A newly assembled board should reach first power-on with a written sequence, defined limits and a named decision owner. A power LED alone does not establish that the assembly is correct. This guide helps an OEM and assembly partner agree what to check before a pilot board becomes a functional-test reference.
1. Define the bring-up boundary
Use this checklist for a low-voltage board whose design and intended operating conditions are available to the engineer. Mains-connected hardware, high-energy battery systems and high-voltage circuits need their own qualified safety procedure. Do not transfer a bench setup from a different product simply because the connector fits.
Record the PCB revision, assembled BOM revision, serial number, fitted options and firmware status. Identify the schematic and test procedure that apply to this exact variant. Keep first power-on separate from shipment acceptance: it is an engineering gate, while production testing checks an agreed set of requirements consistently.
2. Complete the unpowered checks
- Compare fitted parts, DNP locations and component orientation with the released assembly drawing. Resolve conflicting polarity marks before connecting power.
- Inspect the power entry, connectors and accessible solder joints for bridges, damage or debris. Check the agreed AOI or X-ray evidence for hidden connections where applicable.
- Confirm jumper positions, programming headers and fixture pin assignments. Photograph the setup so another engineer can reproduce it.
- Measure resistance to ground on the planned supply rails with power disconnected and capacitors safely discharged. Record the instrument, polarity and settling behavior.
There is no universal resistance value that proves a rail is shorted or safe. Capacitors and semiconductor paths affect readings. Compare with the circuit design, device guidance and a characterized reference board, if one exists. An unexplained reading is a hold point, not a reason to raise the supply current limit.
3. Write the power plan before switching on
List every possible source of power: the bench supply, USB, debugger, communications interface and any battery. A connected interface may energize a rail even when the main supply is off. Analog Devices explains why rail sequencing and signal-driven back-powering can create damaging conditions in AN-932: Power Supply Sequencing. Apply the requirements of the actual devices in your design.
On small screens, scroll the table sideways to compare columns.
| Record | What to specify | Hold condition |
|---|---|---|
| Input source | Connector, polarity, voltage and grounding arrangement | Pin assignment or reference ground is uncertain |
| Current limit | Engineering-approved limit accounting for startup behavior | Repeated current limiting or unexpected consumption |
| Rail sequence | Enable order, required delays and discharge procedure | A rail rises out of order or outside its limit |
| Measurements | Test points, tolerances, instruments and capture method | A result is missing or cannot be reproduced |
| Stop authority | Engineer responsible for investigation and restart approval | An abnormal result has no disposition |
Select limits from design calculations, device documentation and the expected startup waveform. A limit that is too low can cause repeated restart; a high limit can allow excessive fault energy. Record the reasoning rather than supplying an arbitrary current setting for every board.
4. Energize in controlled steps
- Secure the board in a suitable fixture, verify the supply settings and connect measurement leads with power off.
- Apply power using the approved sequence. Observe input current and the first supply rail before enabling additional functions where the design permits this.
- Check voltage at the defined test points, including startup behavior if sequencing or overshoot matters. Record actual measurements rather than “looks good.”
- Confirm resets, clocks and basic communications using the planned tools. Introduce firmware and external loads in the documented order.
- Stop on unexpected heating, current limiting, rail instability or unexplained behavior. Remove power safely and retain the original evidence before changing the setup.
5. Example: a board repeatedly restarts
Hypothetical teaching example: a pilot board starts, loses its power-good signal and starts again. The engineer records the input-current and rail waveforms, then checks the supply limit, cable voltage drop, enable timing and load configuration against the approved plan. The team does not immediately label this a solder defect or increase the current limit.
If a fixture issue is found, document its correction and repeat the affected checks. If an assembly or design issue is found, route it through the relevant change process. Retain the failed result alongside the passing retest so the final record explains what changed. This example is a diagnostic workflow, not a customer result.
6. Turn bring-up evidence into a production handoff
Save a record containing serial number, hardware and firmware revisions, setup photo, instrument identification, limits, measured values, waveform files, exceptions and reviewer approval. Decide which checks belong in the repeatable production test and which remain engineering characterization.
Use the functional-test fixture handoff guide to define production interfaces and the prototype-to-low-volume ramp guide to close issues before releasing the next batch.
Frequently asked questions
Does successful first power-on mean the board can ship?
No. Shipment still requires the agreed inspection, programming, functional testing and release records. Bring-up may not exercise every interface, load condition or protection function.
Should every unit use the engineering bench procedure?
Not necessarily. Convert the approved checks into a controlled production method with clear limits and traceable results. Keep exploratory debugging out of the routine pass/fail sequence.