Current flow direction during matching resistor

[rchemban]

Can u explain what happens if I use the good matching resistors/transistors by common centroid but I have not followed the current in same direction.

Thanks.
Chemban

 

[computaholic]

What do you mean bei different direction? The current should not flow at different angles (e.g. 0° and 90°). To my understanding, this is because the silicon does not behave identical for all direction.

If the current flows the opposite direction, you have to count your normal and mirrowed transistors. Try to have equal values for that. You might want to look up 'chirality of transistors', this will help.

As for resistors, opposite direction shouldnt matter, orthogonal does depending on the resistor type: (n/p)well: dont do!, poly/metal...: shouldnt matter.

 

[erikl]

Common centroid layout techniques very often have to put up with current flow in opposite directions. I think this is only a problem from well proximity effects: View attachment Implications_of_Proximity_Effects_for_Analog_Design.pdf

 

[leo_o2]

I think the electrical and magnectic effect caused by different direction current will have different effect.