PCB Fabricator Comparison: Signals of a Reliable Factory

Most quotations for the same board look similar, and the factories behind them are not. What separates a supplier that produces a predictable result from one that produces surprises is not visible in a price table, and it is only partly visible in a capability page.

A useful PCB fabricator comparison therefore looks past the advertised maximums and asks how the factory behaves when the order is unusual, when a measurement drifts, and when the same board is ordered again six months later.

Capability Statements Versus the Working Range

Every factory publishes a maximum. The maximum layer count, the minimum trace width, the smallest drill. Those numbers describe what has been done at least once, not what is done routinely.

The number that predicts your result is the working range: the trace and space the shop runs every day, the hole sizes produced without special tooling, the registration tolerance achieved across a normal lamination cycle, and the impedance tolerance that can be held without manual intervention.

Process capability is best discussed as a question rather than a claim. Ask which combination of parameters is routine, which requires extra review, and which would be a first for the factory. A supplier that answers that honestly is more useful than one that accepts everything.

The working range also explains pricing differences. Two quotations for the same board diverge when one factory treats the design as routine and the other treats it as an exception that needs additional inspection and a lower yield assumption.

The Engineering Review Is the First Real Signal

An engineering review is the cheapest operation in a fabrication order and the most revealing. The response to a data package shows whether the factory reads the design or only prices it.

A substantive review asks about dielectric thickness, material availability, registration across the lamination cycle, the drill schedule and the finish. It raises the places where the design assumes something the process cannot guarantee, and it returns an assumption list with the quotation rather than after the order is confirmed.

A quotation with no questions is not necessarily a sign of competence. On a complex board it often means the file was priced against a template, and the differences will appear later as clarifications, deviations or a delivery that does not match the drawing.

Quality Records and Batch Traceability

Quality records are what allow a delivered board to be connected to the process that produced it. That connection is what makes an investigation possible when a fault appears in the field.

At a minimum, a batch should be identifiable by a number that links it to the panel, the process parameters, the material lots and the inspection results. For impedance controlled designs, the coupon measurements should be reported rather than assumed. For sequential builds, the cross-section sample rate should be defined.

The retention period matters as well. A product with a ten year service life needs records that survive longer than a warranty period, and the answer to how long records are kept is a reasonable question to ask before a long-life product is placed with a supplier.

Where the fabricator also performs SMT PCB assembly, the records connect to the assembly and test data, which turns a fabrication deviation into something measurable on the finished product rather than a hypothesis.

How a Deviation Is Handled

No process is perfect, and the way a factory handles a deviation says more about its quality records than any certificate.

A board that falls slightly outside an impedance tolerance can be scrapped, reworked or shipped with a note. A batch that shows a registration trend can be stopped or pushed through. Those decisions should be governed by a rule agreed in advance, with the reporting that supports it.

Asking what happens when a measurement is out of specification is a good test of the relationship. A supplier that can describe the decision, the notification and the evidence is one that will not surprise the customer; a supplier that answers vaguely is one where the decision will be made by whoever is on shift.

The same applies to a shortage or a material substitution. The customer should know before the boards are made, not after they are shipped.

Consistency Across Reorders

Batch traceability has a second purpose: making the next order match the first. A build file that carries the stackup revision, the material grade, the copper weights, the drill schedule, the finish and the measured results allows a reorder to repeat a known process rather than reinterpret a drawing.

That record should also state the deviations that were accepted. If a substitute laminate was used on the pilot, the fact belongs in the file so that the next order can either repeat it deliberately or return to the original with a re-qualification.

Consistency is where a cheap supplier often becomes expensive. A factory that produces an acceptable pilot and a slightly different second batch forces the customer to re-verify a product that was supposed to be unchanged.

Questions That Produce Useful Answers

• Which parameter combinations are routine, and which need special handling?
• What does the engineering review cover, and is it returned with the quotation?
• How is a batch identified, and for how long are the records kept?
• How are impedance and other critical measurements reported?
• What happens when a measurement falls outside specification?
• How is a reorder placed so that it repeats the same build?
• What process does the fabricator follow when a material is unavailable?

The answers are short, comparable and difficult to fake, which makes them a better basis for a PCB fabricator comparison than a price list.

How a First Order Reveals the Factory

Capability claims are easy to make and hard to verify, which is why the first order should be designed to test the relationship rather than only to deliver boards.

Send a complete data package and observe how the review is returned. Note whether the questions arrive before the quotation or after the order, whether the assumptions are written down, and whether the delivery date is tied to a defined starting event.

Then examine the shipment. The documentation should identify the batch and report the measurements that were specified. A critical parameter that was agreed should appear as a number, not as a statement that the drawing was followed.

Finally, order the same board again. The second order is the real test of a fabricator, because it shows whether the process was recorded or remembered. A supplier that reproduces the build without renegotiating the specification has demonstrated the only capability that matters for a product in production.

FAQ

Is the cheapest quotation usually a false economy? Not always, but it often reflects a different assumption about the process. The comparison is only meaningful when both suppliers are quoting the same specification.

How can a customer verify capability without visiting? By asking for the working range, the review output and sample records, and by starting with a small order that includes a critical measurement.

Do certifications replace process data? No. A certification describes a system; the process data describes what happened to this board.

What is the most useful first question? Which part of this design is routine for you, and which part is not.

Summary

A meaningful fabricator comparison rests on four things: the real process capability rather than the maximum, a substantive engineering review returned before the order, records that provide batch traceability, and a rule for handling a deviation. Suppliers that answer those questions well tend to be the ones whose second delivery looks like their first.