Every bring-up session on this bench used to start the same way: break the power rail, insert a multimeter in series, and hope the extra 0.3Ω of meter resistance didn't change the board's behavior. It usually didn't. Occasionally it did, and we'd chase a phantom brownout for twenty minutes before realizing the meter itself was the problem.
The design brief was deliberately small
One INA226 current/power monitor IC, one small microcontroller to read it over I2C and push numbers to a display, and a pass-through connector so the monitor sits inline permanently rather than getting clipped on and off. No logging, no Wi-Fi, no app. The whole point was to remove a step from bring-up, not add a new tool to learn.
The INA226 was chosen over the cheaper INA219 for one reason: a lower minimum shunt resistor value option that keeps power dissipation and voltage drop down at higher currents, which matters when the monitor sits permanently inline with boards pulling up to 2A.
What we'd change
The first enclosure was 3D printed in a single piece with the display window sized just slightly too tight. The display's plastic bezel didn't quite clear it, and we ended up filing plastic on a finished part. Version two split the enclosure into a front and back shell with the window sized to the display's actual outer dimension plus a millimeter of clearance, not the active display area.
It's not a product. It's a tool built because we needed it more than three times, which was our informal threshold for 'this is worth a permanent fixture' rather than a one-off jig.