radiated emmission

[amnon]

Hi
I am measuring the radiated emission from a Telco box
It has among other few fast Ethernet ports
which deliver their downstream data via SMII links (125Mhz
While no port is connected, the box measured emission is 30dB bellow
the standard radiated emission allowed limit ,at worst harmony (500Mhz
I have connected all the ports with STP long cables
each cable I have connected contributed about 10dB of noise
These cables are STP , how comes they raise the emission?
What can I do to reduce the added emission levels ?
Can anyone suggest the blackmagic capacitor that will solve the above,
or anything else
Again, emission peaks are seen at Nx125Mhz, where worst results are at
500Mhz

any suggestion will be appreciated

AP

 

[Dbowey]

AP posted:
<< I am measuring the radiated emission from a Telco box
It has among other few fast Ethernet ports
which deliver their downstream data via SMII links (125Mhz
While no port is connected, the box measured emission is 30dB bellow
the standard radiated emission allowed limit ,at worst harmony (500Mhz
I have connected all the ports with STP long cables
each cable I have connected contributed about 10dB of noise
These cables are STP , how comes they raise the emission?
What can I do to reduce the added emission levels ?
Can anyone suggest the blackmagic capacitor that will solve the above,
or anything else
Again, emission peaks are seen at Nx125Mhz, where worst results are at
500Mhz
What have you done with the shields? I suggest you ground them at the cabinet
and float the station end. I presume the cables are all properly terminated(?)

Don

 

[Joerg]

Hello Amnon,

Ferrite toroid cores slipped over the cables can help. Something like 43
material from Amidon. They need to be placed right at the box.

Regards, Joerg

 

[Dave VanHorn]

I have connected all the ports with STP long cables

each cable I have connected contributed about 10dB of noise
These cables are STP , how comes they raise the emission?

This symptom is usually caused by bad PCB layout.
The problem is that high frequency current isn't routed directly back to
it's source, and it's taking a more indirect route, passing by the cable
ground points, to get there.

In the end, the solution may be as simple as a tenth of an inch of PCB
track.
However, re-laying the PCB may not be practical at the moment.
Ferrites on the cables may help, but I don't know if you'll pass testing
that way.
Filtered connectors may help, but they are prohibitively expensive.

 

[John Woodgate]

I read in sci.electronics.design that amnon <[email protected]>

I am measuring the radiated emission from a Telco box It has among other
few fast Ethernet ports which deliver their downstream data via SMII
links (125Mhz While no port is connected, the box measured emission is
30dB bellow the standard radiated emission allowed limit ,at worst
harmony (500Mhz I have connected all the ports with STP long cables each
cable I have connected contributed about 10dB of noise These cables are
STP , how comes they raise the emission? What can I do to reduce the
added emission levels ? Can anyone suggest the blackmagic capacitor that
will solve the above, or anything else Again, emission peaks are seen at
Nx125Mhz, where worst results are at 500Mhz

Most of the emissions from reasonably small boxes come from the cables
connected to them, because the boxes are too small to be efficient
radiators.

You have presumably 'grounded' the box, but at VHF the ground lead is an
inductor, having quite a significant impedance, so the box, and the
cable shields connected to it, are at an r.f. potential to 'true
ground'. You can stop the r.f. current flowing along the cable shields
and being radiated by putting ferrite collars around the cables close to
where they exit the box. You may well need to do this at both ends of
the cables.

The fact that the cables are STP is irrelevant; the r.f. is on the
shields. If you used UTP, the r.f. would be a common-mode signal on the
pairs.

 

[John Woodgate]

What have you done with the shields? I suggest you ground them at the
cabinet and float the station end.

Floating shields is not a good thing to do, unless there is a problem
with low-frequency currents flowing through them. In principle, shields
should act like pipes connecting two tanks of r.f.'water', so they
should be solidly connected at both ends to the **outside surfaces** of
the 'tanks'.

If there are circulating low-frequency currents that can't be diverted
by other means, then the shield may be interrupted at the *receiving*
end by a very low-inductance capacitor. A 10 nF discoidal capacitor has
been used successfully. You put this at the receiving end because
interrupting at the sending end causes differences in STP cable
capacitances, shield to conductor, to degrade common-mode rejection.