Electrical code Compliance - EMI/EMC probe for spectrum analyzer

Author

EMI/EMC probe for spectrum analyzer

blanko

2008-02-13, 8:25 pm

We are having some EMI/EMC problems and I want to rent a spectrum
analyzer to see if changes that I am making are helping or not. I
want to be able to sniff around the board and see the problem areas.

What type of antenna or probe should I use for this? Will I be able
to rent this?

Any other things that I should be considering?

Thanks.

-Mike.

hrhofmann@att.net

2008-02-14, 3:25 am

On Feb 13, 8:08=A0pm, blanko <electro...@yahoo.com> wrote:
> We are having some EMI/EMC problems and I want to rent a spectrum
> analyzer =A0to see if changes that I am making are helping or not. =A0I
> want to be able to sniff around the board and see the problem areas.
>
> What type of antenna or probe should I use for this? =A0Will I be able
> to rent this?
>
> Any other things that I should be considering?
>
> Thanks.
>
> -Mike.

Are the problems radiated or conducted? What frequency range? What
kind of problems?

You will need a Line Impedance Stabilization Network if you are
looking at noise coming out onto the AC power lines, and an assortment
of calibrated antennas if you are trying to meet the FCC of European
Community EMC requirements.

If you are just wanting to see if the noise in the vicinity of an IC
is lower or higher than before, you can use an oscilloscope probe with
a little plate/piece of aluminum foil on the end just to see what
signals are being radiated, but that is only going to tell you if the
noise in that part of the circuit is changing, and not what is the
overall emission from the entire device. Depending on what is going
on, pushing the noise down in one part of a circuit can result in
increasing the noise in another part of the same device.

More info would make it possible to give you a better answer.

A local EMC test house might actually be cheaper to use if you only
need a couple of hours of help making measurements. Where are you
located? I might be able to recommend a near-by reputable EMC test
house. There are IEEE EMC Society chapters in many cities in the USA
and in several non-USA countries. Those folks should also be able to
give you advice on local test houses that are reputable.

H. R.(Bob) Hofmann, past President IEEE EMC Society

Robert Lacoste

2008-02-14, 3:25 am

"blanko" <electrone2@yahoo.com> a écrit dans le message de news:
c51e4ceb-8a9e-46e1-a107-696c57ab4d63@e23g2000prf.googlegroups.com...
>
> We are having some EMI/EMC problems and I want to rent a spectrum
> analyzer to see if changes that I am making are helping or not. I
> want to be able to sniff around the board and see the problem areas.
>
> What type of antenna or probe should I use for this? Will I be able
> to rent this?

If you are looking only for relative / short distance measurements then you
best buy (or rent) would be a pair of H-field and E-field active probes. We
are using HZ-530 kit from Hameg with great pleasure. See
http://www.hameg.com/210.0.html. If your budget is lower then you can also
design your probes yourself, but their frequency response may be a little
more difficult to manage. An E-field probe can be a very short antenna wire
and/or conductive square, it will mainly be used quite far from the product
(say 30cm to 2m depending on frequency) to see relative improvements of the
radiaded. A passive H-field probe can be built with a coaxial cable (build a
small loop with the shielded cable, one or two turns, 1 or 2cm wide, and
solder the inner conductor of the cable end to the shield of the unturned
section - but not the shield to the shield, it would be easier with a
drawing byut I'm sure you will find one on the web), and will be very
usefull to find the sources of the radiation as the H field diminushes very
quickly with distance).

