Jar Testing: How Operators Find the Right Coagulant Dose

It rained hard last night, and this morning your raw water doesn't look like yesterday's raw water. Turbidity is up, the color is off, and the alum dose that produced beautiful settled water all month is suddenly producing haze. Here is the uncomfortable truth at the center of coagulation: nobody can calculate the right dose. Not you, not the engineer who designed the plant, not the chemical rep. There is no formula. What there is, instead, is a row of six jars on a bench — and an hour from now, one of them will have the answer.

The jar test is the oldest piece of process control in water treatment, and it survives because it does something no instrument or equation can: it runs your entire plant — flash mix, flocculation basin, sedimentation — in miniature, six times at once, each with a different dose, and lets you watch which one wins.

Why You Can't Calculate a Coagulant Dose

The dose can't be calculated because coagulation is charge chemistry, not mass balance. The clay, silt, and organic particles in raw water carry a negative surface charge, so they repel each other and stay suspended — that's why raw water can hold its cloudiness for days. A coagulant like alum or ferric chloride works by releasing positively charged metal ions that neutralize that surface charge. Once the particles stop repelling, they can collide, stick, and grow into floc heavy enough to settle.

The right dose is whatever amount of positive charge cancels the negative charge that's actually in the water today. And that depends on turbidity, on organic content, on alkalinity, on pH, on temperature — variables that interact with each other and change with every storm, every season, every algae bloom. A formula would need to know the total surface charge of every particle in the river. The jar test doesn't need to know any of it. It just asks the water.

Get the dose wrong in either direction and the water tells on you. Too little, and particles are only partly neutralized — you get pin floc, tiny specks that form but never grow heavy enough to settle. Too much, and you swing past neutral: the excess positive charge coats the particles and they start repelling each other again — restabilization. That's the counterintuitive one, and it catches operators constantly. When settled water gets worse, the reflex is to feed more coagulant. Sometimes more is exactly the problem.

The Setup: Six Small Plants on a Bench

A jar test rig is a gang stirrer — one variable-speed motor driving a row of paddles — over four to six identical jars. The classic lab beaker works, but most plants use two-liter square jars (you'll hear the trade name B-KER²) because square corners break up the swirl and mix more like a real basin than a round beaker does.

You dose the jars from a dilute stock solution of the same coagulant the plant feeds, measured in with a syringe or pipette. The arithmetic is friendlier than it sounds: with a 1% stock solution, each milliliter added to a one-liter jar raises the dose by 10 mg/L — into a two-liter jar, 5 mg/L. Your plant's bench sheet will have this worked out for its jar size; the point is that each jar gets a known, deliberate dose.

And the doses are the whole strategy. You bracket: put your current plant dose in the middle of the range and step the jars around it. If the plant is feeding 20 mg/L of alum, a first round might run 10 – 15 – 20 – 25 – 30 – 35. One jar is always your current dose, so the test directly answers the question you're actually asking: is what I'm feeding still right?

The Procedure, Step by Step

A jar test imitates the plant in three timed phases — fast mix, slow mix, and stillness — because that's exactly what the water experiences on its way through the real basins. Exact speeds and times vary by plant and SOP; the numbers below are the common convention. What matters most is that you run every test the same way, so results compare across weeks and seasons.

1. Collect fresh raw water. Fill the jars from the raw tap and run the test promptly. Raw water sitting in a bucket starts settling and warming immediately — an hour-old sample is already a different water than the one entering your plant.
2. Load the doses. Draw each jar's dose into its own syringe before anything spins. You're about to need all six doses in the water at the same moment.
3. Rapid mix — about 1 minute at roughly 100 rpm. This is your flash mix chamber. Bring all jars to speed, inject every dose as close to simultaneously as you can manage, and let the violence do its job: dispersing the coagulant instantly through the whole jar. Charge neutralization happens in seconds — the coagulant has to meet the particles before it hydrolyzes away.
4. Slow mix — 15 to 30 minutes at roughly 25 to 35 rpm. This is your flocculation basin. The gentle rolling motion keeps particles colliding so floc can grow — but softly, because floc is fragile, and paddle speed that's too high tears apart everything it builds. This phase is where you start earning information: note when visible floc first appears in each jar, and what it looks like.
5. Settle — 15 to 30 minutes, paddles off. This is your sedimentation basin. Stop the stirrer, pull the paddles if your rig allows, and let gravity run the final exam. Good floc drops visibly, leaving clear supernatant above a blanket of solids. Weak floc hangs.
6. Sample and measure. Draw supernatant from the same depth in every jar — most square jars have a sample port at a fixed height for exactly this reason — and run turbidity on each. Check pH on the leading candidates too. Consistent sampling depth matters: skim one jar at the surface and siphon another near the floc blanket and you've corrupted your own comparison.

