Axii deciding to do a little jiggling all by themselves? What might be the problem is noise pickup on the step control line.

A simple lowpass filter on the step input will make them rock-solid.

I had to do this to all three of my SLA7042 drivers. You can see the capacitor going from the step input of the chipset to ground, and the step input trace has been cut and a SMT resistor soldered in.

Driver boards based on the A3977 chipset can have the same problem. This board was doing a dance when the switching supply feeding the Z-axis was turned on. A 10pf capacitor soldered from the step input to Vcc (5V) combined with a 1K resistor inline with the step input made it much more solid.

The 'C' of a 'RC' filter is generally connected to ground and not the 5V line, but in this case, step and 5V are adjacent, and a 10pf capacitor with 0.1" lead spacing was the easiest to solder in there. For an A/C signal (like noise) ground and Vcc are equivalent.

After testing the above board, and finding 90% of the noise problem had been eliminated, I decided to calculate the proper approximate values of R and C. 10pf is off by a factor of...100.

This is another A3977-based board. Without an input filter, it was unusable, at least when my switching power supply was turned on.. I soldered in a 1000pf (better known as 0.001uF) capacitor with a 1K resistor. No more weirdness!

The cutoff frequency of a RC lowpass filter is given by
f = 1 / (2*pi*R*C)

f = hertz, pi = 3.14, R = ohms, C = farads

If you plug in 1K resistor and 0.001 x 1E-6 capacitor into the above formula you get a cutoff frequency of 159kHz.