Heavy Copper PCB Temperature Rise Test: Conditions and Pass Criteria

Why the copper-weight calculation is not the answer

Most high current boards start with a design chart. You take the current, the allowed temperature rise and the copper weight, and read off a conductor width from IPC-2152 or a similar method. That is the right way to size copper. It is not proof that the finished board runs at the temperature you expect.

The chart describes one conductor in defined conditions. Your product adds things the chart does not see: neighbouring traces that heat each other, plated holes and terminals that carry the same current through a smaller cross-section, solder joints with their own resistance, a heatsink or enclosure that changes airflow, and finished copper that may be thinner or thicker than the nominal weight. Any of these can move the hottest spot away from the place you calculated.

A temperature rise test measures the result of all of them together. That is why it belongs in the qualification plan, and why the conditions have to be agreed before anyone switches on the power supply.

Bare board or assembled board

The first decision is what you actually put on the bench.

• Bare board test. Current is fed through the copper paths only, usually via fixtures or temporary connections. It isolates the board itself: conductor geometry, finished copper and plated holes. It is useful when you compare suppliers or check a new copper weight, because the result is not mixed with assembly variables.
• Assembled board test. The board is tested with its terminals, connectors, busbars or power components fitted, often in its real enclosure or with its heatsink. It shows what the product will do in service, including joint and contact heating.

The two results are not interchangeable. A bare board can pass while a press-fit terminal or a poorly wetted solder joint becomes the hottest point once assembled. If the PCB supplier only builds the bare board, the assembled test is usually run by you or your assembler. Write down who runs which test so that each party knows what its result covers.

Conditions you must fix before the test

A temperature rise number is meaningless without its conditions. These are the items buyers most often leave out:

• Ambient. Still air or forced air, open bench or inside the enclosure, and the ambient temperature at which the rise is measured.
• Mounting. Board orientation, distance to other surfaces, and whether it is fixed to a heatsink or chassis.
• Current profile. Continuous DC, a duty cycle, or a peak pulse. Which nets carry current at the same time matters, because adjacent paths heat each other.
• Connection method. How current enters and leaves the board. Test leads and clamps can add heat or remove it, so the connection should match the product or be documented.
• Sample state. Which revision, which lot, and whether boards have already been through reflow.

If you have a product standard or customer specification that defines these conditions, reference it. If not, write them into the test plan as [N] values that you choose.

Where to measure

Temperature rise is measured at points, and the points decide what you learn. Agree them on a marked drawing rather than in words.

Typical measurement points are:

• The narrowest section of each high current path, where current density is highest.
• Plated holes or via arrays that transfer current between layers.
• Terminals, connector pins, press-fit or solder joints where current enters or leaves.
• Areas where several high current paths run close together.
• One reference point for ambient, away from the board.

State the sensor type and how it is attached, for example thermocouples bonded with a specified adhesive, or an infrared camera with an agreed emissivity setting for the surface finish. Bare copper, solder mask and gold finishes read differently on an infrared camera, so a thermal image without that setting can mislead. Thermal images are useful for finding the hotspot. Fixed sensors are better for the recorded number.

Steady state and the pass criterion

A heavy copper board takes time to stop heating. If readings are taken too early, the rise is understated.

Define steady state in the plan: the reading is taken when the temperature changes by no more than [N] °C over [N] minutes. Then define the pass criterion in one of two ways:

• Rise limit. Measured temperature minus ambient is at or below [N] °C at every agreed point.
• Absolute limit. Measured temperature stays below a value set by the weakest material in the stack, such as the laminate rating, solder mask or a fitted component.

Both values should come from your product requirements and material data, not from the first test result. If a point fails, the plan should say what happens next: repeat with the same setup to rule out measurement error, check finished copper thickness at the failing location, and record the cause before any design or process change.

What goes into the qualification record

A temperature rise result is only reusable if someone can repeat it. Ask for the record to include:

• Board part number, revision and lot or date code.
• Test sample state: bare or assembled, and what was fitted.
• Ambient, mounting, airflow and connection method, with a photo of the setup.
• Current profile for each net, and the power supply and meters used.
• Measurement point map, sensor type and attachment method.
• Temperature-time data to steady state for each point.
• Calculated rise, pass or fail per point, and the person who approved it.

