Production Yield Ramp Planning
A new product does not start at its mature yield, and the difference between the first week and the tenth is the ramp. Planning the yield ramp means knowing what the expected curve looks like, what will be learned at each stage and what the customer will accept in the meantime.
The ramp is a process activity rather than a production activity, and it requires the same discipline as any other process: measurement, a hypothesis and a change. Without it, the yield improves by accident and the knowledge is not retained.
What the Ramp Looks Like
A typical ramp starts low, improves quickly as the obvious problems are removed and then flattens as the remaining causes become smaller and harder to find. The shape is predictable even when the numbers are not.
Planning should be based on that shape, with a target for each stage rather than a single target for the whole period. A target that assumes mature yield from the first batch produces a plan that is met on paper and missed in practice.
The Pilot Build
The pilot is where the process is proved and the design is confirmed to be manufacturable. Its purpose is learning, so it should be built on the production line with production tooling and production documentation.
A pilot built on a laboratory bench proves nothing about the line, and the problems it avoids appear in the first production batch. This is the same reasoning that applies to the first article discipline in first article checks.
First Pass Yield as the Measure
First pass yield is the right measure for a ramp, because it excludes rework and reflects the state of the process. A final yield that includes rework hides the ramp behind the repair station.
The measurement should be taken per station as well as per product, so that the constraint on the ramp can be identified. A ramp that is limited by the inspection station is a different problem from one limited by the printer.
Learning Curve and Effort
Each improvement requires an investigation, and the effort per point of yield rises as the yield rises. Planning the ramp therefore means planning the engineering effort as well as the output.
Where the effort is not planned, the improvement stalls at the point where the easy causes have been removed and the remaining ones need a dedicated investigation.
Containment During the Ramp
While the yield is low, the escapes have to be controlled with additional inspection or additional test coverage. The containment is temporary and it should be designed to be removed as the process matures.
A containment that is not removed becomes a permanent cost, and one that is removed too early releases defects. The exit criterion should be written with the containment plan.
Design Changes During the Ramp
A design change made during the ramp resets part of the learning, because the process that was being characterised no longer exists. The change should be justified against the delay it causes.
Where a change is necessary, it should be introduced with a first article and a fresh measurement, and the effect on the yield curve recorded. This is the interaction between the ramp and the process described for revision handling.
Supplier and Material Readiness
A ramp often stalls because a material is not available in the required volume or because a supplier’s yield is lower at the new quantity. Both should be checked before the ramp starts rather than during it.
Where a second source is required, it should be qualified before the volume increases, since qualifying a supplier takes longer than the ramp itself.
Customer Communication
The customer should know the expected yield curve and the containment that supports it, particularly where the delivery schedule depends on the yield. A ramp that is discussed openly is a schedule risk that is shared.
Where samples are required for approval, they should be taken from a production batch rather than from a hand built unit, so that the approval reflects the product that will ship.
Records and Knowledge Retention
Each improvement should be recorded with the cause and the change, so that the knowledge survives the ramp. A product that ramps successfully and forgets why will repeat the exercise on the next one.
The records should be reviewed with the design team, because some of the causes will be design issues that can be avoided on the next product. This is the same feedback loop that supports the checks in the design release process.
Process Control and Verification
On a design of this kind, yield ramp is the item that decides how the rest of the board is arranged. A stack-up that is drawn rather than described removes most of the ambiguity from a quotation, and it lets the fabricator price the board against the dielectric and copper weights that will actually be used. Where the requirement is not written down, the shop supplies its own default, and the default is chosen for the process rather than for the design.
The measurements that matter are the repeatable ones: conductor width and spacing, annular ring, finished hole size, plating thickness and surface finish are all verifiable on a coupon that travels with the panel. Running a first article through the same checks as the production panel confirms that the two agree, and that comparison is the cheapest form of process control available at prototype stage.
Keeping a sample from the panel turns a dispute into a measurement, because the same coupon can be re-examined by both parties without rebuilding the batch.
Process Control and Verification
On a design of this kind, yield ramp is the item that decides how the rest of the board is arranged. A stack-up that is drawn rather than described removes most of the ambiguity from a quotation, and it lets the fabricator price the board against the dielectric and copper weights that will actually be used. Where the requirement is not written down, the shop supplies its own default, and the default is chosen for the process rather than for the design.
The measurements that matter are the repeatable ones: conductor width and spacing, annular ring, finished hole size, plating thickness and surface finish are all verifiable on a coupon that travels with the panel.
Process Control and Verification
On a design of this kind, yield ramp is the item that decides how the rest of the board is arranged. A stack-up that is drawn rather than described removes most of the ambiguity from a quotation, and it lets the fabricator price the board against the dielectric and copper weights that will actually be used. Where the requirement is not written down, the shop supplies its own default, and the default is chosen for the process rather than for the design.
A first article check confirms that the process and the drawing agree on the points listed above, and that the coupon data supports the values used in the design.
Where a measurement falls outside the expected window, the sample is retained so that the cause can be established before the balance of the batch is released.
A first article check confirms that the process and the drawing agree on the points listed above, and that the coupon data supports the values used in the design.
Where a measurement falls outside the expected window, the sample is retained so that the cause can be established before the balance of the batch is released.
FAQ
How long should a ramp take? It depends on the product and the maturity of the line, and the shape is more predictable than the duration.
Should the ramp target be a yield figure or a defect count? A yield figure is easier to communicate, while a defect count by cause is what drives the improvement.
Is rework acceptable during the ramp? It can be used for containment, and it should not be counted in the yield that measures the ramp.
What ends the ramp? A defined yield target sustained for a defined period with the containment removed.
