Cleaning Bath Concentration Monitoring: Conductivity, pH and Bath Life
A cleaning bath does not fail suddenly; it drifts. Detergent is consumed by the flux load, water is dragged out on the boards, contamination accumulates, and the bath that passed an ionic test in January will leave residues in June unless its bath concentration is measured and corrected on a fixed schedule.
What a Cleaning Bath Has to Remove and Why Concentration Matters
An aqueous cleaning bath removes three things: flux residues, ionic contamination from plating and handling, and particulate matter from the board surface. The detergent works by lowering the surface tension so that water reaches under components, and by saponifying the rosin or the organic acid so that it becomes water soluble. Both mechanisms depend on having enough active chemistry in solution, and both stop working when the bath concentration falls below the point where the micelles break down.
Below that concentration the bath still looks wet and still rinses, which is why the failure is so easily missed. Boards come out of the machine clean to the eye, and the defect appears later as a coating adhesion failure or as electrochemical migration under a component after a few hundred hours of operation. Measuring the bath concentration is therefore a preventive task, not a response to a defect.
Measuring Concentration: Titration, Conductivity and Refractive Index
Titration is the reference method. A sample is titrated against a standard acid or base to an endpoint that reflects the active content, and the result is expressed as a percentage of the nominal concentration. It is accurate and hard to falsify, but it takes several minutes per sample, which is why most lines use it to calibrate a faster method rather than to control the bath in real time.
The two fast methods are conductivity and refractive index. Conductivity responds to the total ionic content, which means it tracks the detergent and the accumulated ionic contamination together, so a reading that is stable can hide a bath where detergent is falling while flux salts are rising. Refractive index responds to dissolved solids and suffers from the same ambiguity. Either method works as a control signal only if it is cross checked by titration at a defined interval, typically once per shift. The practical arrangement is a conductivity probe in the main bath with a display at the machine, backed by a daily titration and a weekly full analysis. Record both numbers side by side. When conductivity rises while titration falls, the bath is loading with contamination faster than it is being replenished, and the answer is a partial dump rather than more detergent.
Setting the Control Window and the Dosing Rule
The control window should be set around the supplier’s recommended working concentration with a tolerance that the process can actually hold. A typical window for a mildly alkaline cleaner is plus or minus 10 percent of nominal, with automatic dosing triggered at the lower limit and an alarm at the upper limit. Dosing on the lower limit rather than on a fixed timer prevents over concentration when the line is running a product with a light flux load.
Dosing needs a mixing delay and a re measurement before the next board enters. Adding concentrate raises the local concentration at the dosing point far above the bulk value, and a board that passes through that zone before mixing is complete sees chemistry that can attack solder mask or aluminium. Give the bath at least five minutes of circulation after a dose, and confirm with a second reading before releasing production.
Saponifiers, Surfactants and the Effect of pH
A saponifier converts rosin flux into a water soluble soap, and it does so only above a certain pH. For most mildly alkaline cleaners that threshold is around pH 10.5, and the reaction slows sharply below it. Since the saponification reaction consumes alkalinity, pH falls as boards are cleaned, and the fall is proportional to the flux load rather than to the number of boards. A line running a high solids rosin flux will drop pH far faster than one running a no clean flux.
Conductivity alone cannot see this. Measure pH with a probe that is rated for the temperature and the chemistry, and calibrate it against fresh buffers every shift, because alkaline cleaning baths foul glass electrodes and the reading drifts downward as the reference junction clogs. Treat a pH that has fallen more than 0.5 units below the start of shift value as an event, and check the titration before adding alkali, since adding caustic to a bath that has already loaded with flux salts can precipitate them onto the boards.
Contamination That Builds Up and Cannot Be Dosed Away
Some contamination cannot be corrected by adding detergent. Dissolved metals, especially copper and tin from the plating and soldering operations, accumulate in the bath and eventually plate out onto the boards or stain the solder mask. Rosin breakdown products and surfactant degradation products accumulate as dissolved solids and reduce the bath’s ability to hold soils in suspension. Neither responds to dosing.
