The packaging-capacity headlines keep coming. Every announcement reads like a continuous-production story — more wafers, more substrates, more volume. The reality on the assembly floor is different: the boards arrive in families, not in one endless stream. And that changes which cleaning machine makes sense.
This article is not a news recap and names no companies. It is a decision frame for the person who has to clean the boards that come off an advanced-packaging line — substrate modules, SiP carriers, mixed board families with different soil types and different traceability demands.
Key takeaways
- Packaging capacity grows in families, not in one continuous stream.
- A 810×560 mm basket covers most substrate and module sizes in one pass.
- Two variables decide the washer: changeover count × changeover loss.
- 0.45 µm recovery filtration is what makes bath sharing safe.
- Traceability means a sampling point per batch, not per board.
Capacity headlines ≠ continuous high volume
When a packaging facility announces expansion, the mental picture is a conveyor running 24/7. That picture is wrong for the boards that actually need cleaning.
Advanced packaging produces high-value, low-volume, high-mix work. Substrate modules and SiP carriers change size, fixture and soil type several times a week. The cleaning problem is not “how many boards per hour” — it is “how many changeovers per week, and what does each one cost.”
That is a different question than the one our earlier planning guide answers. If you are sizing a washer by peak board count and cycle time, that guide (planning cleaning capacity for the ramp) is the right tool. This article is for the line where the bottleneck is changeover, not throughput.
Variable 1: the changeover matrix
Method and scope. The matrix below is a one-page planning table; it defines what changes between batches and which cleaning action each change drives. Machine figures come from the CHUANGQI CQ-C743 specification sheet as published in October 2026. The matrix itself is a process tool — every row is something your line already knows, once it is written down.
| Changeover dimension | Cleaning action it drives | Verification after changeover |
|---|---|---|
| Board size / basket layout | Adjust nozzle height, basket dividers, cycle program | First-batch sample check for coverage |
| Fixture / carrier type | Swap or adjust carriers; check drain positions | Visual check + one ionic-cleanliness test board |
| Soil type (flux family) | Decide whether bath can be shared (see Variable 2) | Bath conductivity read + test-board result |
| Batch traceability level | Assign sampling point per batch, not per board | Record batch ID + result in the log |
Source: CHUANGQI CQ-C743 specification sheet, chuangqitech.com, 2026-10.
The matrix is a one-page table, but it is the single highest-leverage document on the line. When the matrix exists, a changeover stops being a guess and becomes a checklist.
The hardware side matters too: a basket that covers a wide size range absorbs most of the variation. The CHUANGQI CQ-C743 offline washer carries an 810×560×100 mm two-tier basket — most substrate and module sizes fit in one pass, which is exactly what a changeover-heavy line needs (specification sheet, chuangqitech.com, 2026-10).
Variable 2: where cleaning-fluid baths can and cannot be shared
The second variable is the one that quietly decides cost. Every time a bath must be dumped and refilled, the line pays in chemistry, downtime and wastewater. So the question is: which soil types can share one bath?
The rule is not “same bath for everything.” The safe boundary is set by filtration and cross-contamination risk:
- Families with the same flux family and compatible soils → can share the bath, with recovery filtration and scheduled conductivity checks.
- Families with aggressive or incompatible chemistries → must switch the bath or run a separate rinse stage.
The enabler is real filtration, not wishful thinking. The CQ-C743 uses 0.45 µm recovery filtration on the chemistry loop — fine enough to pull solder paste, rosin and flux particles out of circulation so a shared bath stays predictable (specification sheet, chuangqitech.com, 2026-10). Without that, “sharing” is a gamble.
Write the shared-bath boundary into the changeover matrix: bath ID, acceptable soil families, and the conductivity or test-board check that triggers a bath change.
Variable 3: traceability means a sampling point per batch
Advanced-packaging buyers do not ask “is it clean?” — they ask “can you prove it for this batch?”
The traceability model is per-batch, not per-board. That changes the cleaning protocol in three ways:
- A sampling point per batch: one test board per batch goes through the same basket, then gets an ionic-cleanliness check — not every board, but every batch gets an answer.
- The batch ID travels with the record: board family, date, basket number, bath state and result in one log line.
- The sampling point is written into the matrix, so the operator does not have to remember it.
This is a process design, not a machine feature. But the machine supports it: a batch washer with a defined basket and a repeatable cycle makes the batch the natural unit of proof — which is one more reason low-volume, high-mix lines lean toward batch rather than inline.
Variable 4: batch vs inline — the annual changeover count decides
The last variable is the purchase decision itself. The honest way to choose is a calculation, not a preference:
Annual changeover count × single-changeover loss (time + chemistry + rework) vs. the throughput gap of a batch washer.
If you change over 4 times a week, the loss math quickly exceeds the throughput advantage of an inline machine. If you run one family for weeks at a time, the opposite is true. The number that matters is the changeover count, and most lines already know it — they just have not put it into the washer decision.
One more reason batch fits: a batch washer is the same machine that does the process-development work — sample cleaning, first-article validation, small production runs. When the line grows, it does not become useless; it becomes the development and low-volume station while inline handles the steady stream.
Where this frame stops working
This frame assumes the boards are water-cleanable and arrive in identifiable families. If your residue is aggressive, your volumes are truly continuous, or your boards are too large for a 810×560 mm basket, the answer moves toward inline or a different process — run the changeover count calculation before you assume batch is right.
Why does advanced packaging need batch cleaning instead of inline?
Because the boards arrive in high-mix families with frequent changeovers; batch makes the batch the natural unit of proof and changeover management.
How many board families can one batch washer handle?
With an 810×560 mm two-tier basket and a changeover matrix, most substrate and module sizes fit — the matrix defines the families, not the machine.
When must I switch the cleaning fluid between batches?
When soil families are chemically incompatible; compatible families can share a bath with 0.45 µm recovery filtration and scheduled conductivity checks.
How do I prove a batch was cleaned correctly?
One ionic-cleanliness test board per batch, logged with the batch ID — a sampling point per batch, not per board.
Is an offline washer slower than an inline one?
At steady-state throughput, usually yes. The decision is changeover loss vs. throughput gap — run the annual-changeover calculation, not a speed comparison.
Specification values come from the CQ-C743 specification sheet as published in October 2026; the changeover and bath-sharing framework is a process recommendation, not a compliance statement.
Line runs several board families a week? Send your family list and changeover count through the cleaning application form — we reply within 24 hours with a changeover-matrix draft. Start with the machine: offline PCBA cleaning machine (CQ-C743).