PCBs: 9% Yield, Zero Defects. Is Your Supplier Lying?

A PCB supplier may say “zero defects” while reporting a 9% yield. Those two statements can appear together, but they do not describe the same thing.

Yield shows how many boards pass a defined process without rework or scrap. The defect rate shows how many faults were found under a specific inspection method. If the supplier does not explain the calculation, the numbers can create a false sense of control.

I do not treat “zero defects” as a complete quality result. I ask how the supplier measured it, when the test happened, and what happened to boards that failed.

What a 9% yield may mean

A reported 9% yield is unusually low for a finished PCB process, but the number may refer to a limited production stage rather than the complete order.

It could describe:

• A first trial run
• A new material or stack-up
• A single drilling or plating step
• The first pass before rework
• A small sample rather than the full batch
• A calculation based on panels instead of individual boards

A supplier may also use “9%” when the intended figure is “99%.” I would not guess. I would ask for the exact formula.

A useful yield formula looks like this:

First-pass yield = boards that pass without rework ÷ total boards tested × 100

If 90 boards pass inspection out of 100 tested boards, the first-pass yield is 90%. If 9 boards pass out of 100, the yield is 9%. The difference affects cost, delivery time, and product risk.

“Zero defects” needs a clear definition

Zero defects can mean different things at different suppliers.

One factory may mean that no defect was found during AOI. Another may mean that no customer complaint was received. A third may refer only to the final shipment after repaired boards were included.

These are not equal measurements.

When I review a supplier report, I look for:

• The inspection stage
• The sample size
• The inspection standard
• The defect categories
• The number of repaired boards
• The number of rejected boards
• The test equipment used
• The person or team responsible for approval

A statement such as “all boards passed” is useful only when the supplier shows how the boards were tested.

Ask for first-pass yield, not only final yield

Final yield can look healthy after rework. It does not show how many boards failed during production.

For example, a batch of 1,000 boards may have this result:

• 850 boards pass on the first test
• 120 boards pass after rework
• 30 boards are scrapped

The final accepted quantity is 970 boards, giving a 97% shipment yield. The first-pass yield is 85%.

Both figures matter. The 85% result points to process instability, extra labor, longer lead time, and a higher chance that hidden damage remains after repair.

I usually request these figures separately:

1. First-pass yield
2. Final accepted yield
3. Rework rate
4. Scrap rate
5. Customer return rate
6. Defects per million opportunities, when the supplier tracks it

The data should cover the same product, material, quantity, and production period. A number from another product line does not prove the quality of my order.

Review the defect list

Yield alone does not tell me what went wrong.

A 2% defect rate caused by cosmetic marks has a different effect from a 2% defect rate caused by open circuits, solder mask damage, or poor plating.

I ask the supplier to group defects into clear categories:

• Open circuits
• Short circuits
• Drill position errors
• Hole wall problems
• Plating thickness variation
• Solder mask misalignment
• Copper trace width errors
• Delamination
• Surface finish damage
• Warpage
• Contamination
• Dimensional errors

The defect trend matters more than a single report. If the same drill or plating issue appears across several batches, the supplier should show the corrective action and the follow-up result.

Check how the inspection was done

PCB inspection may include AOI, flying probe testing, ICT, functional testing, X-ray inspection, visual inspection, or cross-section analysis.

Each method finds different problems.

AOI can identify many layout and solder mask issues. It may not confirm every electrical condition. Flying probe testing checks selected electrical paths but may not replace functional testing for a complex assembly. X-ray inspection can help review hidden solder joints or internal structures. A visual check may identify surface damage but cannot confirm internal plating quality.

I ask three practical questions:

• What does the test detect?
• What does the test not detect?
• Was the test performed on every board or on a sample?

A supplier that answers these questions clearly gives me a better basis for judging risk.

Review the production records

A clean shipment report does not replace traceability.

For each batch, I prefer to see records linked to:

• Material lot
• Laminate type
• Copper thickness
• Surface finish
• Production date
• Machine or line
• Operator or inspection team
• Test program version
• Rework history
• Shipment quantity

If a field failure occurs, traceability helps identify whether the issue came from material, drilling, plating, imaging, assembly, storage, or handling.

A simple lot code on the packing label can make the investigation much easier. Without it, the supplier may only be able to provide a broad statement about the whole factory.

