Skipping steps, avoiding resonance and the importance of being grounded

At 10:35 PM 12/1/02 -0800, I wrote:

>I finally got the thing running this weekend, and it is not running
>reliably.

Chris L and Jerry J, and Tim G replied with helpful posts about double
checking each axis on its own, systematically varying one parameter at a
time and starting slow and inching up in speed and acceleration, etc.
(Thank you both for your thoughtful posts!)

A first problem I realized, was that I still had "rapid" speed set at
300ipm from when I had been playing with the limits of speed. Sometimes it
was working OK, but some G0's did not work as advertised! :-) I set G0 to
60ipm for X and Y, and 30ipm for Z (I don't have counterbalance springs
installed on the Z carriage yet, and the Sherline milling head plus adapter
plate weigh 12 pounds, so I'm taking it easy on Z for the time being.)

A clue for me came when I was "air cutting"--no spindle motor running, no
tool, no noisy shopvac sucking chips. I was running through the 3D gcode
file and watching the display while the gantry was running. On a move that
was purely X, the Y and Z steppers would take random, varying steps (or
several steps at a time) when they should have been motionless.

I took Tim's suggestion to try another software program. I had played with
TurboCNC in the past, so I downloaded a fresh copy and set it up for the
gantry. It was an eye-opener! X axis would jog, but Y and Z would just
vibrate the motors but not move the axes. I played with the pulse width
along with speeds and accelerations, but could only get one axis to move
with TurboCNC.

I began to think that my problem was electrical. I opened the case of the
controller and noticed several things I had not before. The first one was
that the signal grounds to the three single-axis controller cards were
daisy-chained. I removed them and brought all three grounds to a common
point on the chassis. That solved the problem maybe 75%. But the
"inactive" axes were still taking random steps. I also noticed that the
wires from the DB-25 connector to the step and direction inputs to the
cards were long, thin and maybe not-too-well soldered.

Caudlet posted in another thread at that time:

>One of the areas that gets overlooked is the enemic drive from the
>parallel port of most PC's I highly recommend a buffer card for

I am already using a good Belkin 6' double-shielded parallel cable, so
after some Web reading and searching I ordered an Aten CB-140 parallel port
booster/buffer from Computer Gate:

http://www.computergate.com/products/item.cfm?prodcd=WDECB140

It was the best price I found, $24.95 plus tax and shipping (includes the
wall wart power supply)

While I was waiting for the booster to arrive, I gutted the wiring in the
controller case. I relocated the DB-25 to the side of the case near the
input terminals for the driver cards, and grounded everything to a single
stud in the case. The CB-140 came yesterday, and I tried out the system
this evening. Rock-solid!

Wiring problems aside, I can only assume that the Dragon Driver cards are
more picky about the TTL level of their input signals than Dan Mauch's 2A
driver board, which has always worked reliably from the start.

I am happy to have both the Sherline mill and the new gantry up and running
reliably.

Thanks again for the helpful suggestions, guys!

Best regards,

Randy

Randy Gordon-Gilmore ,----.___________ ______________ _________________
ProtoTrains // = = === == || == == == = || == == == = == =|
Rio Vista, CA, USA /-O==O------------o==o------------o==o-----------o==o-'
zephyrus@... http://www.prototrains.com