Coagulation and the Jar Test

The test needs six jars stirred identically at a speed you can set and change together. Buying one is possible; building one is an afternoon and teaches you what the specification is for. Mount a low-speed geared motor driving a shaft, and hang six paddles from it on a common belt or gear train so they all turn at the same rate. Speed control matters more than power — you need a smooth range from roughly twenty to two hundred revolutions per minute, so use a motor controller rather than a resistor. The paddles should be flat, about seventy-five millimetres wide and twenty-five deep, set at a consistent height above the jar bottom. Consistency between the six is the entire point of the instrument. Any difference in paddle depth or size becomes a difference in mixing energy, and mixing energy is one of the variables you are trying to hold constant while you change the dose. Use square jars if you can get them, or add a baffle to round ones. A round unbaffled jar lets the whole body of water rotate with the paddle, which looks like vigorous stirring and delivers very little actual mixing. Mark a fixed sampling point on each jar, about a third of the way down. Where you draw the sample from changes the answer, so it must be the same place every time.

Collect raw water with something actually in it. A pond, a stream after rain, or tap water deliberately loaded with a suspension of fine clay if you have nothing else. Measure and record its turbidity, pH and temperature before you touch it — temperature matters, because cold water flocculates noticeably more slowly and a test run in winter needs a longer slow-mix. Make a stock alum solution rather than weighing powder into each jar. Ten grams per litre is convenient, and then a dose of twenty milligrams per litre into a one-litre jar is two millilitres of stock. Dosing by volume from a stock is faster, far more accurate at small doses, and is what any works does. Fill six jars with identical raw water. Dose them across a range that brackets where you expect the optimum — zero, ten, twenty, thirty, forty and sixty milligrams per litre is a sensible first sweep. Add all six doses as close to simultaneously as you can, then start the rapid mix. Run the three phases as the flow diagram specifies. During the slow mix, watch and write down what you see: the dose at which floc first becomes visible, how large the particles grow, whether they look like snowflakes or like grains. Floc appearance is real information and experienced operators read it before they read the meter. After settling, sample each jar at your marked depth without disturbing the sediment, and measure turbidity and pH. Plot both against dose.

Your first sweep locates the optimum to within about ten milligrams per litre. Now run a second sweep across that narrow range in small steps and find it properly. This two-pass approach — coarse then fine — is how the test is used in practice and it is much faster than guessing a fine range first. Then change one thing at a time and watch the optimum move. Warm the water ten degrees. Raise the pH with a little sodium bicarbonate. Double the raw turbidity with more clay. Each will shift the answer, and seeing how far is the point: it demonstrates that the dose is a property of today's water, not of the plant. Finally, run a jar deliberately at three times the optimum and look at it. That is the restabilisation the notebook predicted, and watching genuinely dirty water come out of an over-dosed jar is the most useful thing in this blueprint. Where this stops, and this one has a sharp edge. Coagulation removes turbidity, much of the colour, and a good deal of what rides on particles. It does not disinfect. Water that has been beautifully coagulated and settled is clear, looks excellent, and can still contain everything that makes people ill — and it is more dangerous than muddy water precisely because it looks safe. Coagulation is a step in a train, never the whole train, and the train always ends in disinfection with a verified residual. Use this to clarify water for a pond, a workshop, a filter feed, or to understand what your water company does every hour. Do not treat clear as clean.