Watch for rework hidden inside the result

Rework is not always a problem. Some repairs are controlled and acceptable when the product specification allows them. The concern starts when rework is hidden or used to make a weak process look stable.

I ask:

• How many boards were repaired?
• What repair method was used?
• Who approved the repair?
• Was the repaired area tested again?
• Can the repair affect reliability?
• Are repaired boards marked in the internal record?

For a high-reliability product, I may set a separate limit for repaired boards. A supplier should not place repaired boards into the same quality figure as boards that passed on the first test without explaining the difference.

Confirm the acceptance standard

A supplier and buyer can use the same word, “pass,” while applying different criteria.

The purchase documents should state:

• Board dimensions and tolerances
• Layer count and stack-up
• Finished hole size
• Copper thickness
• Surface finish
• Solder mask requirements
• Electrical test coverage
• Warpage limits
• Cosmetic acceptance rules
• Packaging and moisture controls
• Applicable customer or industry requirements

For bare PCB acceptance, many buyers refer to IPC-A-600 as a quality reference. The exact class and project requirements still need to be agreed before production. A standard reference does not remove the need for a clear drawing and inspection plan.

Use a small pilot run

Before moving to full production, I prefer a controlled pilot order when the design, material, or supplier is new.

The pilot should answer practical questions:

• Can the factory build the stack-up as specified?
• Does the supplier hold the required drill and line tolerances?
• Are the test fixtures ready?
• Does the surface finish remain consistent?
• Are the packing methods suitable for transport?
• Does the factory report first-pass yield and rework clearly?

The pilot result should be reviewed against the same criteria planned for mass production. A supplier that provides detailed data during a small run is easier to assess than one that only sends a final “passed” message.

A common reporting mistake

Imagine a factory produces 500 panels. After rework, all accepted panels meet the shipping requirement. The supplier reports:

Final shipment yield: 100%

That statement may be accurate, but it leaves out the production history.

If 70 panels failed AOI, 25 needed electrical retesting, and 10 were repaired after plating problems, the first-pass result tells a different story. The supplier has still shipped accepted material, but the process may be consuming extra time and creating hidden risk.

I would ask for both numbers and the reason behind every major loss. A transparent report may show a lower first-pass yield, yet it gives me more useful information than a polished zero-defect claim.

Questions I send to a PCB supplier

I keep the request short and specific:

1. How do you calculate yield?
2. Is the reported figure first-pass yield or final yield?
3. Does it refer to boards, panels, or batches?
4. How many units were tested?
5. How many units were reworked?
6. What were the top three defect types?
7. Which tests covered the full quantity?
8. Which tests used sampling?
9. What inspection standard did you apply?
10. Can you provide the batch traceability record?
11. How do you control repaired boards?
12. What corrective action was taken for repeated defects?

The quality of the answers often tells me as much as the numbers.

A supplier does not need to promise perfect production. I prefer a supplier that shows the limits of its process, reports failures without hiding them, and links corrective action to later results.

A 9% yield and a zero-defect claim may refer to different stages, different samples, or different definitions. I check the calculation before judging the factory. I compare first-pass yield with final yield, review the defect records, confirm the inspection method, and set acceptance rules before production starts.

Clear data does not remove every manufacturing risk. It gives me a practical way to see that risk before it reaches my product.

When a PCB supplier says “zero defects” while reporting a 9% yield, I would pause before placing an order.

A 9% yield means about 9 out of 100 units pass the stated inspection criteria. Around 91 units need rework, scrapping, or further review. That result does not support a zero-defect production claim unless the supplier is using a different definition of yield.

This is where many sourcing discussions become unclear. “Yield” may refer to:

• First-pass yield after fabrication
• Final electrical test yield
• A small pilot batch
• One process step, not the complete PCB
• Units accepted after rework

“Zero defects” may also refer to shipped boards only. A supplier could remove failed units before shipment and report a clean delivery record, while the production process still produces a high number of defects.

I would ask the supplier to define both terms in writing.

A useful question is:

“Does the 9% yield describe first-pass production, final inspection, or the number of boards accepted after rework?”

The answer changes the risk level.

A PCB buyer should also request data from a real production lot. The report should show the lot size, defect types, inspection stage, rework count, scrap count, and final accepted quantity. A statement such as “our quality rate is 100%” gives little value without the supporting figures.

