Re: [CAD_CAM_EDM_DRO] First note-Help Mariss Help us with shielding???

Tom Eldredge wrote:

> As far as the encoder noise is concerned, what a puzzling thing. We used a
> foil shielded cable for the servo motor drive voltage, and a separate foil
> shielded cable for the encoder signals, and kept them separate all the way.
> We also put a .1 uf cap at the encoder end of the cable between +5 and gnd.
> Cables around 10 feet from drives. We grounded the encoder cable (no
> twisted pairs inside) at the drive end, and left the motor shield
> disconnected. The drives faulted constantly. We cut the ground wire
> connecting to the shield on the encoder cable, and it ran with only an
> occasional fault. We grounded the servo motor drive cable and the faults
> stopped. Never could see much on the scope in the way of noise. Before we
> found success, we also built a little amplifier and put it right in the
> cavity of the servo motor with the encoder. The amplifier was made with a
> 74HC04 (per Marris's advice). We paralleled three of the inverters for A
> and three for encoder B signal. It still faulted every time until we
> shielded the servo motor power cable. Jon Elson has pointed out to us that
> this foil shielding will not last the life of the cable if the cable flexes
> all the time. So, keep that in mind when you select a cable. If I have
> dragged this out too long to keep up, the bottom line is that the only
> configuration that works at this time is to have the servo motor drive cable
> shielded, and leave the encoder cable unshielded. Don't ask me why. I
> don't know. I would have thought it better to shield the encoder cable. If
> we try to ground the encoder cable sheild, it faults out all the time.

What did you ground the encoder cable to? A ground terminal on the
drive, the case of the drive, or some other point? If you ground the shield
to a ground terminal of the drive, then it is injecting all intercepted
RFI current INTO the drive's electronics! Don't assume, just because the
encoder cable is 'causing' the problem, that the encoder's signal is
being disturbed. It sounds a lot more like the Gecko drive's internal
circuitry is being disturbed. Your shield was conducting interference
into it. A better thing to try might be to ground the shield to the drive's
case.

> Electrical noise has been a real pain for me in many projects I have worked
> on.

Many EE's think grounding, shielding and EMI is black magic, and they are
not far from the truth. Although the laws of physics are fairly simple and
totally predictable, when you have complex systems with noise emitters
and noise-sensitive circuits, the paths unwanted signals will travel to
interfere with others can be very hard to figure out. If current can flow
several different ways from one point to another (a "ground loop") it
can be nearly impossible to determine which path most of the current
will follow, and what the effects will be.

There are EMI consultants who charge $250 - 1000 an hour to fly out,
look over your setup and suggest what to do to make the system work
reliably. Thay couldn't get away with charging that unless they were
good at solving problems most engineers couldn't.

Jon