Please, have a look at the image:


The biggest pad at the top is the negative terminal of a battery holder. Through 0 Ohm resitor above it goes to a capacitor across a DC motor (not shown). Below the diode and off the image it goes to a ground of a charger connector.

Just below the pad there is a thin trace, leading to a small capacitor and a 5 pin SOT-packaged something. And a bit lower there is yet another ground connection, that is leading down to transistors and below.

I do agree on keeping DC motor noises as far as possible and it it well demonstrated here via 0 Ohm resistor up. But it is a mystery to me, why would they split ground connection more and more? They could've taken the first split and led it immediately to transistors. Or they could take the second split lower -- near a diode bottom left. Or they could lead a thick trace instead of split.

Is there any real need to split it so many times?

I also wonder, what is the purpose of so many test-points? This ground has 3 testpoints, traces near also have 2 test points. If they are at the same trace, at the same potential, what's their purpose?

  • 1
    \$\begingroup\$ useful search terms : star earth, Kelvin probe, 4 wire probe. \$\endgroup\$
    – user16324
    Jan 7, 2020 at 18:12

1 Answer 1


You are on the right track with your comment about the "same potential." PC board traces have non-zero impedance, but have resistance and (more importantly in this case) inductance. This means that when you are running a load with potentially high, switched currents, the instantaneous potential is not the same at all points along the current path.

Your board has established multiple current paths; some are low-current and these will be pretty close to the same potential because there is little voltage drop associated with the current. The ground for the circuitry on this branch will be the more or less the same between devices, making this ground branch useful for a reference for analog signals and for digital thresholds.

The path containing high, switched currents can have relatively large instantaneous voltage drops along its length; one volt or more is not uncommon at the switching edges. Imagine the ground reference on a processor moving one volt with respect to the digital input signal levels and you can see how this would be a problem.

The strategy is therefore to design so that the high-frequency, high amplitude currents are not flowing in the sensitive, low power circuitry. Similarly, to measure voltage on a device, the test point used for ground reference should be local to the device. You can experiment with an oscilloscope, measuring a voltage at one point in the circuit while moving the probe ground to different points and you will see the effect.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.