Multi-Sourcing High-End Boards Without Losing Consistency
When a class of board becomes scarce, the instinct is to spread the requirement across more than one supplier. The instinct is right, and the execution is where projects come undone, because a second source that produces a board which is not quite the same as the first introduces a new problem in place of the one it solved.
Multi-sourcing therefore has to be planned as a technical arrangement rather than a purchasing one. It works when the design, the stack-up, the approvals and the records are in a form that another plant can follow, and it fails when those exist only in the head of whoever placed the original order.
Why One Source Is Not Enough
A single supplier exposes a programme to that supplier’s own constraints. A plant operating near its capacity will prioritise the work already committed, and a new order, however large, joins a queue that was arranged before it arrived.
The exposure is not only commercial. A process failure, a fire, a material shortage or a change of management at one plant can interrupt a product that has no alternative route, and the interruption is measured in months for a deep stack where material and lamination both have long lead times.
The argument for a second source is consequently about continuity rather than price. The second supplier does not need to take a large share of the work; it needs to be ready, which means qualified and, at least occasionally, used.
What Has to Be Identical
The board itself can be reproduced exactly if the documents are complete. The stack-up, the materials and their approved alternatives, the impedance targets, the finish, the tolerances and the panel arrangement are all things that can be written down, and where they are written down a competent second plant can build to them.
What cannot be reproduced is accumulated habit. Two plants will not run their lines identically, so the assembly that follows the board may see small differences even when the boards meet the same specification. That is one reason to qualify the second source before it is needed rather than during a crisis.
The practical approach is to define what is fixed and what may vary. Material families and impedance tolerance are fixed; the exact process parameters of a lamination press are not, provided the measured result falls within the specification.

Qualifying a Second Source
Qualification is a small batch and a set of measurements. The first boards from the new source are built to the same stack-up, tested against the same limits and compared with the results from the established supplier.
The comparison is what makes the qualification meaningful. Where the impedance measurements, the cross-sections and the electrical test results agree, the two sources can be treated as interchangeable for that product. Where they differ, the difference is identified while volumes are still small.
The qualification also tests the working relationship: how long the review took, what questions were asked, whether the schedule was met. Those are the attributes that will decide the outcome when the second source is used in earnest.
Capacity and the Shape of the Commitment
A second source that has been qualified and never used will not hold capacity for a customer who has never ordered. A capacity reservation is therefore a real arrangement, and it usually means a periodic order rather than a promise.
What matters practically is that the second supplier knows the product, holds the stencil and the programme where assembly is involved, and can be brought up to speed in weeks rather than in a first-time qualification. Where the relationship exists, the ramp is short; where it does not, the qualification has to happen at the moment when time is least available.
For the customer, the cost of maintaining that readiness is small compared with the cost of a gap in supply, and it is worth treating as an insurance premium rather than as an inefficiency.
<img src="https://www.gopcba.com/wp-content/uploads/2020/05/steel_product1.jpg" alt="buffer stock of laminated boards in storage” />
Buffer Stock and What It Is For
Stock held against a supply risk is different from stock held against ordinary variation. The first is a deliberate decision to hold boards, or material, sufficient to cover a defined interruption; the second is a working quantity maintained to keep production moving.
The buffer has to be sized against the real vulnerability. Where the lead time for the board is measured in months and the lead time for a key material is comparable, a buffer sized in weeks is decorative. Where the risk is in the assembly stage rather than the board, the appropriate buffer is assembled units or the material that would otherwise be short.
Its cost is visible: money tied up in goods, storage, and the risk of a design change making the stock obsolete. Against that, the alternative is a stopped production line, and the comparison is usually straightforward, provided the buffer is sized deliberately rather than set by habit.
Design Choices That Reduce the Exposure
Some of the risk is designed in. A stack-up that depends on a single material, a structure that only one plant can produce, or a specification written around one supplier’s process leaves no room to move when that route is constrained.
Designing in a little design margin — an approved alternative material, a structure that two plants can produce, an impedance window that can be met by more than one process — keeps the second source genuinely viable rather than nominally available.
The same principle applies to the assembly. Where the components are chosen with alternatives in mind, and the substitution rules are agreed in advance, a shortage at one device does not stop the batch.
Communication Between Two Suppliers
An arrangement with more than one source creates an information problem as well as a supply one. Changes have to reach both plants at the same time, or a revision that exists at one will be built at the other, and the mistake will surface as a batch that behaves differently from its predecessor.
The control is a single point of truth. Where one document states the current revision, the stack-up and the approved alternatives, and both plants are required to build to it, the risk of divergence is removed. Where each plant holds its own copy of the data, the two will eventually disagree.
It also helps to decide in advance how a change is communicated: who informs both plants, how quickly, and what happens to material already purchased against the previous revision. Those questions are easy to answer in a calm week and difficult to answer during a ramp.
Making the Arrangement Work
Multi-sourcing succeeds when it is documented. A written stack-up, a defined set of measurements, agreed alternatives and an identified contact at each plant are what allow a second supplier to be switched on without a redesign.
It also helps to treat the second source as part of the same system rather than as a fallback. Where the same test limits and the same traceability records apply to both, a delivery from either can be described in the same terms, and the customer is not asked to remember which plant produced which batch. The work concerned is SMT assembly, verification through PCBA testing and the controls that make two batches comparable under quality management.
FAQ
Does a second source need to produce the same board exactly? It needs to meet the same specification and measurable limits. The process may differ, provided the measurements agree.
How is a second source kept ready? By placing occasional orders, so that the product, the documents and the tooling remain current, rather than by an informal understanding.
What makes multi-sourcing fail? Documentation that lives with one supplier, specifications written around a single process, and a second source that has never been used.



