Whole House Reverse Osmosis: When It Is Worth It, and When It Is a Mistake

A whole house reverse osmosis system is the most expensive thing you can plumb into a residential water line. For most of the homes that ask me about one, it is also the wrong purchase. The technology is fine. Point-of-entry RO solves a problem most houses do not have, and it hands you three new ones on the way in: a storage tank, a repressurization pump, and a drain line that runs every time somebody opens a faucet. If the real complaint is chlorine taste, spotted glassware, or nitrate in the drinking glass, cheaper machines exist for all three.

What a whole house reverse osmosis system actually is

Strip the brochure language off and point-of-entry RO is a small water plant bolted to a garage wall. Feed water goes through pre-treatment. A booster pump pushes it against a semi-permeable membrane at 60 to 200 psi. Most of the dissolved salts, somewhere between 90 and 99 percent depending on the element and the feed, never cross the film and leave down a drain line. What does cross is called permeate. It arrives slowly, and at almost no pressure.

So it has to fall into an open tank and then be pumped back up to household pressure by a second pump. That is really the whole article. Everything expensive about a POE system follows from one fact: a membrane trickles and a house gulps. The pressure, recovery and rejection mechanics behind any membrane are identical at a kitchen sink and at the main line. The consequences of a slow production rate are not, once the demand is 300 gallons a day.

One thing buyers rarely hear from a dealer. NSF/ANSI 58 is written for point-of-use RO systems. A whole-house assembly cannot carry that certification as a system, so contaminant reduction claims for POE RO usually rest on the membrane manufacturer’s own performance data rather than a third-party certified system claim. Individual components in the train may hold NSF/ANSI 42 for taste and odor, or 53 for health effects. Ask which certificate covers which box, and ask to see it.

Why you cannot skip the storage tank

A membrane rated 1,000 gallons per day makes about 0.7 gallons per minute. One shower head pulls 1.8 to 2.5 gpm. A washer filling while a toilet refills and somebody rinses dishes will ask for 8 to 12 gpm. No membrane a homeowner would pay for serves that. Permeate therefore drips into an atmospheric tank at zero pressure, and a separate pump draws from that tank and rebuilds 50 to 60 psi for the house.

What homeowners underestimate is the tank itself.

• A 300-gallon polyethylene tank stands roughly four feet tall and three feet across, and it wants floor space near a drain and an outlet.
• It is vented to atmosphere, so it breathes room air. That means a screened vent, an overflow routed somewhere safe, and sanitizing on a schedule.
• The repressurization pump becomes a single point of failure for the whole house, not just the drinking tap. Constant-pressure units run $900 to $2,000 installed, depending on flow rating, and they do eventually die.
• On city water you now lose household pressure in a power outage. You did not before.
• The pump adds electricity, typically $50 to $180 a year depending on household draw and local rates, and it is audible in a finished basement.

Size the tank against peak draw, not against the daily total. Four people showering back to back on a school morning is the number that matters: the tank carries that block alone while the membrane trickles at 0.7 gpm.

The pre-treatment that keeps the membrane alive

A membrane is a consumable with a short temper. Feed it raw well water and you will be buying elements every eighteen months. A working train usually needs a 20-inch sediment stage, 5 micron followed by 1 micron. Then carbon, if there is any chlorine or chloramine in the feed, because thin-film composite membranes are destroyed by free chlorine and the tolerance is counted in ppm-hours rather than years. Iron and manganese have to be held under roughly 0.05 to 0.1 mg/L. Hardness has to be softened out or dosed with antiscalant, or calcium carbonate cements the tail element shut.

Read that list again and notice the irony. Plenty of people look at POE RO precisely because they are tired of hauling salt bags, then discover that a softener sized for their iron and hardness load is the cheapest insurance available on the membrane they just bought. On hard well water you generally end up owning both machines.

Before any of it, get a certified lab panel, not a pool strip. TDS, hardness, iron, manganese, pH, alkalinity, chloride, sulfate, nitrate, plus whatever your county flags locally. A dealer’s free in-home test measures hardness and TDS and sells softeners. It cannot tell you whether a membrane is warranted.

How much water goes down the drain

This is where sources disagree most, and the disagreement has a clean mechanical explanation. Manufacturers publish recovery at standard test conditions: 77 degrees Fahrenheit feed, a set test pressure, low-TDS test water. Your well sits at 50 to 55 degrees. Membrane output falls steeply as water cools, close to half between 77 and 52 degrees, and higher feed TDS raises osmotic pressure and pushes recovery down again. A system advertised at 75 percent recovery routinely lands at 40 to 60 percent in a real basement. Neither number is a lie. They are measured on different water.

Some installers raise the figure by recirculating concentrate back into the feed. That genuinely improves recovery. It also raises the salt concentration the membrane sees on every pass, which shortens element life. Both halves of that argument are true, which is why the number worth asking for is recovery at your water temperature and your TDS, in writing.

family of four, well water
indoor use  4 x 60 gal =  240 gal/day
recovery 50%, feed    =  480 gal/day
to drain  480 - 240   =  240 gal/day
membrane 1,000 GPD at 77 F
same membrane at 52 F ~  500 gal/day
peak draw 8 gpm x 10 min = 80 gal
tank 120 gal, pump 10 gpm at 50 psi

Two hundred and forty extra gallons a day is not an abstraction. On a well it is pump run hours and pump wear. On city water it is metered twice in most towns, once as consumption and again through a sewer charge calculated from that same meter reading. On a septic system it is hydraulic load, and a drain field sized for a four-person house does not keep a spare 240 gallons a day in reserve. Concentrate can often be routed to daylight or a dry well instead, but that depends entirely on local code. I am not a licensed plumber, and this is exactly the question to put to your county before you sign anything.