Friendly yours,
Robert Lacoste
www.alciom.com
The mixed signal experts

hrhofmann@att.net

2008-02-14, 9:25 am

On Feb 14, 2:01=A0am, "Robert Lacoste" <use-contact-at-www-alciom-com-
for-email> wrote:
> "blanko" <electro...@yahoo.com> a =E9crit dans le message de news:
> c51e4ceb-8a9e-46e1-a107-696c57ab4...@e23g2000prf.googlegroups.com...
>
>
>
>
>
> If you are looking only for relative / short distance measurements then yo=
u
> best buy (or rent) would be a pair of H-field and E-field active probes. W=
e
> are using HZ-530 kit from Hameg with great pleasure. Seehttp://www.hameg.c=
om/210.0.html. If your budget is lower then you can also
> design your probes yourself, but their frequency response may be a little
> more difficult to manage. An E-field probe can be a very short antenna wir=
e
> and/or conductive square, it will mainly be used quite far from the produc=
t
> (say 30cm to 2m depending on frequency) to see relative improvements of th=
e
> radiaded. A passive H-field probe can be built with a coaxial cable (build=
a
> small loop with the shielded cable, one or two turns, 1 or 2cm wide, and
> solder the inner conductor of the cable end to the shield of the unturned
> section - but not the shield to the shield, it would be easier with a
> drawing byut I'm sure you will find one on the web), and will be very
> usefull to find the sources of the radiation as the H field diminushes ver=
y
> quickly with distance).
>
> Friendly yours,
> Robert Lacostewww.alciom.com
> The mixed signal experts

Robert gives you good advice.

blanko

2008-02-14, 1:25 pm

Bob - thank you for your comments. Please see my response below:

On Feb 13, 9:01 pm, "hrhofm...@att.net" <hrhofm...@att.net> wrote:
>
> Are the problems radiated or conducted? What frequency range? What
> kind of problems?
>

This is all new to me so please bare with me! Running directly from
batteries with no sensors/cables/power supplies connected we see a
spike at about 820 MHz above
the class B limit. The full range being scanned is 30MHz - 1 GHz. In
this case I believe this is radiated. No when the cables are attached
everything goes crazy. For this case I am not sure if it is radiated
or conducted. What is the best way to determine this?

> You will need a Line Impedance Stabilization Network if you are
> looking at noise coming out onto the AC power lines, and an assortment
> of calibrated antennas if you are trying to meet the FCC of European
> Community EMC requirements.
>

Okay thanks. This is a little over my head.

> If you are just wanting to see if the noise in the vicinity of an IC
> is lower or higher than before, you can use an oscilloscope probe with
> a little plate/piece of aluminum foil on the end just to see what
> signals are being radiated, but that is only going to tell you if the
> noise in that part of the circuit is changing, and not what is the
> overall emission from the entire device. Depending on what is going
> on, pushing the noise down in one part of a circuit can result in
> increasing the noise in another part of the same device.
>
> More info would make it possible to give you a better answer.
>

I'm going to stay with a probe I can purchase. Wiring something up
myself seems error prone!

> A local EMC test house might actually be cheaper to use if you only
> need a couple of hours of help making measurements. Where are you
> located? I might be able to recommend a near-by reputable EMC test
> house. There are IEEE EMC Society chapters in many cities in the USA
> and in several non-USA countries. Those folks should also be able to
> give you advice on local test houses that are reputable.
>
> H. R.(Bob) Hofmann, past President IEEE EMC Society

I am located in San Diego. Any contacts you could give me I'd
appreciate it!

Thank you for your thoughts!

-Mike

blanko

2008-02-14, 1:25 pm

On Feb 14, 12:01 am, "Robert Lacoste" <use-contact-at-www-alciom-com-
for-email> wrote:
> If you are looking only for relative / short distance measurements then you
> best buy (or rent) would be a pair of H-field and E-field active probes. We
> are using HZ-530 kit from Hameg with great pleasure. Seehttp://www.hameg.com/210.0.html. If your budget is lower then you can also
> design your probes yourself, but their frequency response may be a little
> more difficult to manage.
>

Robert - thanks for the information. What about a "regular" probe that
will allow me to take measurements a few meters away as is done in the
testing house? What type are this? Directional? Any thing special to
look for?

Thanks!

