An acceptable flux residue level is a process window you characterize on your own line — not a number you copy from a table. You set it by fixing four variables, then proving the window still holds across batches, shifts and bath age.
This is the characterization method CHUANGQI uses when a customer asks what residue level their process can actually hold. The boundaries below come from our own inline water cleaning machine specification and factory acceptance protocol: rinse water above 15 MΩ, DI water metered at 10–16 L/min, wash–rinse–hot-air dry in a single pass. The article quotes no external numeric threshold on purpose. The single universal ionic limit was withdrawn from IPC J-STD-001 in a 2018 amendment, and acceptance is now built per product through surface insulation resistance testing, service history without residue-driven failures, or humidity-bias testing. (Standard framing follows IPC’s published material on cleanliness acceptance, ipc.org, 2026-06.)
Key takeaways
- An acceptable residue level is a process window you characterize, not a number you copy.
- Four variables set it: board and flux, geometry, rinse water, drying.
- CHUANGQI’s CQ-C9610 inline water cleaning machine holds rinse water above 15 MΩ and meters 10–16 L/min of DI water.
- One passing board proves nothing; the record has to repeat across batches and shifts.
Four Variables Decide the Residue Level Your Process Can Hold
Method: this sequence runs on an inline water cleaning machine in the CHUANGQI workshop over a 72-hour continuous window per run, one board family, one flux chemistry and one wash chemistry at a time, with the variables below held and recorded. The sample volume is the production lot of each run, and the criterion is repeatability across lots, shifts and bath age rather than a single passing measurement. The output is a criterion for your product; it is not a release document, and it replaces neither your customer’s requirement nor your own incoming inspection.
Variable 1 — board and flux chemistry. Residue is not a generic contaminant. A water-soluble flux leaves an ionic load that rinses away with the right water and dwell time; a rosin-based chemistry leaves a film that needs a detergent action first. Fix the board family and the flux class before you measure anything, because the removal rate changes when the chemistry changes.
Variable 2 — board geometry. Geometry decides what the process can physically reach. A low-standoff package traps liquid in a narrow gap, and a tall connector shields the surface behind it from a directed jet. Two boards with identical flux and identical residue mass can sit on opposite sides of the same window for this reason alone.
Variable 3 — rinse water quality and volume. Rinse water sets the floor of the whole process. If the rinse stream already carries ionic load, or the volume per board is short, the board can leave the machine cleaner in the wash stage than at the exit. Resistivity and flow are therefore recorded per run, not assumed.
Variable 4 — drying. A board that leaves the rinse stage wet keeps moving residue until it dries, and any liquid trapped under a component redistributes what was just removed. The dry stage is part of the cleaning criterion, not a convenience step after it.
Before the first run, write the following down. The list is short, and each line exists because we have seen a study become unrepeatable without it.
- The board family, layer count and the lowest standoff component on the assembly.
- The flux class, along with the supplier and the part number.
- The wash chemistry, its target dose, and how the dose is monitored.
- Rinse water resistivity at the start and at the end of the run.
- The dry stage setpoint and the exit temperature actually measured.
- The acceptance route agreed with the customer, in writing, before production.
| Variable | Boundary we control and record (CQ-C9610) | Why it moves the residue level |
|---|---|---|
| Rinse water quality | DI water resistivity above 15 MΩ, inflow 1.0 m³/h | The rinse stream sets the floor; loaded water can re-deposit what the wash lifted |
| Rinse volume per board | DI consumption 10–16 L/min, metered per cycle | Short rinse leaves chemistry behind; extra rinse costs water, not quality |
| Cycle and conveyor speed | 0–1500 mm/min, wash–DI rinse–hot-air dry in one pass | Dwell time decides how much residue the chemistry can lift before the board leaves the zone |
| Wash chemistry dose | Detergent consumption 1.0–2.5 L/h | A drifting dose changes the removal rate while every other setting stays the same |
| Board envelope | Workpiece up to 580 (W) × 450 (L) × 80 (H) mm | Height and density decide whether liquid and air reach a gap in time |
| Drying stage | Hot-air dry inside the same pass | Residue stays mobile while the board is wet and is locked in place once it dries |
Source: CHUANGQI specification sheet for the CQ-C9610 inline water cleaning machine plus our machine acceptance protocol, 2026-10. Run conditions: one board family per run; wash, rinse and dry completed within one conveyor pass; rinse resistivity held above 15 MΩ; detergent dose between 1.0 and 2.5 L/h.
