All computation runs server-side on Pontus /model — this client ships no chemistry.
Makeup analysis — the water itself · feeds every tab
Scaling ions
Source B — second water to blend with the makeup
Plant & dynamics — drives the trajectory, the forecast band, and the cost
Rates & forecast — band width and dollars
Total-analytical basis (no activity-coefficient / ion-pairing correction), so absolutes run high vs a speciated engine — the crossing order is the signal. Library-canonical: calcite is the speciated saturation ratio, not a 10^LSI shortcut.
Mix the makeup (Source A) with Source B by volume. Conservative ions mix linearly; pH is solved through the carbonate equilibrium (mix alkalinity + total inorganic carbon, re-solve) — equilibrium pH, off-gassing aside.
Reclaim usually carries more silica and hardness, so blending it in lowers the cycle ceiling — the chart shows how far.
One model, two parameter sources. Derive: τ = V / blowdown straight from geometry. Set: the same constants unlock as knobs. 2nd order adds the controlled-valve loop — an underdamped controller overshoots, and overshoot can cross a scaling ceiling the steady state never reaches.