A practical example helps.

Suppose a factory produces 1,000 four-layer boards. Only 90 pass the first inspection. The other 910 boards contain issues such as open circuits, solder mask damage, drilling errors, registration problems, or plating variation. After repair and sorting, 985 boards may pass final inspection.

The shipment could contain 985 acceptable boards, but the first-pass yield remains 9%. Those are two different measurements. If I am planning a large order, the rework load, delivery risk, and hidden cost matter just as much as the final shipment count.

I would review the supplier through these steps:

1. Confirm the yield formula

Ask whether yield is calculated as:

Accepted units ÷ total units produced

If the supplier uses another formula, ask for a written explanation. Yield should not change from one report to another without a clear reason.

1. Separate defects from rework

A board that passes after repair is not the same as a board that passes without correction. I would request separate figures for:

• First-pass accepted units
• Reworked units
• Scrapped units
• Final accepted units
• Customer-returned units

This gives a more accurate view of process stability.

1. Check the inspection method

The supplier should explain how the boards are tested. Depending on the design, the inspection may include:

• Automated optical inspection
• Flying probe testing
• Electrical testing
• Microsection analysis
• Solderability checks
• Dimensional inspection
• Visual inspection against the approved standard

A supplier may report zero defects because the inspection scope is too narrow. A bare visual check cannot replace electrical testing when the PCB has dense routing or fine-pitch components.

1. Review defect records

I would not judge a factory only by its best batch. I would ask for defect Pareto data from recent production. Repeated defects reveal more than a general quality statement.

For example, if most failures come from drill registration, the factory may need better panel alignment control. If solder mask openings are often out of position, the artwork, exposure, or cleaning process may require adjustment.

1. Confirm the acceptance standard

The purchase order should name the inspection standard and acceptance criteria. The buyer and supplier need the same reference for issues such as scratches, exposed copper, hole size, annular ring, delamination, and surface finish.

A clear standard reduces disputes after delivery. It also prevents a supplier from calling a defect “acceptable” without prior agreement.

1. Ask for a corrective action report

When yield falls to 9%, I would expect more than a promise to improve. I would ask for:

• The root cause
• Containment action
• Permanent corrective action
• Process owner
• Verification method
• Target yield
• Review date

The supplier should be able to show whether the action worked in a later lot.

A 9% yield does not automatically mean the supplier should be rejected. A new tooling setup, unusual material, complex HDI design, or pilot run can create a temporary result. The key question is whether the supplier understands the cause and can show stable improvement with data.

I would be careful with any supplier that uses “zero defects” as a marketing phrase while avoiding production records. No manufacturing process can provide a useful guarantee without defining the inspection scope, sample size, defect class, and acceptance rule.

My preferred wording in a supply agreement would be more precise:

“The supplier shall meet the agreed first-pass yield target, complete all required inspections, report defects and rework, and ship only boards that meet the approved acceptance standard.”

That wording does not promise an unrealistic manufacturing result. It creates a measurable quality process.

When I evaluate a PCB supplier, I look past the clean shipment count. I want to know how many boards passed without repair, how many were reworked, why failures occurred, and whether the same issue returned in the next lot.

A supplier may deliver an acceptable shipment after heavy sorting. That is not the same as having a stable process. Clear yield definitions and production evidence give buyers a better basis for cost, lead-time, and quality decisions.

When a PCB supplier says “high quality,” I do not treat the phrase as proof. It may describe a reliable process, or it may hide a vague sales promise. For engineers, buyers, and product teams, the real question is simple: what evidence supports the claim?

A quality claim should connect to measurable details such as material type, layer structure, copper thickness, impedance control, solder mask quality, electrical testing, and delivery records. Without these details, a polished website can create more doubt than confidence.

What does “high-quality PCB” mean?

The answer depends on the board’s use.

A simple two-layer control board may need stable dimensions, clean drilling, good solderability, and basic electrical testing. A high-speed communication board may require controlled impedance, low signal loss, careful stack-up design, and tighter process control. A board for outdoor equipment may need stronger protection against moisture, heat, or vibration.

This is why I avoid judging PCB quality through a single phrase. “Premium PCB” does not tell me enough. A useful supplier should explain what quality means for the specific design.

