A tankless unit stores nothing, so it cannot be sized by the gallon. It is sized by how fast it can add heat, and that rate has two inputs, not one. So the question “what size tankless water heater do I need” is really two questions wearing the same coat: how many gallons a minute will the house draw at the same moment, and how many degrees does the unit have to lift them. Almost every sizing guide answers the first properly and waves at the second. The second is the one that moves the answer.
The arithmetic behind it is small. Raising one gallon per minute by one degree Fahrenheit takes about 500 BTU an hour, so the heat you need is flow times rise times 500. What turns one appliance into two is the rise. Incoming water runs near 75 °F in Fort Myers and near 43 °F in Duluth, which means the same box on the same wall hands one house just over eight gallons a minute and the other under five. Nothing about the unit changed. If you have not settled the tank question yet, the storage sizing worksheet covers that side first, because tanks are sized on a different number entirely.
The short answer
- One bathroom, Gulf coast or southern California: the smallest whole-house gas unit on the shelf, around 120,000 BTU/hr, has margin left over. Do not buy up.
- Two bathrooms, warm climate, showers overlapping: 150,000 BTU/hr covers two showers and a kitchen tap at a 45 °F rise with a little room spare.
- Two bathrooms, Ohio valley or mid-Atlantic: you are shopping at 180,000 to 199,000 BTU/hr, which is the top of the residential shelf, not the middle of it.
- Two showers at once north of about the 42nd parallel: the largest residential unit, 199,000 BTU/hr, covers exactly that in January and nothing more. Add a kitchen tap to it and you are past what one unit can do.
- One remote fixture, a garage sink or a guest bath at the far end of the house: stop sizing for the house. Size a small point-of-use unit for that fixture and leave the main heater alone.
Rise is the half that moves
Flow is the easy half, because it is a property of your house and a bucket will measure it. Add up the fixtures that genuinely run together at the worst moment of the day. A shower head sold since the federal cap took effect delivers 1.5 to 2.5 gallons a minute and most sit at 2.0. A bathroom faucet is 0.5 to 1.5, a kitchen faucet 1.0 to 2.2. A tub filler is the outlier at 3.0 to 4.0, which is why a soaking tub quietly doubles the unit you need. Dishwashers and clothes washers mostly heat their own water or run cold now, so leave them out unless yours is old.
Rise is the half you cannot measure with a bucket, and it is set by geology and season rather than by anything you own. Take the temperature you want at the tap, usually 120 °F, and subtract the temperature of the water arriving at the house. In south Texas that subtraction gives you a rise in the mid-forties. In Minnesota it gives you a rise near eighty. The unit’s burner output is fixed, so the delivered flow has to give way. That is the entire mechanism, and it is why a size chart with no climate column is worse than no chart at all.
Side by side
| Running at oncesimultaneous GPM | WARM45 °F rise | MODERATE60 °F rise | COLD75 °F rise |
|---|---|---|---|
| One shower2.0 GPM | 47,000under the smallest unit soldeasy | 63,000still entry level | 79,000entry level covers it |
| Shower plus bathroom sink3.0 GPM | 71,000one bathroom house | 95,000small unit, comfortably | 118,000mid range unit |
| Shower plus kitchen sink3.5 GPM | 83,000the common real peak | 111,000mid range unit | 138,000upper mid range |
| Two showers4.0 GPM | 95,000two bath house, warm | 126,000150,000 class unit | 158,000180,000 class unit |
| Two showers plus kitchen sink5.5 GPM | 130,000150,000 class unit | 174,000top of the shelftight | 217,000past 199,000 BTU/hrtwo units |
| Tub filling plus a shower6.0 GPM | 142,000150,000 class unit | 189,000largest residential unittight | 237,000past 199,000 BTU/hrtwo units |
| Three showers plus a sink7.0 GPM | 166,000180,000 class unit | 221,000past 199,000 BTU/hrtwo units | 276,000well past one unittwo units |
Under 150,000 BTU/hr? Nearly any whole-house condensing unit will do it, so shop on features and warranty instead. Over 199,000? No single residential gas unit reaches it at any price, and the honest answer is staggered showers or a second unit.
Those figures are input ratings, not output. They assume a condensing unit at a UEF near 0.95, because that is what the efficiency does to the sum: 5.5 gallons a minute at a 60 °F rise needs 165,000 BTU/hr of actual heat, and 165,000 divided by 0.95 is 174,000 BTU/hr going in at the gas valve. Run the same row on a non-condensing unit at 0.82 and it wants 201,000, which pushes it off the residential shelf entirely. Efficiency is not only a running-cost question here. It is a capacity question.
What your groundwater is doing
Nobody can tell you your incoming water temperature from a map, and I want to be blunt about that before printing one. A shallow well, a long service line under a hot driveway, a municipal system drawing from a reservoir instead of an aquifer, and a February cold snap all move it. The bands below are the working ranges I use, and the low end of each is the winter figure, which is the one you should size on. Sizing on the annual average is how people end up with a unit that is perfect in September and disappointing in February.
| Regionwinter to summer | INCOMING°F at the meter | RISE TO 120 °Fwhat the unit must lift | 199,000 BTU/hr UNITdelivers, in GPM |
|---|---|---|---|
| South Florida, south Texas, Gulf coast | 72–77barely seasonal | 43–48the easiest in the country | 7.9–8.8three showers and a kitchen tapmost headroom |
| Deep South, southern California, low desert | 64–72mild swing | 48–56still forgiving | 6.8–7.9three showers at once |
| Mid-Atlantic, Ohio valley, mid-plains | 54–64real seasonal swing | 56–66the national middle | 5.7–6.8two showers plus a sink |
| Northeast, Great Lakes, Pacific Northwest | 46–54cold half the year | 66–74where sizing gets hard | 5.1–5.7two showers, nothing elsemargin gone |
| Upper Midwest, northern Rockies, interior Alaska | 40–46coldest supply in the US | 74–80the hardest case | 4.7–5.1two showers is the ceilingsize up or split |
Same appliance, both ends of the country. In Fort Myers the largest residential unit runs three showers together. In Duluth it runs two, and only if nobody touches a tap while they are running.
What size tankless water heater do I need in Ohio?
Four people outside Columbus, Ohio. Two full bathrooms, both used between seven and seven thirty on a weekday. Shower heads measured with a bucket and a stopwatch at 2.0 and 1.8 gallons a minute. Somebody is always at the kitchen sink at 1.5. Worst honest case, all three together: 5.3 gallons a minute.
Incoming water in that part of Ohio sits around 52 °F in October, so the rise to 120 °F is 68 degrees. The heat required is 5.3 times 68 times 500, which is 180,200 BTU/hr. Now run it the other direction, from the biggest unit available. A 199,000 BTU/hr condensing unit at UEF 0.95 delivers about 189,000 BTU/hr of heat, and 189,000 divided by 34,000 gives 5.6 gallons a minute. It fits. Barely.
Then February arrives and the incoming water drops to 48 °F. The rise becomes 72. The same unit now delivers 189,000 divided by 36,000, which is 5.25 gallons a minute, and the household peak is 5.3. Nobody freezes, but the second shower goes tepid for the thirty seconds the kitchen tap is open. That family did not buy the wrong unit. They bought the largest one made for a house and still ran out of margin, which is the real finding: in the northern half of the country, two simultaneous showers plus anything else is the point where a single residential tankless stops being a whole-house appliance.
Buying up costs less than you think
Here is the part that separates tankless sizing from tank sizing, and I think most guides get the emphasis backwards. An oversized storage tank punishes you every day of its life, because standby loss scales with surface area and a tank you do not need is still a tank you are keeping hot. A tankless unit has essentially no standby loss, and it fires only as hard as the draw asks. A 199,000 BTU/hr unit running a single hand wash burns the gas a 120,000 BTU/hr unit would.
So the annual cost of buying up a size class is close to zero, and the arithmetic says so. At this site’s household assumptions the delivered heat load is about 4,000 kWh a year, which is 136 therms of heat at 29.3 kWh to the therm. Divide by a UEF of 0.95 and the input is 143 therms, or $215 a year at $1.50 a therm. That figure is set by how much hot water the family uses, not by how big the burner is. Over ten years it is roughly $2,150 either way. The entire cost of buying one size larger is the sticker gap, a few hundred dollars inside the $1,000 to $2,000 band where whole-house gas units live. Set that against being one shower short every winter morning for fifteen years. The trade is not close.
The number on the carton is not a size
Walk the aisle and you will see units badged 9.8 GPM and 11 GPM. Those are real measurements taken at a temperature rise almost nobody in this country experiences. Work backwards from 11 gallons a minute on a 199,000 BTU/hr unit: 189,000 BTU/hr of heat, divided by 11 gallons, divided by 500, gives a rise of about 34 degrees. That means incoming water at 86 °F. No municipal supply in the United States runs at 86 °F in January. The number is not a lie. It is a laboratory condition, chosen because it flatters the unit. I would not buy a tankless on the figure printed on the carton. Every serious manufacturer publishes a flow-versus-rise curve in the spec sheet, and that curve is the real size of the appliance.
Two more things this method cannot settle for you. It cannot tell you whether your gas meter and line can feed a 199,000 BTU/hr appliance, which is a real constraint on older houses and a question for a licensed plumber, not a spreadsheet. And it cannot make an electric whole-house unit work in a cold climate: the same 5.3 gallons a minute at a 68 °F rise needs about 53 kilowatts, since kilowatts equal GPM times rise times 0.1465, and that is a service most existing homes do not have. Electric against gas tankless is really that one question in disguise. Ask an electrician before you fall in love with a model.
If you want the read-off version without the reasoning, the tankless size chart lays fixture flows and required capacity out as a grid. If you are still weighing whether the appliance suits the house at all, the pros and cons checked against a real house covers the cold-water sandwich and the minimum activation flow, which are sizing problems dressed as comfort problems. Once you have a BTU number the choice narrows to models, and that is where Rinnai against Rheem becomes the useful comparison.
Sources
- DOE Energy Saver, sizing guidance for demand-type water heaters, source of the flow-rate and temperature-rise method used here.
- AHRI Directory of Certified Product Performance, for rated input in BTU/hr and UEF on residential gas instantaneous units.
- ENERGY STAR certified product finder, tankless gas water heaters, for the UEF bands behind the 0.95 and 0.82 figures.
- Manufacturer specification sheets, specifically the published flow-versus-rise curves sitting behind the headline GPM on the carton.
- FTC EnergyGuide label, for the annual energy cost basis on the units it covers.
- US EIA residential energy prices, for the check on the $1.50 per therm and $0.17 per kWh constants.
The sum is two lines long and it is the only thing that matters here. Measure your fixtures, decide honestly which of them run together, find your winter incoming temperature, and multiply. Everything else on the box is marketing. If you want to see how these models are built, which constants they use and where they stop being reliable, the method page lays it out.
1 comment