Re: I hate limit switches!!!

Gah. I feel your pain. I'm only days behind where you are now.
Noise and grounding are the black arts of electronics. You have to
backtrack to the basics and really put your brain in gear.

First, any voltage variation will create a voltage noise and any
current variation will create a current noise. Unless a potential
antenna is more than a quarter-wavelength away, you're in the "near
field" area like the secondary of a transformer and can interact with
the electric and magnetic fields independently.

Ground returns for signals or drive power should be close to the
outgoing path to avoid creating a big loop antenna for generating an
E- or B-field. Twisted-pair and coaxial work.

Better yet, enclosing these in a shield which returns via a single
path to a single ground point. Naturally, motor cases are a big noise
source and beg to be attached to the other end of the shield path. A
"guard trace" on a PCB or separate shield on a cable should be
connected only at the source end--say, case to braid to grounding
*point*. Multiple ground paths risk creating some nasty ground loops.

Your limit switches as, say, twisted pair, would keep the loop area
small for minimum noise pickup and change the polarity of the incoming
noise with each half-twist. Running a current through your limit
switches of, say, 100mA lowers the effective impedance and makes it
harder for noise to swamp the limit signal. Also, as a current-driven
signal, an E-field has little effect.

An ideal solution is sending power to each motor and putting all your
power drivers right there at each motor, connecting the logic via an
infrared link or something.

Good luck! Wish me luck, too. We'll both need it!

Don

--- In CAD_CAM_EDM_DRO@yahoogroups.com, "andrebarclay" <andre.
barclay@v...> wrote:

> I finally got my limit switches hooked up but not until days of
> messing around with the wiring. It turns out that I have so much
> noise in my machine wiring that I'm surprised I could hear myself
> think. I did a search on this site and others to find a solution,

but

> it seems that everyone has their own ideas about noise and I'm

afraid

> I'll die a natural death before I get around to trying all of them.
> My setup is as follows. I have 3 non-metallic flex conduits (bad
> start). One carries the shielded cable for the 3 motor encoders
> lines, the second carries 3 shielded cables for the power to the
> servos and the third carries shielded cables for the limit switches.
> All cables have a drain wire and a foil shield. The cables that

carry

> the encoder wires have an additional wire "netting" around them. I
> have tried grounding the drain wires from all the cables to the
> machine, the computer case and every other configuration I could

think

> of and, although I get some improvement here and there, nothing

seems

> to eliminate in completely. The cables for the encoders have 8 wires
> in them not including the drain wire. Since I am only using 4 of
> those wires. I just folded back the extra wires and taped them off

on

> both ends to the insulation. Could these wires be picking up noise
> and if so, what is a quick fix rather then replacing them?
>
> Thanks
> Andre Barclay