I usually check these points:

• Base material and material grade
• Board thickness tolerance
• Copper weight and finished copper thickness
• Minimum trace and spacing
• Hole size and position tolerance
• Surface finish
• Solder mask alignment
• Impedance control method
• Electrical test coverage
• Inspection records
• Traceability of batches

A supplier that can answer these questions clearly is easier to assess than one that only uses broad praise.

Claim 1: “We use high-quality materials”

This statement sounds positive, yet it needs more detail.

The material may be standard FR-4, high-Tg FR-4, aluminum, polyimide, or another type. Each option fits a different design need. A board used near a heat source may need a material with stronger thermal performance. A flexible circuit needs a different construction from a rigid board.

I ask the supplier to provide the material brand, grade, Tg value where relevant, dielectric data when needed, and supporting documentation. The material name should match the design requirement and the production record.

One case I have seen involved a small controller board that passed basic assembly but showed warping after exposure to heat. The original order only stated “FR-4 material.” A later review found that the chosen material did not fit the temperature range of the product. The issue was not caused by one dramatic manufacturing error. The purchase description was simply too general.

Claim 2: “Every board is tested”

Testing is useful only when the test method matches the risk.

A flying probe test may suit a low-volume order with changing designs. A fixture-based electrical test can support larger repeat orders. Automated optical inspection can check component-side assembly or bare-board features, yet it does not replace every other inspection method.

I ask four direct questions:

1. What does the test check?
2. Is the test performed on every board or on a sample?
3. What happens when a board fails?
4. Can the supplier provide a test report linked to the production batch?

For bare PCBs, common checks may include continuity, isolation, open circuits, short circuits, hole condition, surface finish, and visual appearance. A high-speed design may need impedance testing as well.

A supplier does not need to promise that every board will fit every use. I prefer a supplier that explains the test limits and shares a clear response process for failed units.

Claim 3: “We meet international standards”

A standard name alone does not confirm that the product meets your needs.

Ask which standard applies, what scope it covers, and whether the certificate belongs to the actual production site. A management certificate may show that a system exists, while it does not prove that every board in your order meets a particular electrical or dimensional requirement.

I also compare the certificate details with the quotation and factory information. The company name, site address, validity period, and scope should make sense together. If the supplier avoids simple questions about the document, I treat the claim with care.

A better statement looks like this:

“Production follows the stated quality system. The order will be checked against the agreed drawing, material list, test plan, and acceptance criteria.”

That wording is less dramatic, but it gives me something I can verify.

Claim 4: “Zero defects”

This is a warning sign for me.

PCB production involves many process stages, from imaging and etching to lamination, drilling, plating, finishing, and testing. A supplier can reduce defects through process control, inspection, and corrective action. No marketing sentence can replace a clear acceptance standard.

I ask the supplier to define:

• Acceptable cosmetic marks
• Hole and board outline tolerances
• Copper thickness limits
• Solder mask requirements
• Electrical test conditions
• Rework and replacement terms
• Handling of repeated defects

A practical quality promise describes control methods rather than making an absolute claim.

A simple way to review a PCB supplier

I use a short review process before placing a larger order.

Step 1: Send complete design data.
Include Gerber files, drill files, stack-up details, material needs, surface finish, copper weight, dimensions, quantity, and testing requirements.

Step 2: Request a written review.
Ask the supplier to identify unclear points, possible manufacturing risks, and any changes needed before production.

Step 3: Compare the quotation with the claim.
Check whether the quoted price includes electrical testing, impedance testing, inspection reports, packaging, and any special process.

Step 4: Review a sample order.
Measure board thickness, inspect the finish, check drilling, compare the board outline with the drawing, and review the test report.

Step 5: Track repeat performance.
Record defect types, delivery condition, response time, and whether the supplier follows the approved design.

This process does not need a large quality department. A shared checklist and clear records can prevent many misunderstandings.

What should buyers look for in a quality claim?

A useful claim contains evidence, limits, and a way to check the result.

