Unusual EMI frequencies

[Matt]

Hello,

We currently have a system in environmental testing and it is
having an issue passing a particular EMI test (RE-102) and we have
spikes above the acceptable limit at and around 22 MHz and 29 MHz. The
odd thing is is that these spikes occur at completely different
systems in the same test (different types of power supplies and
enclosures). Could the issue be with the labs input power (isn't this
frequency range within the PLC communications frequency range?). I
should also mention that the spike grows as we turn more components in
the system on, i.e. it increases with the load.

Any help or ideas would be great.

Thanks,
Matt

 

[Joerg]

Hello,

We currently have a system in environmental testing and it is
having an issue passing a particular EMI test (RE-102) and we have
spikes above the acceptable limit at and around 22 MHz and 29 MHz. The
odd thing is is that these spikes occur at completely different
systems in the same test (different types of power supplies and
enclosures). Could the issue be with the labs input power (isn't this
frequency range within the PLC communications frequency range?). I
should also mention that the spike grows as we turn more components in
the system on, i.e. it increases with the load.

Any help or ideas would be great.

What else is in those systems?

You should be able to narrow down exactly where that comes from, using
EMCO near field probes or the like. A possibility is that something
intermodulates. For example one strong TV signal with another, inside
your unit. But no matter what, you've got to figure out where that
happens or they won't give you that coveted compliance certificate.

 

[Rich]

What else is in those systems?

You should be able to narrow down exactly where that comes from, using
EMCO near field probes or the like. A possibility is that something
intermodulates. For example one strong TV signal with another, inside
your unit. But no matter what, you've got to figure out where that
happens or they won't give you that coveted compliance certificate.

Joerg,

Thanks for the reply. We have a variety of different peice parts
in each system. For instance, one uses a linear power supply (AC/DC)
and the other uses a switching. There aren't and RF communication
hardware or monitors (except in one system, which was turned off and
we still saw the spike albeit lower than before. I am heading up to
the lab tomorrow so I will try using a near field probe on the system.
I'll report my findings.

Thanks,
Matt

 

[Matt]

Joerg,

    Thanks for the reply. We have a variety of different peice parts
in each system. For instance, one uses a linear power supply (AC/DC)
and the other uses a switching. There aren't and RF communication
hardware or monitors (except in one system, which was turned off and
we still saw the spike albeit lower than before. I am heading up to
the lab tomorrow so I will try using a near field probe on the system.
I'll report my findings.

Thanks,
Matt- Hide quoted text -

- Show quoted text -

It seems Rich had used my computer while I stepped away. The message
above was posted by me. Sorry for any confusion.

Matt

 

[Joerg]

It seems Rich had used my computer while I stepped away. The message
above was posted by me. Sorry for any confusion.

That's what I thought

Anyhow, what I meant with other stuff: Is there some kind of micro
controller or other circuit that needs and has a clock source? Or any
OEM parts that might have something like that inside? If so, then I'd
first look for anything that is common between the linear and the
switcher system.

Switch mode supplies rarely generate noise above a few MHz although I
have seen a serious blunder by a major power supply manufacturer where
it emitted 250MHz at full bore. Probably an oscillating opamp or
comparator that nobody had noticed until I did the EMI, for years abd
years. In other words one cannot necessarily trust OEM components when
it comes to EMI no matter what their claims on the datasheets are.

 

[John Popelish]

Hello,

We currently have a system in environmental testing and it is
having an issue passing a particular EMI test (RE-102) and we have
spikes above the acceptable limit at and around 22 MHz and 29 MHz. The
odd thing is is that these spikes occur at completely different
systems in the same test (different types of power supplies and
enclosures). Could the issue be with the labs input power (isn't this
frequency range within the PLC communications frequency range?). I
should also mention that the spike grows as we turn more components in
the system on, i.e. it increases with the load.

Any help or ideas would be great.

Sounds to me like a room resonance. Are these measurements
taking place inside a shield room?

 

[Matt]

Sounds to me like a room resonance.  Are these measurements
taking place inside a shield room?

John,

Sorry for such a late delay, I was up at the lab testing quite a
bit. The room was a shielded room and the input power was running
threw line filters before it reached the EUT. We believe at this
moment that it has something to do with an unshielded power cable
because we were able to run the system using internal battery power
and we didn't see the same noise level at the 22 MHz frequency range.
Unfortunately, we got booted from the EMI chamber so the company we
were using could perform an audit. I will have to try a few more tests
next week.

Thanks,
Matt