Hi Joe.
You asked for an explanation as to why the difference between the near and
far field results. I think the replies so far have probably answered that
question. I've tried here to give some help with the real problem of
solving the excess emissions.
From my own experience and discussions with colleagues, I've found you
definitely need to do some (if not most) of the trouble-shooting while at
the test site. Finding a problem then just returning to the lab to solve it
usually leaves you with a lot of questions unanswered. That might not be
much help this time but perhaps next time?
The following is how I'd go about tackling the problem. I'm curious to see
if there's anyone in the group who disagrees with my approach.
As with any EMC problem, you've got to consider the source, the
transmission medium and the victim. Obviously there's nothing you can
change about the victim (the test antenna) but you should be able to narrow
it down to work out the real source, and the means by which it is being
radiated.
For clues to the problem's cause to begin with I usually ask:
(1) For the problem frequency, what's the most likely source?
(2) For the problem frequency, what's the most likely source antenna? At
400MHz the wavelength is a bit under 1m (3x10^8 / 400x10^6 = 75cm) so any
short cables (or at this frequency, maybe even long PCB track - like
back-plane tracks?) that might make nice 1/2 wavelength or 1/4 wavelength
dipole antennas would be the first I'd check out. Could also be a slot
antenna effect in your enclosure - any seams or gaps in the box that are in
this ball-park?
Usually I'd try isolating the source by either disconnecting cables,
turning off or unplugging cards, attenuating cable emissions with copious
amounts of ferrite clamps etc and get the test engineer to do a spot check
at the problem frequency as I tried eliminating each suspect. This is where
the buckets of ferrite cable clamps, rolls of aluminium foil, shielding
mesh and earthing straps come in to play. Here's where that near-field
probe might come in handy too. This kind of troubleshooting though often
requires a fairly intimate understanding of the way the equipment under
test works so you can be confident about your assumptions and the
conclusions you draw from the observed results. If the design engineer
isn't actually at the test site, she/he should at least be accessible by
phone to discuss the problems and make suggestions as to what to try.
Using this technique, you can usually narrow it down fairly quickly to the
source and antenna. If there's time, and its practical then I'd try some
quick modifications to the problem circuit that's the source of the noise
in order to get some reference of what changes cause what kind of reduction
in the emission levels.
Quite often though, you have to be aware that a change may solve the
emission problem at the frequency you're working on, but result in the
energy appearing elsewhere in the radiated spectrum causing the equipment
to exceed the limit at some other frequency, especially if you've just
modified the source antenna and not the signal causing the emission. Also
note - although its difficult when you're rushing to get the problem fixed,
it pays to make good records of what you change and what the results are -
can help a lot later on.
If you have the time at site to try a few different options (that are
repeatable later), and get the highest 3 or 4 emission levels for each
option at site, then if you can't find a solution you're happy with at the
test site, it gives you a reference to work with back in the lab.
For example, say that you found that:
Design Change #1 resulted in 6dB reduction in the emission at 400MHz with
other peaks (below the pass/fail limit) at 200MHz (3dB under), and
800MHz(8dB under)
Design Change #2 resulted in 20dB reduction in the emission at 400MHz but
caused the peak at 200MHz that went over the limit by 6dB with the peak at
800MHz reducing to 10dB under the limit.
Design Change #3 resulted in 3dB reduction in the emission at 400MHz with
other peaks (below the pass/fail limit) at 200MHz (9dB under), and
800MHz(7dB under)
Then when you get back to the lab to try and find a good permanent fix, by
repeating the changes you made at site, and comparing the emissions levels
you observe for each in the lab with the results at the test site, you can
get a reasonable feel to tell if your measurements in the lab are going to
be indicative of what you'll see at site.
If you have access to a spectrum analyser and an antenna that covers the
frequency range you're interested in you can probably get a better feel for
the effect of your changes than when using a near-field probe.
If the emissions are high enough to fail the test, there's a good chance
you can get a reasonable indication of the result of changes by comparing
the emission results measured in someone's back yard(know anyone who owns a
farm/ranch?) or perhaps the roof at work, to those measured at the test site.
Use the antenna and analyser to measure the peaks with your equipment under
test switched off first so you have an idea of what the ambient peaks are,
then measure the emissions from the unmodified equipment, set up as close
as practical to how it was at site. Although far from ideal, if you can see
a decent correlation between what you measure and what was measured at site
for the same equipment under test, then I'd have a lot more confidence in
the results of this kind of measurement than for those taken in the near
field. Although the results would only be ball-park, that's often all you
need anyway. Given the expense of official OATS testing (unless your
company happens to won its own site), I'd want to be pretty confident of a
fix before booking in again.
Hope that's of some help.
Regards,
Peter Poulos
-------------------------------------------------------------------------
Please note: The views, opinions and information expressed and/or
contained herein do not necessarily reflect the opinions or views of
Foxboro, the organisation/s through which this communication was transmitted
nor any other third party, unless explicitly stated so.
Peter Poulos (Hardware Design Engineer)
Foxboro Australia
At 03:07 AM 15/09/2000, marti...@appliedbiosystems.com wrote:
I am having a difficult time answering the following question for a
non-technical person. Hopefully, someone can put the answer into a language
that a non-technical person can understand.
We have a 400 MHz clock and are failing radiated emissions at 10 meters by
10 dB
at 400 MHz. We bring the product back to our lab and start making
modifications
on the clock circuit and taking measurements with a near field probe. With
these modifications and measuring with a near field probe, we realize a 10 dB
reduction in emissions at 400 MHz. Why would we not see the same
reduction when
taking the product back to a 10 meter site?
Your help is appreciated.
Regards
Joe Martin
marti...@appliedbiosystems.com
-------------------------------------------
This message is from the IEEE EMC Society Product Safety
Technical Committee emc-pstc discussion list.
To cancel your subscription, send mail to:
majord...@ieee.org
with the single line:
unsubscribe emc-pstc
For help, send mail to the list administrators:
Jim Bacher: jim_bac...@mail.monarch.com
Michael Garretson: pstc_ad...@garretson.org
For policy questions, send mail to:
Richard Nute: ri...@ieee.org
-------------------------------------------
This message is from the IEEE EMC Society Product Safety
Technical Committee emc-pstc discussion list.
To cancel your subscription, send mail to:
majord...@ieee.org
with the single line:
unsubscribe emc-pstc
For help, send mail to the list administrators:
Jim Bacher: jim_bac...@mail.monarch.com
Michael Garretson: pstc_ad...@garretson.org
For policy questions, send mail to:
Richard Nute: ri...@ieee.org