Good signs include:

• Clear process descriptions
• Specific material information
• Test reports linked to an order
• Defined inspection points
• Transparent tolerance data
• Photos or records from the actual production site
• A written response to nonconforming boards
• Consistent communication between engineering and sales

Warning signs include:

• Repeated use of broad praise
• No material grade or test method
• Certificates with unclear ownership
• Pressure to approve files without review
• Different answers from sales and engineering
• No written acceptance criteria
• Promises that sound too wide for the stated price

I do not reject a supplier because its website uses simple language. I look at whether the team can support its claims with documents, measurements, and clear answers.

A PCB supplier earns trust through repeatable work, not through the strongest slogan. When I review a quality claim, I turn each promise into a question: What is being checked? What is the limit? Who records the result? What happens when the board does not meet the agreed requirement?

Those questions move the discussion away from marketing language and toward production facts. That is where a PCB quality decision becomes more useful, more practical, and easier to defend.

“Zero-defect PCB” sounds reassuring. I still pause when I see that phrase on a supplier report.

A perfect result may describe a small sample, a limited inspection scope, or a production lot that has not faced field use. The phrase alone does not tell me how the number was measured. When I buy printed circuit boards, I need more than a clean percentage. I need evidence that connects the result to my product, process, and risk level.

The useful question is not only, “Did the supplier report zero defects?”

I ask, “What was inspected, how was it tested, and what could the process still miss?”

What “zero defects” may mean

A supplier may use the term to describe several different situations:

• No defects found during visual inspection
• No failures detected in electrical testing
• No defects found in a selected sample
• No customer complaints within a stated period
• No defects recorded in one production lot

These statements do not carry the same weight.

A visual check may identify scratches, missing markings, exposed copper, or poor solder mask coverage. It may not reveal an internal via crack, a weak laminate bond, or a conductor issue hidden between layers.

An electrical test can confirm continuity and isolation for tested nets. It may not confirm long-term thermal stability, plating strength, or assembly performance.

A sample inspection can reduce testing time. It cannot provide the same level of confidence as a test applied to every board.

When I review a quality claim, I look for the exact inspection method behind the number.

The sample size changes the meaning

Imagine a supplier checks 30 boards from a lot of 10,000 and finds no defects.

That result is useful, but it does not mean all 10,000 boards are defect-free. It means no defects appeared in the selected sample under the chosen inspection conditions.

A better report may include:

• Total quantity produced
• Quantity inspected
• Sampling plan
• Defect classification
• Inspection equipment
• Test coverage
• Acceptance limits
• Lot identification

Suppose another supplier tests all 10,000 boards through automated optical inspection and electrical testing. The second result gives me more information, though it still does not prove that every possible defect has been found.

Quality data should show its limits. Clear limits make a report more trustworthy.

Defect escape matters more than a factory percentage

A board can pass a factory inspection and still cause trouble after assembly.

For example, a PCB may pass a basic continuity test but fail during thermal cycling because of a weak plated through-hole. A board may look clean under a camera while its impedance falls outside the target range. A connector pad may meet a visual standard but lose solder strength after repeated assembly cycles.

This is why I track defect escape.

Useful metrics include:

• Defects found before shipment
• Defects found during assembly
• Customer returns
• Field failures
• Defects per million opportunities
• First-pass yield
• Rework rate
• Test coverage by defect type

The right metric depends on the product. A simple control board and a high-layer-count board for industrial equipment should not use the same quality discussion.

Ask what the inspection can detect

Every test has a detection range.

AOI can review many surface features at high speed. It may not identify every internal construction issue.

An automated electrical test can check selected circuit paths. It may not measure every mechanical or material property.

An X-ray inspection can help review hidden solder joints and certain internal structures. It adds cost and does not replace other tests.

A cross-section analysis can show layer alignment, copper thickness, hole quality, and plating condition on a sample. It examines selected sections, not every board.

I prefer a quality plan that links each major risk to a suitable control. A supplier should be able to explain why a test is included and what type of defect it is designed to find.

Review the number behind the claim

Before accepting a “zero-defect” statement, I ask for several details:

1. What is the defect definition?

   Does one cosmetic mark count as a defect? Does a borderline measurement count? Are functional failures separated from appearance issues?

2. What was the inspection level?

   Was every unit tested, or was a sample selected?

3. What tools were used?

   The report should identify the inspection system, tester, measurement method, and calibration status where relevant.

4. What was the production period?

   One successful lot does not describe every future lot. I want to see whether the result covers a stable production history.

