CAD CAM EDM DRO - Yahoo Group Archive

Re: Shielding

Posted by reed_irion
on 2002-05-31 22:22:30 UTC
Seeing as how I spent several years employed as an EMI/RFI engineer I
cant help but wonder a few things...

I have seen several posts saying the shield should be tied to the
chassis of the driver and left unconnected at the motor... I cant
help but think this would make a marvelous antenna.

I would THINK (yes I may be wrong) something like as follows: You
would want to make sure the ground of your driver board is WELL
connected to your chassis. You probably need to make your driver
board with a "ground plane." Even though the FET's (or BJT's or
tubes or whatever) are switching in the kilohertz or less region, the
fast rising edges and amounts of current involved probably create a
lot of RF energy at harmonics of the switching frequency. The top
frequency of the harmonics depends on a lot of things, one of which
is a resonant structure in the system, such as an ungrounded cable.
I have seen harmonics from a power supply switcher (400 kHz) show up
at 400 kHz intervals out to 300 MHz+. Bear in mind also that it only
takes picoamperes of current loose on shield to fail the worldwide
CSIPR22 standard.

So I digress... I would THINK (yes I may be wrong) that the Drivers
need a ground plane and need to be well grounded to the chassis with
a plated thru hole and a conductive standoff. If possible, you would
want the shield of the cable to be grounded as close to the drivers
as is possible. I would think you would want the other end of the
shield connected to the case of the motor. If you still had a
problem, passing the shield through a ferrite core (of appropriate
composition) or even wrapping several turns of the shielded cable
around a ferrite donut might help.

The idea is that you want to present a ground as close a possible to
the source of the radiation. The smaller the area of the the
radiation, the less EMI.

How do you know if you have an EMI problem ? Spurious beahvior of
your system. Maybe extra steps... The old lady next door
complaining that she cant see Oprah every time you turn on that dang
router thingy... An agent from the FCC at your door...

Otherwise dont worry about it.



--- In CAD_CAM_EDM_DRO@y..., JanRwl@A... wrote:
> In a message dated 31-May-02 17:35:02 Central Daylight Time,
js3mc@a...
> writes:
>
>
> > There has been discussin on the shielding of the motor cables but
I'm
> > wondering what else should/needs to be shielded as well AND why?
>
>
> Jay: The currents flowing in wires to stepper-motors are "chopped"
and
> generate HUGE quantities of electrical nasties called "EMI".
Actually, in
> well-done assemblies, were it not for the electrical noise in the
motor-drive
> conductors, other than just plain good wiring-layout, no other
shielding
> should be necessary.
>
> However, it is probably NOT a bad idea to used shielded cable for
the
> printer-port-cable as well, and at least dress the wires to the
STEP and
> DIRection inputs on your motor-drives next to the chassis-metal,
and do NOT
> run these "logic-level" input-wiring adjacent/parallel to any power-
supply
> and motor-cables. Any "ribbon-cable" inside the box containing
your drives
> and/or motor-power-supply should be laid FLAT on the chassis-metal,
and
> covering that will adhesive-backed aluminum foil wouldn't be a bad
idea, if
> you... Nevermind.
>
> You see, the currents in "logic input lines" are very small,
particularly
> when CMOS logic is involved. That's one reason the Gecko's use OPTO-
ISOLATOR
> inputs, as the LED-inputs on those require MUCH more current-in (16
mA) than
> "logic inputs" such as LS or CMOS gate-inputs. However, the
currents in the
> motor-lines are not only AMPERES, but this current is "chopped" at
tens of
> kilohertz, making a "spray" of EMI into the air near these wires.
So, if you
> were to lay these input-conductors parallel and "touching" the
motor-cables,
> often, enough electrical energy from this "spray" of EMI can be
INDUCED into
> the input-conductors so that the drive-inputs "see" these signals
as inputs,
> and things go nuts!
>
> See Education; Experience; Years of Trial-and-Error Breadboarding;
Patience
> with "Where's THAT Coming From???; Large Quantities of Aspirin...
>
> Lotsa luck! Jan Rowland, Old Troll
>
>
> [Non-text portions of this message have been removed]

Discussion Thread

JanRwl@A... 2002-05-31 17:23:03 UTC Re: [CAD_CAM_EDM_DRO] Shielding reed_irion 2002-05-31 22:22:30 UTC Re: Shielding Steve Blackmore 2002-06-01 01:09:37 UTC Re: [CAD_CAM_EDM_DRO] Re: Shielding Alan Rothenbush 2002-06-01 09:32:34 UTC Re: Re: Shielding