Why Your PCB Resets When a Motor Starts
A bench-first guide to finding voltage dips, ground shifts and noise in low-voltage motor-controller boards.
When a motor starts and the PCB resets, check the controller supply, reset pin and ground during the same startup event before changing components.
Start by finding out what actually reset
A controller works on the bench until the motor starts. The display blinks, the microcontroller reboots and the startup sequence begins again. It's tempting to add a larger capacitor and move on. But a reset that only happens under load may be caused by a voltage dip, a disturbance on the reset pin, or a watchdog timing out. Those faults call for different fixes.
Start by logging the reset-cause register as early in boot as possible, before startup code clears it. Check the microcontroller's documentation: some flags persist across several events, and not every reset leaves a useful flag. A brownout flag points you toward the supply; a watchdog flag tells you a task stopped responding, not why. First confirm that the controller rebooted. A display reset or a lost communications link can look similar.
1. Capture the motor start and the logic rail together
Record the start command or motor current on one channel and the controller supply at its local decoupling capacitor on another. Use a short probe ground connection at that capacitor. A long ground clip can pick up switching noise and make a clean rail look worse than it is. Use a properly rated differential probe to compare points that aren't at the same ground potential. Never clip an earth-referenced probe ground onto a switching node or lift the scope's protective earth.
Make two captures: one wide enough to see the whole start and another zoomed in around the switching edges. If you have more channels, also record the regulator input and the reset pin, each referenced appropriately. Compare the lowest supply voltage and dip duration with the controller's specified operating range and configured brownout threshold. Compare a failed start with a successful one under the same load. A multimeter's steady reading can miss the event entirely.
2. Follow the current through the board
The motor can draw its largest current just before it gets moving. That current flows out through the driver and back through the supply return. Look for parts of that path shared with the controller: a long cable, a small connector, a narrow trace or a plane neck. A shared impedance can shift the controller's local ground during startup even when the supply looks fine at the connector.
If both the regulator input and the logic rail dip, look upstream at the source, cable and connectors. Check whether the supply has hit its current limit. If the input stays steady but the logic rail dips, inspect the regulator's transient response and the placement and value of its recommended capacitors. If the local logic rail is stable but the reset pin glitches, check the reset supervisor, pull-up and routing near the switching loop. Stable power and reset pins with a watchdog flag? Check for a stalled task, communications errors or a peripheral losing power.
What the scope trace tells you:
| What you see | Where to check first |
|---|---|
| Regulator input and logic rail dip | Source current limit, cable and connectors |
| Regulator input steady; logic rail dips | Regulator transient response and local capacitors |
| Logic rail steady; reset pin glitches | Reset routing, supervisor and switching loop |
3. Fix the path that your measurements implicate
If the source is sagging, shorten or improve the power path, or use a driver's current limit or soft start if the load can still reach the required torque. Properly sized bulk capacitance on the driver's supply can help with short transients, but it won't make up for an undersized supply during a sustained load. Check capacitor voltage rating, ripple and inrush. For a regulator, follow the datasheet's capacitance, ESR and layout requirements rather than adding parts at random.
If the reset follows ground bounce or a switching edge, shrink the driver's high-current loop and keep its return from sharing a narrow path with the logic. Keep the controller's reference plane continuous; don't cut it just to separate labels called "power" and "digital." Route reset and other sensitive signals away from the switching node. Re-run the same capture after each change so you know what actually helped.
4. Prove the fix outside the easy bench setup
Repeat starts with the enclosure, production harness and actual load. Test at the lowest supported supply voltage and at relevant temperature limits. Try repeated starts, commanded stops and reversing if the product does that in service. For a blocked-rotor test, use a current-limited setup and stay within the hardware's specified limits. Watch for driver faults as well as controller resets.
Write down what passes: no unintended reset or lost control, supply and reset signals within their specified limits, and predictable recovery when the driver trips. Keep a before-and-after trace with the power source, harness, load and firmware version noted. That gives the next revision a repeatable test, not a bench anecdote.
Frequently asked questions
Is a bigger capacitor always the right fix?
No. It can help with a short voltage dip, but not an undersized supply, a disturbed reset input or a reproducible software stall. Measure first and verify that inrush remains acceptable.
What should I probe first if I only have two channels?
Trigger on the motor-enable signal and measure the controller's supply at its nearest decoupling capacitor. Then repeat with the reset pin or regulator input on the second channel, depending on what the first test shows.
Can software fix motor-start resets?
It can help when the problem is a watchdog timeout or when controlling motor acceleration reduces a measured surge. It cannot replace adequate voltage margin. Don't disable brownout protection to hide a dip.
Ardra Lab develops electronics products in Vadodara, Gujarat, India. Contact us to discuss power integrity, embedded hardware and PCB validation.
Contact Mr. Prit Varmora · service@ardralab.com · +91 93164 94418 · www.ardralab.com
Office: 402, Sampatti Complex, Sardarbaug Lane, Race Course, Vadodara, Gujarat, India – 390007
Further reading: Texas Instruments — Best Practices for Board Layout of Motor Drivers
原文:https://www.a8q9.cn/en/article/why-your-pcb-resets-when-a-motor-starts-813d2e.html