5. Were any units repaired or reworked?

   A shipped board may pass after rework, but the rework rate still tells me about process stability.

6. What happened after shipment?

   Customer feedback, assembly yield, and return data can reveal issues that factory testing did not catch.

A practical example

A purchasing team receives two supplier reports.

Supplier A states:

“10,000 PCBs produced. Zero defects.”

No sample size, test method, or defect definition appears on the page.

Supplier B reports:

• 10,000 boards produced
• 10,000 boards tested for continuity and isolation
• AOI applied to all boards
• 120 boards reviewed through cross-section sampling
• Three cosmetic issues found and accepted under the agreed standard
• No electrical failures
• Rework rate: 0.4%
• Lot number and test records available

Supplier B has not claimed perfection. I may still trust the second report more because it shows how the result was created. The small number of cosmetic findings does not automatically make the process poor. A transparent report gives me a better basis for judging risk.

Use a quality checklist before approval

I keep the review simple:

• Confirm the drawing revision
• Confirm material and layer count
• Check controlled impedance requirements
• Review hole, copper, and plating specifications
• Define acceptable cosmetic limits
• Match each major risk with a test
• Set the sampling or full-inspection plan
• Request test records by lot
• Review rework and failure data
• Approve a process-change notification rule

This approach helps prevent a polished slogan from replacing technical evidence.

A zero-defect result can be a good sign. It should not be treated as proof that every board is perfect under every condition. I trust PCB quality numbers when the supplier explains the scope, method, sample size, limits, and follow-up data.

The strongest quality message is not “nothing ever fails.” It is a clear record showing what was checked, what was found, what was corrected, and what the process can reasonably control.

When I compare PCB suppliers, I do not start with the lowest quote. I start with one question:

“What is your first-pass yield for a board with specifications similar to mine?”

This question can reveal more than a price sheet. A supplier may offer a low unit cost, yet repeated rework, scrap, and delayed shipments can raise the total cost. If a factory reports a 91% first-pass yield, 9 out of every 100 boards may need rework or may not pass the first test cycle. The number needs context, but it gives me a useful starting point.

A supplier should be able to explain how yield is calculated. I ask whether the figure covers:

• Bare board production
• Electrical testing
• Assembly and soldering
• First-pass inspection
• Reworked boards
• Scrapped boards

Some suppliers report only the boards that pass the final inspection. That figure may look strong while hiding defects found during drilling, plating, solder mask printing, or electrical testing.

I also ask for yield data from a product with similar requirements. A simple two-layer board and a high-density multilayer board do not create the same production risks. Fine traces, microvias, tight impedance limits, heavy copper, blind vias, and mixed materials can all affect the result.

A useful question is:

“Can you show yield records from a board with a similar layer count, line width, copper weight, and surface finish?”

The supplier may not share customer names or confidential documents. A general report with process data can still help me judge whether the answer is based on actual production work.

The next question focuses on design review:

“What does your DFM review check before production?”

A good DFM process should look at trace width, spacing, hole size, annular ring, solder mask clearance, copper balance, panel layout, and stackup design. It should also identify areas that may create problems during etching, drilling, plating, lamination, or assembly.

I prefer suppliers that raise design concerns before releasing the job. A short message such as “this via size may reduce drilling stability” can prevent a larger issue after production begins. I do not expect every supplier to approve every design without changes. I expect clear feedback, practical options, and written approval before the files move to production.

Material control deserves direct questions as well.

I ask:

• Which laminate brands do you use?
• Can you provide material certificates?
• How do you control material substitutions?
• What happens if the requested material is not available?
• Will I be notified before a substitute is used?

A material change can affect dielectric thickness, impedance, thermal performance, and soldering behavior. A substitute may appear similar on paper but perform differently in production. I want the supplier to request approval before making that change.

Process control also affects yield. I ask how the factory monitors drilling, plating thickness, etching, lamination pressure, and solder mask registration. The answer should include inspection methods, not only general statements about quality.

For example, I may ask whether the supplier uses:

• Automated optical inspection
• Flying probe testing
• Electrical testing
• X-ray inspection for selected structures
• Microsection analysis
• Dimensional checks
• Impedance testing when required

Each method serves a different purpose. Automated optical inspection can identify surface pattern issues. Electrical testing can find open and short circuits. Microsection analysis can show problems inside plated holes and layer structures. The right combination depends on the board design.

