Telecommunications performance related to magnetic fields?

Status
Not open for further replies.
B

b.toepper

Member
Location
Rio Rancho, NM
I have a client who runs Cat5e throughout their factory. They have a general "guideline" that tells us to design the layout of this cabling "...not near large magnetic fields ... such as transformers and fluorescent lighting panels."

Ok, I get what they are driving at and, in general, we typically design our cabling away from large transformers. Now that I've found this requirement (and being of a curious nature) I set out to find real measured values of magnetic fields and their effect on performance of data communications. EIA/TIA do have a threshold of 3 V/m, but that is a measure of E field, not magnetic. Magnetic fields are measured in Gauss (or Wb/m2).

Do any of you know if someone has actually tested these situations? For example, taken a length of Cat5e, placed it in a magnetic field of X milliGauss, and tried to obtain gigabit performance? I did try looking in IEEE for white papers on the topic with no avail. Maybe I did not look hard enough?

I'm just looking for a gentle nudge in the right direction.

Thanks in advance.
 
b.toepper said:
I have a client who runs Cat5e throughout their factory. They have a general "guideline" that tells us to design the layout of this cabling "...not near large magnetic fields ... such as transformers and fluorescent lighting panels."

Ok, I get what they are driving at and, in general, we typically design our cabling away from large transformers. Now that I've found this requirement (and being of a curious nature) I set out to find real measured values of magnetic fields and their effect on performance of data communications. EIA/TIA do have a threshold of 3 V/m, but that is a measure of E field, not magnetic. Magnetic fields are measured in Gauss (or Wb/m2).

Do any of you know if someone has actually tested these situations? For example, taken a length of Cat5e, placed it in a magnetic field of X milliGauss, and tried to obtain gigabit performance? I did try looking in IEEE for white papers on the topic with no avail. Maybe I did not look hard enough?

I'm just looking for a gentle nudge in the right direction.

Thanks in advance.
Click to expand...

I really doubt that you are going to be able to find a #.
As I am sure you know, twisted pair cable is designed to
minimize any such effects since the magnetic fields
theoretically cancel. I worked in the datacomm industry
at a large equipment manufacturer, where we were very
involved with the IEEE standards and other industry
efforts, and am unaware of anyone in the industry who
did such tests. Part of the problem as I am sure you
aware is that any attempt to quantify this would vary
with the quality of the cable -- the standards specify
minimum parameters, but that is really a theoretical
standard that doesn't exist as an actual cable/xcvr/etc.
The guideline you describe makes good sense from
a design point of view, but FWIW, in about half a dozen
years of being involved at a manufacturer that had
hundreds of manufacturing customers, I never heard
of any incidents of the equipment not working because
of large magnetic fields -- this may be because the
problems were solved in the field by common sense
measures such as the guidelines you describe, or it
may be because it's a non-issue, I suspect it's more
of the latter, but I can't prove it to you. If there were
such a failure, it would likely be "catastrophic", i.e.
your data transmission would approach zero, not some
major fraction of a gigabit, because while there is some
redundancy built into the data transmission protocols,
they basically assume a reasonably clean channel. If
you haven't found it by googling around IEEE and
EIA/TIA, I suspect it isn't out there.
 
I suspected as much, but I guess I wanted to hear it from another source. Thank you for your input, it is much appreciated. I suspect that as higher speeds are figured and when industrial / commercial environments become more and more dense, this will become an issue to be considered.

Or maybe the price of fiber (and fiber equipment) will come down to the point that this becomes a non-issue. :smile:

Again, thanks for your input.

To anyone else, if you know of a study or white paper, please let me know!
 
Our group performed detailed magnetic field testing for a major player(has 3 letters in their name) in this field.
Their data cables were running parallel to 3000 amp single conductor power cabling.
Their "shielded" data cables "picked up" similar induced levels compared to unshielded cables.
Sorry,study is confidential however,I can say they spent over $200k.relocating the data cables.:smile:
 
081204-1941 EST

b.toepper:

Run your own experiment.

Get a length of unshielded CAT-5 cable. Maybe 25 to 50 feet. At one end short all the wires together. At the other end pick one twisted pair and connect a Fluke 27 meter in AC MV across this pair.

Go to an open field and measure the induced voltage. My guess is 0.0 MV. A Fluke 87, true RMS, has a somewhat higher zero reading because of internal noise in the meter.

Now take this test cable and meter near whatever object you want to test for magnetic field effects.

I have run communication tests with unshield CAT-5 and had to get within about 4 ft of a MIG welder's welding cable to cause errors. This is probably mostly magnetic coupling.

Note: after about 25 to 40 kHz a Fluke 27 or 87 reads funny and by 100 kHz drops out. An older Simpson 260 is good into several hundred kHz.

I have started to setup an experiment with an automotive ignition coil and spark gap to test this as a noise source. But this is not a high priority.

.
 
gar said:
I have started to setup an experiment with an automotive ignition coil and spark gap to test this as a noise source. But this is not a high priority.
Click to expand...
So, you're not going to run networking in your truck until next year? :D
 
081211-0951 EST

Larry:

It is just another noise source to provide an additional means to evaluate noise susceptibility. The major coupling would be capacitive and RF.

The spark gap noise generator may be more impressive to show someone than when I put 1000 V RMS 60 Hz voltage on the interconnect cable between my RS232 isolators.

But I do like your joke.

.
 
gar said:
081211-0951 EST

Larry:

It is just another noise source to provide an additional means to evaluate noise susceptibility. The major coupling would be capacitive and RF.

The spark gap noise generator may be more impressive to show someone than when I put 1000 V RMS 60 Hz voltage on the interconnect cable between my RS232 isolators.

But I do like your joke.

.
Click to expand...

Gar,
For a more calibrated test you might try something like this:

http://www.thermo.com/com/cda/products/product_application_details/1,1063,11367,00.html


Or if you perfer old school type spark gaps the old " Nema Shower Arc" was always fun.
 
Status
Not open for further replies.
Top