Board Bring-Up: A First-Power Checklist for New Hardware
A new board straight from assembly has never been powered. It may have a short, a part fitted the wrong way round, or a design error that sends the wrong voltage to an expensive chip. Connecting it to a full supply and hoping risks damaging the one prototype that would have shown the problem. A methodical bring-up checks the board step by step, from passive measurements to running firmware. This article sets out that sequence.
At a glance
- Ohms first
- Measure each rail to ground before power is applied, to find shorts safely.
- Limit amps
- Use a bench supply with its current limit set just above the expected draw.
- One rail
- Bring up power in stages where the design allows, checking each before the next.
- Log it
- Record every measurement, so later boards can be compared with the first.
The problem
An assembly fault such as a solder bridge or a reversed diode can draw enough current to damage traces and parts in a moment if the supply is unlimited. Worse, the damage often destroys the evidence of what went wrong.
Design errors are equally possible on a first board: a regulator set to the wrong output, a missing pull-up that holds a chip in reset, or a sequencing problem between rails. Each is easy to diagnose in isolation and hard to untangle if everything is powered at once.
Measure before you power, limit before you switch on, and change one thing at a time.
The sequence that works
Start unpowered: inspect under magnification for orientation, bridges and missing parts, then measure resistance from each rail to ground and compare it with the expected value or a known-good board. A very low reading usually means a short, though large capacitances read low briefly while the meter charges them.
Then apply power through a current-limited supply, starting low, and watch the current as you raise it. Check every rail voltage, then clocks and reset lines, then connect the debugger and confirm the processor responds before loading application firmware and testing each interface in turn.
What good looks like
Prepare a bring-up plan listing every rail with its expected voltage, resistance to ground and current, and every clock and reset signal to check.
Work through it on the first board with current limiting, a check for hot components at each step, and a written log.
Record fixes and rework on the board and in the design files, then repeat a shorter sequence on later boards.
Bring-up sequence
- Inspect: Orientation, bridges, missing parts
- Measure: Rail resistance to ground
- Power: Current-limited, rising slowly
- Check: Rails, clocks, reset
- Connect: Debugger and test firmware
- Exercise: Each interface in turn
| Check | How | If wrong |
|---|---|---|
| Resistance to ground | Meter, unpowered | Look for bridges, reversed parts, wrong values |
| Supply current | Bench supply readout | Stop and find the hot component |
| Rail voltage | Meter at the load | Check regulator setting and feedback divider |
| Ripple and start-up | Oscilloscope, short ground lead | Check capacitors, layout and sequencing |
| Clock | Oscilloscope on a clock output pin | Check crystal, load capacitors and enables |
What it changes
Protected Prototypes
Faults are found before they cause damage.
Faster Diagnosis
Each stage isolates problems to a small part of the circuit.
Design Feedback
Errors are recorded and fixed in the next revision.
Repeatable Process
Later boards are brought up quickly from the same plan.
Where to go next
Design for bring-up: test points on every rail, links to isolate rails and an accessible debug header make this far easier.
Write simple test firmware that exercises each interface on its own. It separates hardware faults from application software problems.
Problems this prevents
- Prototypes damaged by an unlimited supply
- Evidence of a fault burnt away at first power
- Chips powered before rail voltages were checked
- Prototype fixes not carried into the design
Discuss your project
Email support@aavico.com or use Get in Touch when the website finishes loading.