Measure the accumulation indirectly by tracking the amount of detergent added per unit of board area against the amount predicted from the flux load. When the actual addition rate climbs above the prediction and the bath still tests low, the bath is saturated and needs a partial or full replacement. A simple total dissolved solids measurement or a specific gravity reading taken at the same temperature each time gives an early indication, and it costs almost nothing to collect.
Rinse Water Quality and the Final Ionic Load
The rinse stages decide the final ionic load on the board, not the wash stage. A perfectly controlled wash followed by a rinse that is carrying 20 microsiemens of dissolved solids will leave a conductive film behind, and the board will fail an ionic contamination test even though the detergent did its job. Monitor rinse water conductivity at the final rinse overflow, and set the dump threshold on the conductivity of the incoming water plus an allowance for the soil carried over.
Counterflow rinse design matters here. Where the final rinse overflows back into the preceding rinse, the water quality in the earlier stage is worse than the water going in, and the gradient is what removes the last of the detergent. Check that the flow rates are set so that water moves from clean to dirty rather than the reverse, and verify with a conductivity reading at each stage rather than at the supply only.
Bath Life, Dumping Criteria and Changeover
A cleaning bath should be dumped on criteria rather than on a calendar. The criteria are a detergent concentration that can no longer be maintained within the window, a dissolved metal concentration above the process limit, a total dissolved solids value above the specified maximum, or a rinse water consumption rate that has become impractical. Any one of them is sufficient.
Changeover is the highest risk activity in the cleaning process. A freshly charged bath is at its most aggressive, so the first boards through it are the ones most likely to be attacked, especially around aluminium parts and unplated edges. Dilute to the working concentration and circulate to temperature before releasing production, and run a known good panel through the new bath as a verification. Record the date, the reason for dumping and the verification result in the same log as the routine readings.
Monitoring Hardware: Sensors, Sampling Points and Calibration
Sensor placement determines whether the reading describes the bath or a pocket of it. Install the conductivity probe in the main circulation line rather than in a stagnant corner, keep it downstream of the dosing point with enough pipe length for mixing, and mount it where it can be removed and cleaned during production. A probe that cannot be cleaned without breaking the line will be cleaned far too rarely.
Calibrate conductivity against standard solutions at the bath operating temperature, not at room temperature, because the temperature coefficient of an alkaline cleaner is significant. Calibrate pH against two buffers that bracket the working range, and replace the electrode on a schedule rather than when it fails. A pH electrode in an alkaline bath has a working life measured in months, and its failure mode is a slow downward drift that looks like a process change.
Records and the Link to Ionic Contamination Testing
The cleaning records and the assembly cleanliness test results have to be read together. A trend of rising ionic contamination on finished boards, matched against a trend of stable bath concentration and stable rinse conductivity, points at the board design or the flux rather than at the bath. Conversely, a bath that has been inside its window all month while the boards fail points at the sampling plan or the test itself.
For each production day, keep the bath concentration, pH, final rinse conductivity, dose quantity, board area processed and the ionic contamination result. Ten numbers per day are enough to reconstruct what happened when a coating adhesion problem appears three months later, and reconstructing it from memory is not possible.
FAQ
How often should a cleaning bath be tested? Titrate once per shift as the reference measurement, and read conductivity continuously with a display at the machine. Record pH once per shift as well, because it responds to the flux load rather than to the number of boards and is the earliest indication that a saponifying bath is losing its ability to work.
Why does conductivity stay normal while boards get dirtier? Conductivity measures total ionic content, so it cannot distinguish detergent from accumulated flux salts and dissolved metals. The bath can be loaded with contamination while the reading looks stable. Titration against a standard reveals the difference, which is why it remains the reference method.
When should the bath be dumped? Dump on criteria, not on a calendar: when the concentration can no longer be held in its window, when dissolved metals exceed the process limit, or when rinse water consumption becomes impractical. Any one criterion is sufficient, and the first boards after a changeover should be verified with a known good panel.