Reading the Jars

The turbidimeter gives you the ranking, but the jars themselves tell you why — and the why is what makes you better at this. Four things are worth reading in every round:

• Time to first floc. Jars with adequate dose show visible floc within the first few minutes of slow mix. A jar where floc is slow to appear is telling you the charge chemistry is starving.
• Floc character. Good floc looks like snowflakes or broken bits of tissue — discrete, well-formed, growing. Pin floc looks like dust that refuses to organize. Huge fluffy floc that shears apart when a paddle passes may be real but fragile — watch whether it survives to settle.
• Settling behavior. The best jars show floc dropping decisively once the paddles stop, building a defined blanket with sharp, clear water above it. Floc that hangs in the middle of the jar for ten minutes will hang in your sed basin too — and carry over to the filters.
• The supernatant numbers. Turbidity is the score. But read pH alongside it, because coagulant dose and pH are not independent — which brings us to the catch.

Picking the Winner (It's Not the Clearest Jar)

The winning jar is the lowest dose that meets your settled-water target — not the clearest jar outright. If 25 mg/L and 30 mg/L both settle to 1.5 NTU, the extra 5 mg/L is buying you nothing but chemical cost, extra sludge in the basin, and a bigger bite out of your alkalinity. And it's nudging you toward the restabilization wall. Optimum means enough, not most.

Then there's the second check that separates a jar-test operator from a jar-test technician: look at the pH of the winning jar. Alum is an acidic salt — every milligram you feed consumes roughly half a milligram of alkalinity (as CaCO₃) as it hydrolyzes. In well-buffered water that's background noise. In soft, low-alkalinity water, a big dose increase can drag finished-water pH down toward — or through — the floor your corrosion-control program depends on. A jar that clarifies beautifully at 40 mg/L but lands the pH at 6.6 hasn't solved your problem; it's traded a turbidity problem for a lead-and-copper problem.

From Jar to Plant: Scaling the Dose

The beautiful thing about mg/L is that it doesn't care about volume — the winning jar's dose is the plant's dose. What changes is the feed rate, and that's one formula:

If jar four won at 25 mg/L and the plant is treating 2.0 MGD:

Set the feeder, then verify the plant agrees with the bench: watch the flocculation basin for the same floc character you saw in the jar, and watch settled-water turbidity come down over the next detention time. The jar test is a prediction. The plant is the grade.

When to Run One

Run a jar test whenever the raw water changes or is about to. In practice that means:

• After storms and runoff events — the classic trigger. New turbidity, new organics, new charge demand.
• Seasonal transitions — spring runoff, summer algae, fall lake turnover. Each brings a different water to your intake.
• When settled turbidity drifts with no equipment explanation — before you blame the basin, ask whether the water changed underneath your dose.
• Before any chemical change — switching alum to ferric, adding a polymer, adjusting coagulation pH. Test it in a jar before you test it on the town.
• On a routine schedule — many plants run one weekly or monthly regardless, so there's always a current baseline to compare a crisis against.

The Mistakes That Quietly Ruin Jar Tests

The procedure is simple; the discipline is where jar tests go wrong. Five failure modes account for most bad bench data:

• Stale raw water. The sample you collected before lunch is not the water in your intake now. Collect fresh, run promptly.
• Bench conditions that don't match the plant. If your flocculation basin gives 20 minutes of detention, a 30-minute bench floc flatters every jar. Match your mixing times — and speeds, as close as your rig allows — to what your plant actually delivers.
• Reading too early. Floc that looks promising five minutes into settling can stall. Let the settle phase finish before you rank anything.
• Testing turbidity and ignoring pH. The clearest jar can be a corrosion-control liability. Read both, every round.
• Changing two variables at once. A round that varies dose and coagulant and pH answers nothing. One variable per round; run another round if you have a second question.

Frequently Asked Questions

The Bottom Line

Coagulation is the one process in your plant where the water gets a vote — and the jar test is how you count it. Six jars, three timed phases, one disciplined hour, and you walk away with the answer no formula can give you: the lowest dose that makes this morning's water settle clean. Bracket around your current dose, mix like your plant mixes, read the floc and not just the turbidimeter, check the pH before you crown a winner, and scale the champion with mg/L × MGD × 8.34. The plants that do this routinely aren't guessing when the storm hits. They're confirming.