Re: [CAD_CAM_EDM_DRO] wire question

JanRwl@... wrote:

>In a message dated 5/28/2003 2:32:05 PM Central Standard Time,
>davemucha@... writes:
>
>
>
>>I'm looking at some wire and am wondering if I should get unshielded for
>>the steppers (16 condutor) and shielded home/limit/low votage and then
>>unshielded for the 110VAC for the routers.
>>
>>
>
>Dave: This sounds OK if you "dress" things logically (don't have tye-wraps
>every two inches, pulled so tight the PVC flows). I have HEARD that shields
>should NOT be used to conduct current (as in "return current" from LED's, etc.),
>and they should be "tied to ground" only at ONE end (the
>controller/power-supply end, NOT the "machine end"). Perhaps a good idea, but I have never seen a
>situation where "problems went away" as soon as one end of the shield was
>disconnected! Of course, I probably never had more than a few mA of current
>flowing along the shield, so no matter. Someone should do some "field tests" on
>this point, one day, and "publish" here!!! Jan Rowland
>
>

The logic of this is when you are trying to get the lowest induced noise
in a
piece of equipment, such as audio, telephone or measurement instruments.
However, if you are trying to exclude strong interfering signals, or prevent
radiation from wires in the cable, then a continuous shield from the housing
of box A to the housing of box B often works better. This is one of
those black
arts where there actually are no hard and fast rules, just general
guidelines.
The reason is that the particular path a signal (or the signal's return
current) takes
when there are many possible paths, is very hard to predict in advance.
Inductance
and contact resistance make significant contributions to the
distribution of currents
in these cases. Any case where you have sensitive signals (like encoder
signals) and
powerful currents with fast rise/fall times (like stepper motor
waveforms) in cluse
proximity, it can get complicated.

Jon