The Window Is Repeated Evidence, Not One Clean Board
A single clean board is an anecdote. A window becomes a criterion only when the same result repeats while conditions change by themselves: a different production lot, a different shift, chemistry that has been in the bath longer, a different operator running the same recipe.
That is why the acceptance record is built in stages rather than as one final measurement. Each stage removes a different way for the process to fail, and each stage leaves a document behind — which is also what lets a customer audit the decision without re-running the study.
| Stage of the acceptance record | What it proves |
|---|---|
| Leak and spray-pressure calibration | Liquid actually reaches the surfaces the process claims to reach |
| Drying temperature uniformity test | No wet pockets remain that would redistribute residue after the rinse |
| 72-hour continuous run | The window holds beyond a warm-up rather than within the first hour of a demonstration |
| Ionic cleanliness verification on a test board | The rinse stage performs as specified at the point of acceptance |
Source: CHUANGQI acceptance protocol for inline water cleaning machines, 2026-10 — four stages: pressure and leak calibration, drying temperature uniformity, a 72-hour continuous run, and ionic cleanliness verification on a test board.
Limits of applicability. This method characterizes a machine-level window and the record that proves it; it is not a substitute for testing your own production boards with your own flux and your own acceptance criterion, and the numbers above belong to the CQ-C9610 specification rather than to your process. It is not suitable for a line whose board family, flux class or wash chemistry turns over week by week: with no stable variables there is no window to characterize, and the work belongs in continuous monitoring instead. Where a customer-specific ionic target applies, that target stays the customer’s to set; this method only tells you whether your process can hold it. A machine acceptance record also says nothing about the board population you have not run yet — the first new board family opens the question again.
Two existing guides cover the decisions on either side of this one. If the criterion itself is still open on your side, how clean is clean enough for a PCBA explains how acceptance lines get set per product. If you would rather characterize on your own boards first, the sample cleaning test route produces a before-and-after record you can hand to your customer. Where the window points to a wash step, our CQ-C9610 inline PCBA water cleaning machine is the platform whose process boundaries are tabulated above.
Frequently asked questions
How do you set an acceptable flux residue level?
Characterize it. Fix the board family, the flux class, the wash chemistry and the rinse water, hold them constant through a production lot, and record what the process delivers. The level you can accept is the upper edge of the range your process repeats, not a number taken from someone else’s table. Then write the acceptance route down with your customer so the criterion survives a change of engineer.
Does IPC give a residue number I can use directly?
Not as a single universal figure. The old unified ionic limit was withdrawn from IPC J-STD-001 in a 2018 amendment, and the current version of the standard sets out routes for establishing acceptance per product instead — surface insulation resistance testing, service history without residue-driven failures, or humidity-bias testing. Read the standard as a framework for building your criterion rather than as a table to look up.
How many boards do we have to test?
Enough to see the window repeat under conditions you do not control: several production lots, more than one shift, and a bath that has aged past its freshest state. A single passing board only shows that one board passed. The stage that most often catches a drifting window is the one that runs longest, which is why a continuous run sits inside our own acceptance chain.
What if the residue level looks fine but the field still fails?
Then the criterion is measuring the wrong thing for that product, usually because the residue is sitting where the measurement cannot see it — under a low-standoff package, beneath a shield can, or in a region the coating later seals. The answer is not a tighter number; it is a criterion tied to the failure mechanism, verified on the geometry you actually build.
Next step: send the board family, the flux class and the residual level your customer applies, and we will identify the variables that decide the window for that assembly and outline what a characterization run would record — start with a process characterization request.