A squarewave at 12MHz will cause radiated emission at 12MHz, 36MHz, 60MHz and so on. But sometimes emission can be seen at other frequencies, where the 12MHz can be seen as delta frequencies between them, e.g. 468MHz, 480MHz, 492MHz. Why is that? What should one be looking for on a PCB when this is displayed on the spectrum analyzer?

Another case is when doing radiated emission measurements a large "bump" on the dBµV/m graph can be seen. Why they are a "bump" and not individual frequencies are probably because the RBW and video BW settings, but what should one be looking for on a PCB when this is the case?

 

A square wave can be produced by adding odd harmonics to a fundamental frequency sine wave – which is why you are seeing emissions, the frequency of which gives an odd number when divided by the fundamental frequency (12MHz in this case).

So fix the 12MHz emission, and the harmonics will disappear.

See this wikipedia article on square waves, in particular the Fourier analysis:-
https://en.wikipedia.org/wiki/Square_wave

 

Well, the first sentence was just a statement before the two following questions.

Why are 12 MHz as delta at these high frequencies?

And why is the "bumps" commonly seen from the SA? This last one was just a general question not related to 12MHz. I usually see them at 30-200M radiated emission.

 

Well, the first sentence was just a statement before the two following questions.

Why are 12 MHz as delta at these high frequencies?

And why is the "bumps" commonly seen from the SA? This last one was just a general question not related to 12MHz. I usually see them at 30-200M radiated emission.

If it is not a "perfect" square wave with a 50.00% duty cycle then there will be other harmonics as well. And unless every element of the electrical circuit is perfect, additional harmonics due to the non-linear functionality will appear as well. Perfect waves only appear in textbooks.

 

Good point. I guess bad decoupling, long and narrow traces can explain this?

 

Good point. I guess bad decoupling, long and narrow traces can explain this?

Those would contribute to the results.