Remineralisation and blending

Puts a controlled amount of mineral back into water that has had it removed, so what reaches the machine sits inside the target band instead of below it.

Two mechanisms reach the same destination and they are not interchangeable. BLENDING diverts a measured proportion of the untreated or carbon-filtered supply around the treatment stage and recombines it afterwards; the output is a weighted average of two waters, which means everything in the bypassed stream — hardness, alkalinity, chloride, whatever the test found — arrives in proportion. DOSING passes the treated water through a medium that dissolves as water contacts it, adding mineral that was chosen rather than inherited. Blending costs nothing to run and gives you back a smaller version of the problem you paid to remove. Dosing costs a consumable and gives control over composition rather than only over quantity. Which is right depends entirely on why the water was stripped in the first place: if the supply was merely over-mineralised, a blend is the honest and cheap answer; if it carried something you specifically wanted gone, a blend reintroduces it and dosing is the only real option. The target in either case is the same and it is two-sided — enough dissolved mineral for flavour compounds to bind to and carry, and enough buffering that the water is not aggressive toward the metal it spends its life inside. THE FAILURE HIDDEN IN BOTH is drift. A blend valve delivers a proportion of a feed that changes with season, source and tank turnover; a dosing cartridge depletes from the day it is fitted. The output moves in both cases, and it moves silently, because the water looks and flows exactly the same the entire time. Nothing about this stage announces itself. It is verified with a reading or it is not verified at all.

• Decide the target before you decide the method, and use the one the site already publishes rather than inventing a house profile. The destination is a TDS of 75–250 ppm with 150 ppm named as the ideal, and a general hardness of 17–85 ppm as CaCO₃ with 68 ppm named as the ideal. Everything below is about how you get there and how you prove you are still there.
• Blend where the supply was merely over-treated, and dose where the supply carried something you deliberately removed. A blend returns a proportion of everything in the bypassed stream, so blending back into water you stripped because of a specific ion reintroduces that ion in miniature — cheaper, and pointless.
• Remineralise for TWO reasons and do not forget the second, because the second is the expensive one. Everybody quotes extraction: water with nothing dissolved in it has nothing for flavour to bind to and the cup reads hollow. Fewer people mention corrosion: unbuffered water attacks copper, brass and boiler metal, and that bill arrives long after the taste complaint was solved by other means.
• Verify at the tap the machine drinks from, never at the cartridge and never from the specification sheet. A stage that has been fitted is not a stage that is working, and the whole point of this device is a number you can only confirm by measuring it downstream.
• Put the re-check in the maintenance schedule with an owner and a date, alongside the cartridge changes. This is the only stage in the chain whose failure is entirely invisible — no taste of chlorine, no drop in flow, no discolouration. If nobody measures it, nobody finds out.
• Match the water to what it actually feeds. Espresso and filter are not obliged to drink the same water, and ice is a third case again: ice made from under-mineralised water dilutes an iced drink toward nothing as it melts. Decide per line rather than treating the bar as one outlet.
• Be very careful about moving the blend to fix a taste complaint without a reading. Opening a bypass because the coffee tastes thin will raise TDS and hardness together, and hardness is the one that comes back as scale months later. A taste adjustment made blind at this valve is a scale decision made blind at the same time.
• If you cannot commit to measuring it, do not build a system that requires measuring. A simpler treatment chain reliably run and honestly understood beats an elegant one nobody verifies — the failure mode of the elegant chain is silent, and the failure mode of the simple one is at least visible in the cup.
• Confirm the plumbing, the housing position and any backflow requirement with a licensed plumber before the counter is built, and let the local authority govern. A remineralisation stage added after the joinery is finished is a stage that ends up somewhere nobody can reach.