analysing differential line cross talk with VNA

[john23]

Hello ,I have a situation as shown below were there are 3 differential lines one is a victim and two agressors.
I have one agressor is far away then the other agressor .
what is the best way to see the crosstalk between the differential ines.
buttom line We have digital data which is getting distorted in the time domain.
Differential mode cross talk is a mode that calculates the crosstalk from differential line.
maybe its better to analise the cross talk by common mode Scc21?
What is the best method to do in VNA?
Thanks.

[mtwieg]

What is the best method to do in VNA?

So you have twelve total single-ended ports. So assuming you have just a two port VNA, you would need to characterize it with multiple two port measurements (78 to be exact). Then you can combine those into a single 12x12 S matrix, and then convert that to a mixed-mode S matrix.

maybe its better to analise the cross talk by common mode Scc21?

Depends. Do the aggressors actually create significant common mode? Is the victim sensitive to common mode?

[john23]

Hello Mtweig, we have digital data getting distorted by the agressor.I have simulated a basic structure shown in the photos which ilustrated the effect of far aggressor and close aggressor
 can I see if the common mode is the dominant factor of distorting the digital signal?
Thanks.

[mtwieg]

can I see if the common mode is the dominant factor of distorting the digital signal?

Not from network parameters alone. You have to measure how much of each mode the aggressor is actually injecting into its ports.

[john23]

Hello Mtweig,I know that I need to do IFFT for the impulse responce and analyze it.
Two questions:
1.suppose I have SDD21 where differential port 1 is the agresssor and differential port 2 is the victim.
When I do IFFT the time domain represents the signal travelim on the victim or the agressor?
For example I have an Impulse at 0.1ns is on the victim?(which i need to multiply by the velocity to find exact location)
 
2.I can find nowehere on the internet the formulas for how much energy got coupled to the victim in the impulse.could you please give me some manualon the subject?
Thanks.

"you have to measure how much of each mode the aggressor is actually injecting into its ports."

[mtwieg]

Hello Mtweig,I know that I need to do IFFT for the impulse responce and analyze it.

I don't see any reason to transform to time domain...

1.suppose I have SDD21 where differential port 1 is the agresssor and differential port 2 is the victim.
When I do IFFT the time domain represents the signal travelim on the victim or the agressor?

IFFT of the network parameters simply transforms from a frequency-domain representation of the network to a time-domain (impulse response) representation of the network. It still does not tell you how much of any mode is being injected into your system. That requires other measurements with an oscilloscope or spectrum analyzer.

Do you actually have a real system you're characterizing, with real crosstalk, or is this purely a hypothetical simulation? Or a homework problem?

[Kalvin]

Here is nice article describing how to make differential line measurements using 2-port VNA:

https://www.signalintegrityjournal.com/articles/4102-using-ultra-broadband-baluns-to-perform-differential-s-parameter-measurements-using-single-ended-2-port-vna

Terminate the transmission lines using the impedances your actual circuit is using when measuring crosstalk.

[john23]

Hello Mtweig,could you please share a  manual of how do we use spectrum analyzer the see the location and quantity injected by agressor in the victim net?
 Its for general knowledge very intresting.

 That requires other measurements with an oscilloscope or spectrum analyzer.

Do you actually have a real system you're characterizing, with real crosstalk, or is this purely a hypothetical simulation? Or a homework problem?

[mtwieg]

Hello Mtweig,could you please share a  manual of how do we use spectrum analyzer the see the location and quantity injected by agressor in the victim net?
 Its for general knowledge very intresting.

Without knowing more about your actual setup, I can't give more specific advice. You need to provide more info on the actual hardware being tested.

[tszaboo]

I would first try to simulate this.
You don't need a field solver, based on the picture you posted. You can simply treat this as a coupled microstrip. You enter the physical parameters into a simulator, especially your prepreg and distance, terminate it, and then simulate the Sxy.
You should be able to do that for example in qucsstudio, if you don't have a more expensive simulator, this one is free. Then you have a good feeling about it, if it even makes sense to measure it.
But I would take a guess and say that's it's not coupling but something on your power supply. You are at what, -40dB?

[john23]

Hello Mtweig, Its a basic passive PCB with differential lines going threw it.
The ends of the differential transmission lines are terminated.
The cross talk is from same side agressor NEXT.

I see very high values of NEXT in the frequency domain.
I just want to know where exactly the most problematic point where the most energy gets coupled.
How can I do that?
Is there a manual I could use?
chatGPT gives me theorys that  you said are wrong so I am looking for manuals.