I also ask how the supplier handles a failed batch.

“Will you send a root-cause report when the yield falls below the agreed level?”

The report should explain the defect type, affected quantity, suspected cause, containment action, and correction plan. I do not need a long document filled with general language. I need information that helps me decide whether to change the design, the process, or the supplier.

A common example is a plated-through hole that fails electrical testing. The cause could be poor drilling, insufficient desmear, weak copper deposition, or contamination during processing. Each cause requires a different response. Replacing the boards without identifying the cause may leave the same problem in the next batch.

Traceability is another area that I check before placing an order. I ask whether the supplier can link finished boards to material lots, production dates, equipment records, operator checks, and test results. This information can help isolate a problem if a defect appears after delivery.

I also request a sample production plan. It should show the expected review points, pilot build process, inspection stages, and shipment conditions. A supplier that can explain the production path usually makes communication easier during development.

Price still matters, but it should sit beside yield, lead time, testing, material control, and response speed. A cheaper quotation may not be cheaper after freight, rework, replacement boards, assembly delays, and engineering time are included.

I use a simple comparison table with these fields:

• Quoted price
• Stated first-pass yield
• Yield definition
• DFM review scope
• Test coverage
• Material control
• Traceability
• Response to defects
• Pilot build support
• Delivery record

The supplier that gives the lowest number is not always the best fit. The supplier that answers clearly, supports the answers with process details, and accepts measurable quality terms may reduce more risk across the full project.

Before choosing a PCB supplier, I ask about the 9% that does not pass on the first attempt. I want to know where that loss comes from, how it is measured, and what the factory does to reduce it. That conversation often tells me more than a polished website or a low initial quote.

When a PCB supplier tells me, “We have zero defects,” I do not accept the statement at face value.

A clean quality record sounds reassuring, but the phrase can mean different things. It may refer to one production batch, one inspection method, or a customer return rate. It may also describe a target rather than a measured result.

For me, the useful question is not “Do you claim zero defects?”

It is:

How do you define a defect, how do you find it, and what evidence can you show?

Ask what “zero defects” means

A PCB supplier may use the term to describe:

• No defects found during final inspection
• No customer complaints within a set period
• No defects above an agreed acceptance level
• No defects in a small sample
• A production goal used by the quality team

These statements do not carry the same meaning.

A board may pass a visual check and still have an issue with solderability, impedance, via reliability, copper thickness, or electrical continuity. A small sample may also miss a fault that appears only once every few hundred boards.

I would ask the supplier to define the claim in writing. The definition should include the product type, batch size, inspection stage, defect categories, test method, and reporting period.

A clear statement is more useful than a broad promise.

Review the supplier’s quality records

I normally ask for records that show how the supplier tracks product quality. Useful documents may include:

• Incoming material inspection records
• In-process inspection reports
• AOI results
• Electrical test records
• X-ray inspection results for suitable assemblies
• Final inspection reports
• Customer complaint records
• Corrective action reports
• Yield and rework data

The goal is not to collect paperwork for its own sake. I want to see whether the supplier measures the process and responds when the numbers change.

A supplier that reports a small number of defects and explains how they were handled may be more trustworthy than one that reports perfect results without supporting data.

Check the inspection coverage

A PCB inspection system is only as useful as its coverage.

For bare PCBs, the supplier may need to check:

• Trace width and spacing
• Hole size and position
• Plating thickness
• Solder mask registration
• Surface finish
• Board thickness
• Warpage
• Impedance
• Electrical continuity and isolation
• Layer alignment
• Copper exposure
• Contamination

For PCB assemblies, the inspection plan may also include:

• Component placement
• Solder joint quality
• Polarity
• Missing parts
• Wrong parts
• Tombstoning
• Bridges
• Insufficient solder
• Functional performance

AOI can detect many visible placement and solder issues, but it does not replace every other test. X-ray can help inspect hidden solder joints, yet it may not identify all functional faults. A flying probe or fixture test can check electrical behavior, while functional testing can show whether the assembled product performs as expected.

I would ask the supplier to map each key risk to a specific inspection method. If a risk has no clear test, the quality plan has a gap.

Look at the acceptance standard

“Pass” has meaning only when both sides use the same standard.

