Kelvin Connection Errors

[dracotonisamond]

Recently i had a prototype board manufactured and it came with a manufacturing error in one of the kelvin connections where the wire is fused to the pad early.
what sort of errors in measurement should i expect with this defect or will it be trivial?

im trying to figure out if its worth assembling or not because these chips are not easy to come by anymore.

[ejeffrey]

That should make no difference.  The entire back edge of the pad should be nearly equipotential and carrying little current.

[ConKbot]

https://www.analog.com/en/analog-dialogue/articles/optimize-high-current-sensing-accuracy.html

While there is no direct mirror of your situation there, you can see that the error can be quite large, IF you are using an extremely low value current shunt like in the paper there. (They are using a 500uohm resistor)

Since you neck the trace out to the shunt resistor asymmetrically on side that wasn't etched properly, if you are running a very low value shunt, there could be "some" error.  If the voltage dropped over the shunt swamps any field gradients in width of the trace, then error could be negligible.

If it were a straight shot in and out of the resistor pad, with the trace the same width, I'd agree that the back edge should be equipotential, but with the practicalities of an actual layout, it's solid into "it depends" territory.

[dracotonisamond]

the pad in question is a 5mohm low side shunt for a battery fuel gauge.
the defect side is to the negative terminal of the battery pack.

i believe this gauge(MAX17320) should correct for the mismatch since the defect will affect current in both directions.
and the sense resistor's terminal doesnt sit on the edge of the exposed pad.

ill probably just go ahead and assemble the prototype board.

[Kleinstein]

The layout on the current side for the shunt looky like it would distribute the current resonable well. So there should not be very much voltage over the length of the pads.  How much would also depend on the details of the shunt and how evenly the solder is.
5 mOhms is also not yet super low in resistance.
The error that would happen is of the type of adding a small part (likely in the lower µohms range) of the copper (and solder) resistance to the shunt. So the the type of error would be a slightly more positive TC of the shunt resistance. It would effect both directions in the same way, so no problem there. The current terminal to the right is long enough so that the part outside the picture can no do much in disturbing the balance.