Treadmill motors and interference

I just got another axis converted to the Geckodrive step/direction servo
drive for a total of 2 axis's on servos and Z still on a stepper. I have one
of the Surplus Center variable speed treadmill motors driving the mill
portion of my machine (a Shoptask 3 in 1). For the moment the connection for
the step and direction signals is from a parallel cable to a connector on
the driver. Inside the box for the driver the connector is hooked to a
junction board and then I have separate wires with ground, step, and
direction for each axis. Currently the Z axis is still going to the board,
but the other 2 axis's are not pulled out through the lid of the box and
connected to shielded cable that takes the step and direction signal to the
servo drives.

The problem I am now having is when I turn on the treadmill motor the axis's
all start jumping around. The amount and direction is related to the speed
and load on the treadmill motor. I first went and hooked the drain wires
from the shield on the encoder leads to ground and then also hooked the
drains of the wiring heading to the step and direction signal in the stepper
driver to ground. Did not seem to make any difference. The only piece of
wire that is not shielded is a +5 coming from the computer and acting as the
common for the step and direction signals to the servos. The real odd part
is I also am getting this random motion on the Z axis (the one that is still
hooked to the stepper). Makes me think the problem is not the encoder signal
getting scrambled, but signals being fed into the step and direction line.
What I can't follow is hoe the interference is getting to the stepper line
as it is not using the +5v unshielded line as a common, but is using a
ground in the shielded parallel cable.

I have tried putting a 1 uf cap in the treadmill motor power leads and it
seems to stop the problem. Looking at the design of the speed control there
is no caps downstream of the rectifier, so the motor is getting very dirty
DC from what I can figure. I am surmising that the addition of the little
cap is just enough to keep the ripple from creating enough interference to
cause ghost signals to appear on the step and direction lines.

I guess the question is, should I accept the cap as being a realistic and
reliable solution?

Tim
[Denver, CO]