For many PCB projects, the quality agreement may refer to IPC standards, customer drawings, approved samples, or project-specific requirements. The document should explain acceptable cosmetic marks, solder conditions, dimensional limits, electrical values, and other product requirements.

I would also check the defect classification:

• Critical defect
• Major defect
• Minor defect

The limits should match the product’s use. A cosmetic mark on a hidden area may not have the same effect as a short circuit near a power section.

A supplier should not decide acceptance rules after production. The buyer and supplier need to agree on them before the order enters manufacturing.

Ask how the supplier handles hidden defects

Some PCB problems do not appear during a quick visual inspection. Examples include:

• Internal delamination
• Weak plated-through holes
• Copper cracks
• Poor solderability
• Incorrect dielectric thickness
• Impedance outside the required range
• Thermal stress damage

I ask whether the supplier performs tests such as microsection analysis, thermal stress testing, solderability testing, or impedance testing when the design and application require them.

The right test depends on the board. A simple two-layer control board may need a different quality plan from a high-density multilayer board used in industrial equipment.

A supplier should be able to explain why a test is included or excluded.

Study traceability

When a problem appears, I need to know where the affected boards came from.

Good traceability may link the finished PCB to:

• Material lot
• Laminate batch
• Copper foil batch
• Production date
• Production line
• Machine settings
• Operator or shift
• Inspection records
• Test results
• Shipment details

A traceability system helps the supplier isolate affected products instead of checking every shipment without a clear direction.

Here is a common production example. A batch of assembled control boards passes AOI, but several units fail during system testing. The supplier reviews component lots, placement files, reflow profiles, and functional test records. The review shows that one connector position had a process variation on a single production shift. The supplier can then hold the related boards, inspect the matching batch, and correct the process.

Without traceability, the investigation may become a long exchange of guesses.

Pay attention to how the supplier discusses mistakes

I learn a great deal from a supplier’s response to a difficult question.

A reliable quality conversation may include:

• What happened
• How the issue was detected
• Which products may be affected
• The suspected root cause
• Immediate containment
• Corrective action
• Verification after the change
• The person responsible
• The planned completion date

A short answer such as “This cannot happen in our factory” gives me little useful information. Manufacturing involves materials, machines, people, and changing conditions. A supplier that can discuss risk in a calm and specific way is easier to work with.

I also ask to see a redacted corrective action report. It can show whether the supplier solves the cause or only sorts the affected boards.

Use a receiving inspection plan

Even when a supplier has a strong quality system, I still maintain incoming inspection for the risk level of the product.

The plan may include:

• Quantity check
• Part number check
• Packaging condition
• Visual inspection
• Dimensional sampling
• Electrical test review
• Sample functional testing
• Documentation check

The sample size should reflect the product value, failure risk, and production history. A simple visual sample may be enough for one order. A safety-related or high-cost assembly may need a wider inspection plan.

Incoming inspection is not a sign that the supplier has failed. It is part of managing the whole supply chain.

Decide whether the claim is supported

I do not need a supplier to promise perfection. I need clear requirements, suitable testing, stable records, and a response plan when a problem appears.

Before approving a PCB supplier, I would compare these points:

1. Definition of “zero defects”
2. Inspection methods and coverage
3. Applicable quality standards
4. Test reports and production data
5. Traceability from material to shipment
6. Corrective action process
7. Sample approval and change control
8. Incoming inspection expectations

“Zero defects” can be a useful quality target. It should not replace evidence.

The safest decision comes from matching the supplier’s claim with its inspection data, process controls, and ability to explain risk. A supplier that says less but can show how quality is managed may deserve more confidence than one that makes a perfect-sounding statement.

We welcome your inquiries: mr.xu@lingchaopcb.com/WhatsApp +8613780181891.

IPC International 2020 IPC-A-600K Acceptability of Printed Boards

IPC International 2020 IPC-6012E Qualification and Performance Specification for Rigid Printed Boards

IPC International 2012 IPC-2221B Generic Standard on Printed Board Design

International Organization for Standardization 2015 ISO 9001:2015 Quality Management Systems Requirements

Douglas C Montgomery 2019 Introduction to Statistical Quality Control Eighth Edition

Joseph M Juran and A Blanton Godfrey 2016 Juran’s Quality Handbook Seventh Edition