Do you need one at all

Minnesota Department of Health puts the line at greater than 7 grains per gallon or 120 mg/L.

That is a comfort, scale and appliance-life threshold. It is not a health threshold — hard water is not a health hazard, and anyone selling softening on health grounds is selling something else.

For context, USGS classifies hardness as soft to 60 mg/L, moderately hard to 120, hard to 180, and very hard above that. The softening threshold sits exactly at the boundary between moderately hard and hard.

Checking that threshold yourself

The two units get used interchangeably and rarely explained. Here is where the conversion comes from, using NIST’s published factors:

1 grain      = 64.79891 mg     (NIST SP 811, grain to milligram)
1 US gallon  = 3.785412 L      (NIST SP 811, gallon to litre)

1 grain per gallon = 64.79891 mg / 3.785412 L = 17.118 mg/L

So 120 mg/L divided by 17.118 is 7.01 grains per gallon — which reproduces Minnesota Health’s 7 gpg figure exactly. The two numbers in that guideline are the same number in different units, and now you can prove it rather than assume it.

What softening actually does

It is an exchange, not a removal. Minnesota Health: “Resin beads inside the softener trap the calcium and magnesium and exchange them for sodium or potassium.”

Something comes out and something goes in. That has two consequences most product pages skip:

  1. Softened water carries added sodium. If anyone in the household is on a sodium-restricted diet, Minnesota Health’s suggestion is to “regenerate your softener with potassium chloride instead of sodium chloride (salt).”
  2. Minnesota Health recommends you keep an un-softened tap for cooking and drinking. That advice appears in a state health department FAQ and almost never in a softener listing, for obvious reasons.

Sizing

Capacity is measured in grains of hardness removed between regenerations.

capacity = hardness (gpg) x 75 gallons per person per day x people x days per cycle
Household Hardness Grains/day 7-day cycle Buy
2 people 10 gpg 1,500 10,500 16,000 grain
4 people 15 gpg 4,500 31,500 32,000 grain
4 people 25 gpg 7,500 52,500 64,000 grain
6 people 20 gpg 9,000 63,000 64,000 grain

Longer cycles need more capacity but use less salt per grain removed. Undersizing is the more expensive mistake: a unit regenerating daily wears out its resin and its valve far sooner than the rating suggests.

What it costs to run, which is not the sticker price

Two softeners with the same grain rating can cost very different amounts to own. The difference is the efficiency rating, and it is rarely on the box.

EPA WaterSense publishes the criterion: look for a unit with “a salt efficiency greater than 3,350 grains of total hardness exchange per pound of salt.” EPA adds that this is verifiable through the voluntary efficiency rating under NSF/ANSI 44.

That figure turns your sizing arithmetic into an annual bill.

grains per year = hardness (gpg) x gallons per day x 365
salt (lb/year)  = grains per year / salt efficiency (grains per lb)

Four people at 15 gpg, using 300 gallons a day:

4,500 grains/day x 365      = 1,642,500 grains per year
1,642,500 / 3,350 grains/lb =   490 lb of salt per year
1,642,500 / 2,500 grains/lb =   657 lb of salt per year

The same water, the same household, the same capacity. An inefficient unit costs an extra 167 pounds of salt every year — around four more bags, every year, for as long as you own it.

The water side of the bill

Regeneration uses water as well as salt. EPA’s guide looks for units consuming “5 gallons (18.9 liters) of water or less per 1,000 grains of hardness removed.” Its fact sheet from the same month says 4 gallons for the same criterion. We quote both because both are EPA.

At 1,642,500 grains a year, 5 gallons per 1,000 grains is about 8,200 gallons. At 4 gallons it is about 6,600 gallons.

For scale, EPA states a typical softener uses “20 to 70 gallons of water per regeneration cycle, although some larger systems can use up to 200 gallons.”

