Verdin iMX8MP - Radiated emission is too high at 4GHz

For our customer we have developed a carrierboard. During pre-compliance Radiated Emission measurements there was seen the emission was quite high around 4GHz, a few dB over limit.

After investigation by using of a near field probe weakest point is around the RAM/traces:
image

We have used two different som modules and results is the same.

Is this issue known? What can we do to reduce the emission level?

Used SoM:
Verdin IMX8MP Q 4GB IT V1.1A.

Hello @iwk

Did you make a comparison test using one of our carrier boards? The test with nearfield probes is usually showing spikes with the frequency you see in the radiated emission test somewhere on the board but in most cases, this does not mean that there is a problem. e.g you will see a spike of 25 MHz over the crystal oscillator but it is most likely not the source of radiated emissions of multiple of 25MHz.

The diagram you send only shows the range from 3 to 5Ghz. Do you also see harmonics of this frequency in the lower band ?

Best Regards,

Matthias

Hello Matthias,

At a testhouse a pre-compliance was performed 1-12GHz:

Marker shows a value over the limit line.

Remark: during the test Toradex heatsink was mounted (this part is directly connected to GND). I think this should be correct.

Hello iwk,

i like to take a look at your carrier design.
can you sent the schematic to support.eu@toradex.com?
Or we can set up a Zoom call so that I can look at you design.

Best Regards,

Matthias

Hello Matthias,

I prefer to set up a Zoom call. So than we can show you schematic and PCB design.

When do you have time for a Zoom call?

Wiebren

Hello @iwk

I send you a image with Spread spectrum enabled. Did you had time to test that?

Best Regards,

Matthias

Hello @matthias.tx ,
I received the image. Tommorow I will verify this image. My proposal is to test this on a Verdin board and compare this to the early used image. The image is succesfully download.

Did you also reproduce this issue on a board. A few months ago you had plans to reproduce the issue on antoher carrier board (oher customer).

Other question has other customer perform radiated emission tests at 3GHz and higher?

Best Regards,
Wiebren Keultjes

Hello Matthias,

I verified the emission on a Verdin Development Board:

Difference in emission is 14dBm lower for spread spectrum image, big difference in emission. How is the performance and stability when spread spectrum is configured? Or what are risks? What are next steps?

Currently we are working on a update of the carrier board. We expect to have a newer revision in approx. 2 months and than perform a new Pre-compliance test with focus radiated emissions. It is possible to participate the pre-compliance we can invite you when we have a booking.

How we can add the spread spectrum to our image?

Best Regards,
Wiebren Keultjes

Hello @iwk,

great to hear that. I will discuss the next steps with the Team.

Best Regards,

Matthias