I have no experience with that board, so my comments are pure guesswork.
When one of the axes is experiencing this problem, can you probe the test points for that axis to see whether the voltage value at each test point is as expected, or whether it is significantly different? If you have a defective or possibly intermittent potentiometer, the test voltage will be incorrect, which could possibly lead to improper functioning of the associated axis driver circuitry, such as greatly reducing the motor winding current, which could give you the symptom that you describe.
Do you have a schematic for the board? Can you identify the component with the small heatsink, and determine its function?