What low mineral water does to copper pipe

Search this topic and you will find dealers calling aggressive RO water a myth while corrosion engineers treat it as routine plumbing chemistry. The mechanism reconciles them. A membrane removes calcium, magnesium and bicarbonate alkalinity. Dissolved carbon dioxide is a gas, so it passes straight through and reforms carbonic acid on the permeate side. What comes out the other end typically sits around pH 5.5 to 6.5, with almost no buffering capacity and nothing available to lay down a protective carbonate film inside the pipe.

Water does not hunger for minerals. That part is folklore. The real problem is duller and worse: unbuffered, slightly acidic water standing overnight in copper tube, brass fittings and old lead solder will pick up metal, and nothing in the water holds the pH steady while it happens. EPA’s action level at the tap is 1.3 mg/L for copper and 15 parts per billion for lead, and a house plumbed before the 1986 solder ban is the one to worry about. If a POE RO goes in, permeate should be corrected before it reaches the plumbing, through a calcite contactor or a measured raw-water blend. Our guide to what remineralization changes and what it costs covers the targets. Budget $300 to $900 for the vessel, driven mostly by tank size, and $30 to $80 a year in media.

What the installed price actually covers

There is no national price for this, and anyone quoting one is guessing. Regional labor rates, well versus city feed, whether pre-treatment already exists, and how far the drain and the electrical panel sit from the equipment all move the number by thousands. What follows is a line-item band, so you can read a quote instead of reacting to a total.

Then the running cost, which is where POE RO quietly separates itself from a softener. Membrane elements last two to five years and run $200 to $900 each, and a two-vessel skid holds two of them. Prefilters three or four times a year, $60 to $200 annually. Calcite media, pump power, softener salt. Call it $250 to $700 a year in consumables on a typical residential build, more if the feed water is rough. Element life is the variable that swings hardest. Good pre-treatment and a conservative recovery setting can double it, and neglect can halve it.

The three homes that genuinely need it

Point-of-entry RO earns its keep when the dissolved load damages the house, not just the taste of one glass of water. Three profiles qualify.

Very high total dissolved solids

EPA’s secondary standard for TDS is 500 mg/L, an aesthetic guideline rather than an enforceable health limit. Above roughly 1,200 to 1,500 mg/L, everything the water touches suffers. Water heater elements crust over. Dishwashers streak no matter what you load into them. Fixtures spot within a day of being cleaned, and the taste turns salty or metallic at every faucet in the house. A softener will not fix this. Ion exchange trades calcium and magnesium for sodium and leaves total dissolved solids roughly where they started. People who soften high-TDS water usually report that nothing changed except the lather.

Brackish and salt-intruded wells

Coastal wells with seawater intrusion, and inland aquifers carrying high chloride and sodium, present a problem no cartridge solves. The secondary standard for chloride is 250 mg/L. Past that you get corrosion at fixtures and a taste nobody in the house will drink. Carbon does not remove chloride. Softener resin does not remove chloride. A membrane does, and in the residential range there is nothing cheaper that does the same job. This is the one profile where I would not argue with a POE RO quote on principle, only on its line items.

A contaminant stack, not a single contaminant

One membrane can address arsenic, uranium, nitrate, sulfate and hardness in a single pass. If a lab report comes back with four or five parameters over their limits, the alternative is four or five dedicated media vessels, each with its own media cost, backwash schedule, drain connection and way of failing quietly. Stack those up and the RO quote stops looking extravagant. One contaminant does not qualify. Nitrate over the 10 mg/L MCL is a drinking exposure rather than a bathing one, and a $400 under-sink unit handles the water you actually swallow.

Bacteria are a separate matter. A membrane is not a disinfection barrier you should stake your family’s health on, because a pinhole in the film or contamination downstream of it leaves you nothing to detect. Coliform in a well means a certified lab test first, then UV certified to NSF/ANSI 55 or shock chlorination, and a call to your county health department, which usually runs a testing program and keeps a list of certified local labs.

What most readers should buy instead

Split the job. Treat the whole house for what damages the house, and treat one faucet for what you drink. A sediment filter sized so it does not choke flow, a carbon tank for chlorine and taste, a softener if hardness runs above about 7 grains per gallon, an iron filter if there is iron, UV if the well has ever tested positive for coliform. Our comparison of how to judge whole house filtration systems properly walks through sizing that stack against peak flow rate rather than against marketing capacity claims.

Then put an under-sink RO unit certified to NSF/ANSI 58 at the kitchen sink for $200 to $600, plus $60 to $150 a year in filters, depending on stage count and household use. It wastes a few gallons a day instead of a few hundred. It needs no 300-gallon tank, no booster pump, no pH correction bed. It does not put household pressure at the mercy of a pump motor on a Sunday night.

The whole stack usually lands between $1,500 and $3,500 installed, which is less than the pre-treatment section alone on most POE RO quotes. Spend the difference on a proper lab panel and a plumber who will sit down and read it with you.