-Mike

hrhofmann@att.net

2008-02-15, 3:25 am

On Feb 14, 11:31=A0am, blanko <electro...@yahoo.com> wrote:
> On Feb 14, 12:01 am, "Robert Lacoste" <use-contact-at-www-alciom-com-
>
> for-email> wrote:
you[color=darkred]
We[color=darkred]
..com/210.0.html. If your budget is lower then you can also[color=darkred]
e[color=darkred]
>
> Robert - thanks for the information. What about a "regular" probe that
> will allow me to take measurements a few meters away as is done in the
> testing house? What type are this? Directional? Any thing special to
> look for?
>
> Thanks!
>
> -Mike

Mike:

I suggest you contact:

Mark Frankfurth
Cymer Inc.
Manager of Compliance Engineering
17075 Thornmint Ct.
San Diego, CA 92127
Phone: +1 858 385 6558

He is the chairman of the San Diego chapter of the IEEE EMC Society
and may be able to help you if you call him and give him a better
explanation than you can do via e-mails on this site. Please be sure
to tell him that I suggested you call him.

Bob Hofmann

Robert Lacoste

2008-02-15, 9:25 am

"blanko" <electrone2@yahoo.com> a écrit dans le message de news:
a53d2522-8b72-4e91-b7d4-551f9e9a997e@e6g2000prf.googlegroups.com...
> On Feb 14, 12:01 am, "Robert Lacoste" <use-contact-at-www-alciom-com-
> for-email> wrote:
>
> Robert - thanks for the information. What about a "regular" probe that
> will allow me to take measurements a few meters away as is done in the
> testing house? What type are this? Directional? Any thing special to
> look for?

Well, it's called an antenna... The only difficulty is that antennas for EMC
measurements must be very wideband and calibrated. Usual ones are
log-periodic antennas in the lower frequencies and bridged horns in
microwaves. However if you are only doing relative measurements then a
non-calibrated and not-so-wideband antenna can be used. For example if you
are only looking at a specific spurious radiation then you can use a simple
diploe antenna tuned for this specific frequency, as long as you just want
to compare two versios of your product at this frequnecy only.

Robert

Wimpie

2008-02-16, 5:25 pm

Hello Mike,

You spoke about an 820 MHz peak. Is it a real narrow band peak (for
example just one peak with the RBW setting of the analyzer at some
kHz), or is that peak a summation of several spectral components close
together? Are there sub-harmonic components, etc? Is it close to a
band of interest, or far outside?

The reason for such questions is that sometimes from the knowledge of
your design, in combination with the signals on the analyzer, you can
determine the source of the unintended radiation (radiated or
conducted).

When you expect more EMC challenges (and have some time to study), you
could buy an EMC book with a practical approach (such as "EMC for
product designers" by Tim XXXXXXXX).

Just for relative measurements, probe construction is not that
difficult. Several books on EMC describe how to do it. Not everything
you buy is worth the money (from own experience).

A very primitive alternative for the spectrum analyzer (only when you
know where the level is to high) is to use a so called "wide band
receiver". These, mostly hand-held, receivers are used by radio
hobbyists and radio amateurs. Many modern designs go to 2 GHz and
beyond and have an signal strength indicator. In that case you can see
whether your circuit change does reduce the unintended radiation.

With regards to measurements some meters away from the product, at UHF
you will have many reflections that will influence your measurements
with far over 6 dB. It will be difficult to maintain the same RF
environment (even your own movements will give different readings). To
avoid reflections, test ranges are outside or in anechoic chambers.
For comparative measurements, nearby measurements may give better
results.

I hope this contributes a bit to the solution of your EMC problem.

Best regards,

Wim
PA3DJS
www.tetech.nl (Dutch).

blanko

2008-02-16, 8:25 pm

Thanks for the great posts everyone! I am here in my lab adding
shielding to my plastics. Seems like things are getting better!

-Mike

LinkBot