The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> coupled capacitance for differential signal layout
https://designers-guide.org/forum/YaBB.pl?num=1285239023

Message started by newic on Sep 23^rd, 2010, 3:50am

Title: coupled capacitance for differential signal layout
Post by newic on Sep 23^rd, 2010, 3:50am

As i know, differential signal lines should be drawn as close as possible in the layout. The extracted coupled capacitance (CC) will be large due to close proximity. Is it ok? In the parasitic net table, the total cap (C+CC) value will be dominant by CC value. Am i doing the right thing?

Title: Re: coupled capacitance for differential signal layout
Post by love_analog on Oct 11^th, 2010, 6:24am

You place diff signals close together so that noise coupling on both lines is the same.
If you are worried about CC, you can place them further apart. Just make sure there is no differential noise coupling. One way is to shield both lines.

Title: Re: coupled capacitance for differential signal layout
Post by rfmagic on Oct 11^th, 2010, 8:10am

As long as the coupled capacitance CC is taken into acount (as you probably did) you are doing the right thing. Just make sure that both differential lines follow route i.e. any potential noise should be coupled to both lines

Title: Re: coupled capacitance for differential signal layout
Post by newic on Oct 12^th, 2010, 5:15am

Which approach is better?

1. put differential lines close together and get large CC
2. put differential lines apart like 2x min spacing, and get lower CC

Title: Re: coupled capacitance for differential signal layout
Post by raja.cedt on Oct 12^th, 2010, 6:11am

i feel put as near as posb and try to route a thin vss line in between

Thanks.

Title: Re: coupled capacitance for differential signal layout
Post by rfmagic on Oct 12^th, 2010, 7:06am

I agree. put them as close as possible but in your simulation add CC and dont forget the capacitance to gnd

Title: Re: coupled capacitance for differential signal layout
Post by RobG on Oct 12^th, 2010, 7:52am

I've never liked the ground line between the signal lines because your cap increases if you keep the separation the same.

As a sort-of oldtimer, I was stunned by how much the sidewall cap dominates the overall cap in modern processes. Metal sheet resistance is very large too.

Has anyone seen studies of how close lines have to be before differential noise becomes an issue? Seems like it would be on the order of microns given older processes seemed to work.

rg

Title: Re: coupled capacitance for differential signal layout
Post by thechopper on Oct 12^th, 2010, 5:34pm

newic wrote on Oct 12^th, 2010, 5:15am:

Which approach is better?

1. put differential lines close together and get large CC
2. put differential lines apart like 2x min spacing, and get lower CC

Having a large CC migth not be harmful, depending on the nature of your differential signal: if your signal is a low pass one most probably that CC will help filter it out and it ends up being beneficial. On the other hand if your diff signal is high frequency and / or diff clock signals you should look for minimizing CC.

Best
Tosei