How to use this before you buy

  1. Work out your grains per year from the formula above.
  2. Ask the seller for the salt efficiency in grains per pound, and the water used per 1,000 grains. Both are measurable, and both are on an NSF/ANSI 44 efficiency rating.
  3. Divide, then multiply by what salt costs you locally.
  4. Compare that annual figure across units, not the purchase price.

A cheaper softener that burns four extra bags of salt a year stops being cheaper, and the crossover usually arrives well inside the life of the unit.

Before you size anything: two disqualifiers

Penn State Extension is specific about when a softener is the wrong purchase. Softeners are recommended only where pH is above 6.7, hardness is between 3 and 20 gpg, and dissolved iron is below 5 mg/L.

And the failure that ends softeners early: “Oxidized forms of iron and manganese will foul the softener resin.” If your iron is already oxidised — particles visible the moment you draw a glass — it must be removed upstream. The iron page has the test that tells you which form you have.

Where iron is going through the resin, add roughly 5 grains of capacity per gallon for every 1 ppm of dissolved iron. At 2 ppm that is the equivalent of another 10 gpg of hardness, which is usually the argument for removing it first instead.

Mistakes that cost the most money

Undersizing. A unit regenerating daily wears out its resin and valve far sooner than the rating suggests. It is the more expensive error by a distance.

Ignoring the efficiency rating. Two softeners with the same grain rating can differ by roughly 167 pounds of salt a year at the same hardness.

Sizing from a regional map. USGS states its widely-reproduced data are not designed for decisions at the scale of an individual property.

Softening every tap. Minnesota Health recommends keeping an un-softened tap for cooking and drinking, which softener listings rarely mention.

Buying before checking iron and pH. Dissolved iron above 5 mg/L or pH below 6.7 makes a softener the wrong first purchase.

Three situations, and what changes

Municipal supply, hardness only. The straightforward case. Size on the formula, check the efficiency rating, and buy the next standard capacity up.

Well water with iron present. Add roughly 5 grains of capacity per gallon for each 1 ppm of dissolved iron, or remove the iron upstream, which is usually the better trade.

Sodium-restricted household. Regenerate with potassium chloride, which Minnesota Health suggests directly, and keep an un-softened tap.

The objections worth answering

“Isn’t hard water a health problem?” No. It is a comfort, scale and appliance-life issue. Anyone selling softening on health grounds is selling something else.

“Why not buy the biggest unit available?” Oversizing costs more up front and does not improve the result. The sizing formula exists so you buy the next standard capacity above your requirement, and stop there.

“Does softened water taste salty?” Softening exchanges calcium and magnesium for sodium or potassium rather than adding salt as such. If sodium is a concern, potassium chloride and an un-softened tap address it.

Who this is for

Households above roughly 7 grains per gallon sizing a residential softener, on a private well or municipal supply, who want the running cost before the purchase.

Choose a softener if

  • Your hardness is above 7 grains per gallon, or 120 mg/L.
  • Your pH is above 6.7 and dissolved iron is below 5 mg/L.
  • You can discharge brine and there is a drain at the installation point.
  • Scale, soap use or appliance life is the complaint you want fixed.

Choose to wait if

  • Your hardness sits below 7 grains per gallon. Most households do not recover the running cost at that level.
  • Your iron or your pH falls outside the window above. Fix that stage first, then size this one on the water the softener will actually see.
  • Someone in the house is sodium-restricted and you have not yet decided between potassium chloride and keeping an un-softened tap.

Salt-free alternatives

A salt-free “softener” is a different product with a different objective, and there is an objective test for the distinction — NSF/ANSI 44 is scoped to salt-regenerated cation exchange, so a saltless device falls outside it structurally.

→ Salt-free vs salt-based, and the certification test → Why a high TDS reading is usually hardness in another unit → How much sodium softening actually adds, worked

Where it sits in the system

Fourth. After pH correction, after iron and sulfur removal, before UV. The full sequence and the reasoning is worked through in the sizing method.

→ Radium: the claim 78 hardness-certified softeners do not carry