drain" losses, eh?

More than you ever wanted to know about tank vs tankless electric HW heating & distribution losses under high & low use profiles can be found here:

http://www.nrel.gov/docs/fy03osti/32922.pdf

Dana1,

Thanks for your explanation.  I never thought about gas/propane or electric would make so much difference but it all makes sense.

From the information we gathered, electrical tankless water heater seems work, but the efficiency is not much improved from electrical tanked heater.  Now we understand why. 

In our case, efficiency is not our major concern.  Our biggest concern is that if it really works as it says, not just burn a lot BTU's for some luke-warm water. 

The cabin will be used on mostly weekend and holidays.  We plan to shut off the water and switch off some circuits every time before we leave (just for peace of mind).  Occasionally, we may let friends stay there so we would like to keep the operation as simple as possible.  We also would like to have hot water anytime we walk in, no matter it's late Friday night or early Saturday morning.  So when we first learned about tankless water heater, everything seems fit our need. We designed this house with tankless heater in mind and asked our electrician to install two full panels (only costed a little more than one panel).  We also looked into propane tank but it doesn't fit on the site.

Most of the plumbing are on interior walls except the kitchen sink and future washer.  We should insulate the pipe to the sink.  We have community water.  There is a filter at the incoming water.  It was frozen during the last storm.  We learned that we have to shut off at water meter and drain the pipe.  This winter is unusually cold.

Thanks for the link.  I bookmarked it and will spend some time to read it.

Thanks again - very helpful information.

Hi all. I hope bumping an old thread is ok. Thanks Dana1 for the link. This was the kind of information I've been looking for. Some of the stuff i have been wondering: What is the biggest water draw in an average home? 3GPM for laundry seems to be the answer in that link...still correct? How much of a draw does a normal shower take? 2GPM seems to be the number they've used. Here in Calgary, AB I have built the foundation for my house addition (16' extension/36' width(south facing) + 8'x12 stairwell area+/-) out of 8"concrete core 13 1/4" thick Advantage ICF. There are 3 Gienow windows across the 36' dimension (south facing and 5'x4' with Sol-R glass) The foundation height is 9'7, all out of ICF. with two stories to be stick framed on top of the foundation. The walk-out basement will be a secondary suite to be rented to a single person. Am I right in thinking that an electric demand type tankless water heater could be a good choice? My reasoning being: Single person-should only need one water draw at any one time. No pilot light or venting required, keeping the tight seal of the ICF in tact. Electricity portion of the utilities to be split- The tenant pays their own power bill. Thanks for any replies and corrections.

Bathtub fills can easily be 4, 5, 6, even 8gpm, depending on the water pressure and the size of the distribution.  (IIRC, my old-skool antique ceramic shower head measured in at ~4gpm, before I throttled it back from an amazing gusher to a more reasonable warm-rain.) 

With a standard 2.5gpm shower head, 120F hot water and 50F cold water you get about 2gpm of hot for a shower. With colder incoming water it takes more gpm on the hot, with warmer incoming water, less.

Electric on-demand hot water heaters take a very large amount of current.  In Calgary you may see incoming water temps as low as 40F, and to get a minimum 2gpm of flow @120F output, which takes ~23,400 watts of power, which is ~100amps @ 240V.  Is it worth upgrading the electrical service & panel by an extra 100A just for hot water?  For a single person, a 40 gallon tank would be more than enough unless they love endless hot showers or are filling a 50 gallon soaking tub, etc.  Electric hot water tanks have very low standby losses compared to their gas-fired tank cousins- it may take decades to make up the difference in installation costs on the very marginal difference in operating costs (which are paid by the tenant anyway.) 

Insulating all the plumbing from the tank ,including the cold side plumbing, and the pressure & temperature safety valve downspouts, etc. to a minimum of R4 (5/8" wall closed cell pipe insulation) can bring an electric tank up to very near the efficiency performance of an on-demand, since the standby losses of an electric tank are well under that of the near-tank plumbing.  If you insulate ALL of the rest of hot water distribution plumbing to at least R2 (the 3/8" wall pipe insulation) you will have cut the distribution losses to a fraction of the typical ~15% down to 5% or less, for a lot less cash outlay than an upgraded electrical service/panel + tankless hot water heater.  If there's space for the tank, that's the route I'd recommend from a cost & efficiency point of view.

For showering you can extend the apparent capacity of any tank heater with a drainwater heat exchanger, but that's a difficult retrofit in many applications, and the available options for first-floor or basement slab showers are limited.  With an on-demand it's like having an extra 8-12kilowatts of heater, and can give some margin to otherwise on-the edge situations. The cost is comparable to an upgraded power service + tankless vs. a 40 gallon tank + pipe insulation, but the return on investment is higher.  See:  http://www.renewability.com/uploads/documents/en/analysis_dwhr_minnesota.pdf

It's not necessarily worth doing it for your single-tenant, but it could be for the main unit if you're heating water with electricity, or even propane.  There's an online calculator which gives you some idea as to the payback for a few different models, based on average use patterns and average water temps for many Canadian cities online here:

http://www.ceati.com/calculator/

I have a Tempra 29 installed in my house for over a year now. Panel wise you should be fine as I'm in a single 200 amp entrnace and have no issues. If all fixtures are as close as you say I would skip the recirc. The main issue will be minimum activation flow, which my only beef with my tankless. The way I have dealt with it is to lower the setpoint to where when I want to use hot water the hot water taps will be fully open. My current setting is 42C (~115F).

Thanks Dana! I have a lot to digest now...
PS sorry for the no-paragraph approach. I don't know hoe to format my posts properly.

Thanks AngleOfDebt for sharing your experience.

We ended up with a T36.  Installed in the summer and just fired it up for the first time before Thanksgiving and it seems work.  The water is not super hot but good enough for shower, etc.  I noticed that every time we have to turn hot on full for about 10 seconds before warm water comes.  I guess we'll get use to it.

We are working on the bathroom this winter.  When we have the shower, we'll find out if we really like the tankless heater.