Signal integrity - Crosstalk between microstrips

How to use openEMS. Discussion on examples, tutorials etc

Moderator: thorsten

Post Reply
jockeosth
Posts: 13
Joined: Wed 11 Jan 2017, 20:24

Signal integrity - Crosstalk between microstrips

Post by jockeosth » Mon 15 May 2017, 23:10

Hi!

I start to make a setup to calculate the coupling between two transmission lines. I like to use the S-parameters to convert to time-domain crosstalk values.
From the simulation I get approximately the expected results. When using spacing == line width for 50R line I should get about 4.4% near-end crosstalk, when spacing = 2*line_width about 1.9% and when spacing = 3*line_width about 1% according to the book "Signal and power integrity - simplified" (figure 10-27).
In simulation I get around 3%, 1.2% and 0.68%. Around 50% difference, I have no other source to compare to at the moment so I think reasonable values as a starting point.

However I have some observations:
The S-parameters at DC may be off from the expected ideal values. For example S31 should be 0 and S21 very close to 1. I know the correct DC and low-frequency values are important for conversion back to time-domain. I think the small near-end crosstalk that still is present long time after the pulse have travel over the transmission line is due to this, the final value are very close to S31 at DC/low freq.
Hence, is there a good way to improve the S-parameter resolution in the simulation?
When I use the calcPort() function I think it extrapolate points?

Best regards
Joakim
Attachments
Microstrip_CouplerAirBoxLossyGND.m
(8.85 KiB) Downloaded 211 times

thorsten
Posts: 1388
Joined: Mon 27 Jun 2011, 12:26

Re: Signal integrity - Crosstalk between microstrips

Post by thorsten » Mon 15 May 2017, 23:36

Hi,

sounds complicated and I'm not sure I can really help you.
I don't know if I can find some time to look at your model.
But what you can try is play around with the Gaussian pulse parameter to have more (or less?) power at low frequencies.
Furthermore you need to make sure the energy in your system decays really good and far down (maybe to -60 or more?). Some realistic losses to your substrate can help here.

regards
Thorsten

jockeosth
Posts: 13
Joined: Wed 11 Jan 2017, 20:24

Re: Signal integrity - Crosstalk between microstrips

Post by jockeosth » Fri 19 May 2017, 16:18

Hi again,

I found a comparative study of different EM-tools that include crosstalk as well. Using the same setup as shown here:
http://www.pcdandf.com/pcdesign/index.p ... ation-1512
I got the same results with OpenEMS (about 4.2%), as expected :)

Before I found the comparative study above I try to model in Sonnet-Lite too with a close match to the results from OpenEMS.
The small difference after the pulse where the crosstalk should be 0 seems to be due to the conversion from S-parameters to time-domain is very sensitive for phase/amplitude at low frequencies.

Thanks again for advice/support and a very good tool!

Best regards
Joakim

Post Reply