On 01/10/2016 12:44 PM, John Thornton wrote:
> Well none of the suggestions about star grounding did any good. In
> fact they made the problem worse than before, so bad I could not even
> get 2.6 to run.
That may suggest a couple of things being wrong simultaneously:
1) you may still have loops in the 0VDC which are crossing power
domains. These /are/ hard to detect with many devices connected together;
2) there are impedance mismatches in the signal wiring (this one often
works together with 1);
3) shields may pick up noise they should reject.
The question is how to reduce the practical problems, not how to
eliminate all theoretical problems. That is the hard part.
> Off this list I was sent the AB servo best wiring practices pdf
> which explained how to reduce noise in a servo drive enclosure. I
> followed the advise as much as possible and after removing all my ground
> wire antennas I was back running with an occasional sserial error. The
> short explanation is to terminate the shields to a ground plane as close
> as possible to the point where you take the shield/drain wire from the
> cable. I still have a cable bundle running to my switch box that is
> parallel wires and I might replace that with some twisted pairs. Photos
> at 11.
You refer to this document?
http://literature.rockwellautomation.com/idc/groups/literature/documents/at/motion-at004_-en-p.pdf
That doc explains very well how to separate into different power zones.
Also, it makes a good case of shielding and how to do it properly. BTW,
note that it uses protective ground as shield potential and _not_ 0VDC.
That is also why the argument of two-sided termination of the shield is
appropriate (see page 21).
But all this does not save you from impedance problems. The sserial
errors are interesting because it is probably something you actually can
measure with an oscilloscope.
You should be able to see the signal integrity and whether you have
reflections of the signal at either end of the cable.
The sserial connection has a great potential for a loop. See attached
image for illustration. The image has a shielded connection between two
devices, which is good. However, the 0VDC connection (aka GND), which is
embedded in the shielded connection is shorted outside the cable via a
protective ground connection through the PSUs. This means that part of
the signal return may circumvent the constraints of the signal cable and
that creates an unbalanced path. Effectively, it means that the shield
is now part of the cable's impedance, and that is guaranteed to be off
by a factor.
You can check for this scenario by disconnecting the sserial connector
at one side and measure 0VDC vs protective ground on the connector on
the other side (and the other way around). You are in trouble when
measurements on both sides indicate a connection between 0VDC and
protective ground.
If you only have one PSU for both devices, then you are potentially also
in trouble. The PSU will then act as the loop facilitator (through two
distinct 0VDC connections). However, the effects are more subtle and
depend on coupling effects.
The rule for a shielded cable is that all energy must be contained
within the cable to be effective. In other words, the sum of all
currents in the cable's wires must be zero. Bypassing some of it may
cause impedance and therefore signal integrity problems.
--
Greetings Bertho
(disclaimers are disclaimed)
------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