Keep this record with the first article or sample approval. It becomes the reference when you compare a later lot or a second supplier.

When the test has to be repeated

A passed test covers the design and process that were tested. Write the triggers for retesting into the PO or quality agreement. Common triggers are:

• Change of copper weight, conductor width or layer allocation of the high current paths.
• Change of laminate, surface finish or plated hole structure.
• Change of terminal, connector or press-fit part.
• Transfer to another factory or a second supplier.
• An increase in rated current or a new operating environment.

Whether you also sample lots in production is a cost decision. Many buyers do a full test at qualification and rely on finished copper thickness checks and electrical resistance checks per lot, then retest only on a trigger.

PO wording you can adapt

The text below is a starting point for your own terms. It is not legal advice and not PCB168 standard terms. Replace every [N] with values you have chosen.

> Temperature rise verification. The board shall be tested for temperature rise per the buyer test plan [document no., revision]. Test sample: [bare board / assembled board per drawing no.]. Executed by: [supplier / buyer / assembler]. Ambient [N] °C, [still air / forced air at N m/s], mounting per plan. Current per net as listed in the plan, applied [continuously / per duty cycle]. Measurement at points T1 to T[N] per marked drawing, using [sensor type and attachment]. Steady state when change is ≤ [N] °C over [N] min. Pass when rise above ambient is ≤ [N] °C at all points. A failing point places the lot on hold pending root cause and buyer disposition. Retest required on any change listed in the plan's change triggers. The test record shall include the items listed in the plan and be supplied with the sample approval package.

Keep the plan as a separate controlled document and reference it by number and revision. That way the PO stays short and the plan can be updated through change control.

How this fits with the rest of your heavy copper RFQ

The temperature rise test checks decisions made earlier in the RFQ:

• How you derived copper weight from current and allowed rise, covered in deriving copper weight from current and temperature rise.
• How much current your via arrays have to carry, covered in via current capacity RFQ inputs.
• How current enters the board through terminals and connectors, covered in high current connection interface RFQ.
• How acceptance evidence is listed in the purchase specification, covered in heavy copper PCB purchase specification and acceptance.

For build options, see the heavy copper PCB product page. When the test plan is ready, send it with your Gerber files, stack-up and current table through the quote request, so the supplier can say which part of the test it will run before pricing.

Checklist before you send the RFQ

• Test sample state chosen: bare, assembled, or both
• Party responsible for each test named
• Ambient, airflow and mounting defined
• Current per net and duty cycle listed
• Connection method matched to the product or documented
• Measurement points marked on a drawing
• Sensor type, attachment and infrared emissivity setting stated
• Steady-state rule written as numbers
• Rise or absolute limit written as numbers you chose
• Failure handling and lot hold rule stated
• Record content defined
• Retest triggers listed

FAQ

Is a calculated temperature rise enough for approval?

It is enough to size the copper, not to approve the product. The calculation does not include terminals, solder joints, neighbouring heat sources or the real enclosure. A test under agreed conditions shows the combined result.

Should the PCB supplier run the temperature rise test?

It can run a bare board test if both parties agree on the plan. A test of the assembled product is usually run by the buyer or the assembler, because it depends on parts and housings the PCB supplier does not control. Name the responsible party in the PO.

Can I use an infrared camera instead of thermocouples?

An infrared camera is good for finding hotspots. For the recorded value, fixed sensors at agreed points are easier to repeat. If you use infrared for the record, state the emissivity setting for each surface finish.

How long should the test run?

Until the agreed steady-state rule is met at every point. Write that rule as a temperature change over a time window rather than a fixed duration, because heavy copper boards with heatsinks can take longer to settle.

Who sets the allowed temperature rise?

You do, based on your product requirements and the ratings of the laminate, solder mask and fitted parts. Agree it in writing before testing.

Do I need to retest every production lot?

Not necessarily. Many buyers test fully at qualification, check finished copper and resistance per lot, and retest when a listed change happens. Lot testing is a deliberate cost choice.