Methodology

Results are compared to published EPA drinking water standards. The distinction between the two kinds matters and we keep it explicit throughout.

• Primary standards (MCLs) are health-based and enforceable for public systems. Nitrate at 10 mg/L as N, arsenic at 0.010 mg/L, coliform bacteria absent.
• Secondary standards (SMCLs) are aesthetic. Iron 0.3 mg/L, manganese 0.05 mg/L, TDS 500 mg/L, sulfate and chloride 250 mg/L, pH 6.5 to 8.5.
• Private wells are regulated by neither. These are the right benchmarks to hold your water to, but nobody is enforcing them on your behalf.

• Hardness conversion: 1 grain per gallon = 17.1 mg/L as calcium carbonate.
• Household use: 75 gallons per person per day by default, adjustable in the tool.
• Iron and manganese compensation: 4 gpg of additional load per mg/L. The commonly published range is 3 to 5; we use the midpoint.
• Target regeneration interval: 7 days.
• Usable capacity: 75% of rated capacity, because ratings are measured at maximum salt dose.
• Iron limit for a softener alone: 3 mg/L, above which we recommend a dedicated iron filter.

• Salt efficiency: 4,000 grains of hardness removed per pound of salt at an efficient dose. At maximum dose this falls to roughly 2,500.
• Regeneration water: about 50 gallons per cycle.
• Water cost defaults to zero, because well owners pump their own — only the electricity is theirs to pay.

• Drawdown factors: 30% at 20/40 psi, 25% at 30/50 psi, 21% at 40/60 psi.
• Minimum pump run time: 1 minute below 1 HP, 2 minutes at 1 HP and above.
• Required drawdown = pump GPM x minimum run time; tank volume = drawdown / drawdown factor, rounded up to a standard size.
• Air pre-charge: 2 psi below cut-in pressure.

This is deliberately more precise than the widely repeated “tank = GPM x runtime x 1.5” rule, which ignores the pressure switch setting entirely even though that is what determines how much of the tank you can actually use.

• Total dynamic head = vertical lift + pressure head + friction loss.
• Pressure head = cut-out psi x 2.31 feet per psi.
• Friction loss by the Hazen-Williams equation with C = 150, the standard coefficient for plastic (PE or PVC) pipe. Checked against published pipe tables: 10 GPM through 1 inch computes to 6.99 ft per 100 ft here against a published figure of about 7.0.
• Water horsepower = GPM x head / 3960.
• Shaft horsepower = water horsepower / pump-end efficiency, which rises with flow: 28% at 3 GPM, 38% at 5 GPM, 46% at 7 GPM, 52% at 10 GPM, 56% at 20 GPM, interpolated between those points.
• Motor size is the smallest standard motor whose rated horsepower times its service factor covers that shaft load: 0.5 HP x 1.6, 0.75 HP x 1.5, 1 HP x 1.4, 1.5 HP x 1.3, 2 HP x 1.25, 3 HP x 1.15, 5 HP x 1.15.
• Vertical lift uses the pumping water level, not the static level.
• Design flow is capped at the well’s sustainable yield where one is given.

A motor’s horsepower rating is its shaft output, so the efficiency that belongs in this step is the pump end’s alone; motor losses show up on the electric bill, not in the motor size. The efficiency points and the service-factor rule were calibrated against the Goulds GS selection charts (5 to 25 GPM series, 1/2 to 3 HP): a 1 HP pump in the 10 GPM series delivers 10 GPM at about 300 feet of total head, and the tool recommends 1 HP at that flow up to about 290 feet, just on the safe side of the chart. Efficiency is still the softest number here and varies between models, which is why the tool reports water horsepower separately and presents motor size as a starting point for reading a manufacturer’s pump curve rather than as a selection.

The yield cap is a deliberate departure from the published guidance, which sizes pumps to household demand alone. A pump cannot produce water the well does not have, and running dry destroys the motor, which is cooled by the water passing it.

The pump diagnostic scores each candidate cause against the symptoms selected, using fixed weights, and normalises the best match to 100%. The weights encode one claim in particular: rapid cycling is weighted most heavily toward a waterlogged pressure tank rather than a failing pump, because that is where it usually leads and because the cost of getting it wrong is asymmetric.

The percentages are a ranking aid, not probabilities. They express how well a cause explains the symptoms you ticked, nothing more.

The hardness-by-zip lookup and map are built from USGS groundwater samples published through the Water Quality Portal. Each well counts once (its own median), counties are summarised by the median across wells, and samples from landfill and contamination monitoring wells or above 2,000 mg/L are excluded. The full filter log, sample counts and caveats are published with the data, along with the date it was last downloaded.

Hardness dataset methodology→

Equipment cost ranges are planning figures compiled from published pricing for typical residential systems. They are not quotes and they are not based on our own purchasing. Installed cost varies widely with plumbing, water chemistry and region.

Nothing on this site is a substitute for a certified laboratory test or advice from a licensed professional. Where a result touches health — bacteria, nitrate, arsenic — the right next call is your county or state health department, not a website.