Anyone who has tried to pick between two closed-loop NEMA 17 steppers knows the spec sheets rarely settle anything, especially once microstepping enters the picture. Proto G Engineering tackles that exact problem with a head-to-head torque comparison of two such motors, measuring holding torque in pound-inches across several microstep settings.
The bench arrangement is well thought out. An ESP32 on a breadboard drives both motors in the test. Motor A is an MKS unit and Motor B comes from BigTreeTech, and both carry driver and magnetic encoder boards mounted on the rear, along with OLED screens and pushbuttons for menu navigation. 3D-printed plastic couplers link the two shafts for the head-to-head rotation runs, and a digital torque wrench held in locking pliers on the bench takes the readings. With a Rigol DP832 supply, multimeters, oscilloscopes and printers around the workspace, this looks like a properly equipped test bench.
The numbers are what make it interesting. At 1/16 microstepping Motor A reads 3.5 lb-in against 4.1 for Motor B, yet at 1/8 and 1/32 Motor B stalls and logs 0.0 lb-in while Motor A still manages 3.4 and 2.9. Each motor was connected individually to the wrench, with results noted in a notebook beside a printed chart of theoretical microstep torque loss percentages, and handwritten notes listing the standard 0.4 to 0.7 N·m range for NEMA 17s sit nearby.
The open question is Motor B. It comes out ahead at default settings, but it repeatedly stalled against the wrench at two of the three settings, and the maker suspects firmware issues. Calling Motor A the more reliable choice across microstepping configurations seems well earned given the data, though a retest of Motor B after that firmware suspicion is addressed would be worth seeing. The maker also offers to send out a link to the ESP32 code, which is a generous touch.
Project by Proto G Engineering. Watch it on their YouTube channel: