How to fix interconnect(wire EM) EM violation in vias using Magma Tool?

[vlsi_deepa]

Hi friends,

For SOC design, I am performing EM analysis.Current density-peak.

While I select average current density,there is no cell and wire em analysis.For peak current density there is no cell Em analysis.But for interconnect(wire EM) EM analysis,I got slack of –2.06mA/cut for via12.Nearly 100000 nets got interconnect EM violation.I am using Magma Tool.

Q: How to fix interconnect(wire EM) EM violation in vias using Magma Tool?

Its somewat urgent.

Thanks and Regards,
Deepa.

 

[praneshcn]

Re: EM ANALYSIS

hi did u find a solution for the question u had asked long ago. hope u have and if so could u please share the problem and solution in more detail?



how about increasing the number of vias at places where em is occuring. I once did tat for solving via reliability issue. I added some additional vias manually.

 

[vlsi_deepa]

Re: EM ANALYSIS

Hi praneshcn,

I have fixed all EM violations for current density Average.... Most of people said,Thats enough... No need to concentrate in Peak Current Density ...Because the check should be done using Average Current Density... Its a actual Value of Current Flowing.. Peak wont give the Actual value of Current Flowing...
Soln for EM:
*We can increase the Metal Layer to solve EM Issue...
*If we get Violation on Vias,We can increase the VIA size (If possible)... Refer Ur Rule File,And then U can try this..... U can add additional via manually,Wat u did for solvin via reliability....
*Check the Powerplanning datas also... If need,increase the Width of Meshes...

(Correct me,If i am wrong)

Regards,
Deepa...:|

 

[santuvlsi]

Re: EM ANALYSIS

Hai deepa,

What u have said is excatly correct,

Power EM problems are solved by increasing wire widths and

inserting vias, so that RMS current density through wires is maintained below the threshold.

Santhosh

 

[vlsi_deepa]

Re: EM ANALYSIS

Hi Santhosh,

But In some cases,Even though we are inserting vias and Wire sizing,It Doesnt eliminating EM issues... RMS current density is not maintained the thershold level...In my deisgn,I got the same prob.. Couldnt eliminate EM for currnet density RMS... Wat wil be soln? We have to go back and Restructuring the powerplan?

Thanks,
Deepa....

 

[master_picengineer]

EM ANALYSIS

Hi,
What about spacing between wires ?

 

[Asha_eda]

Re: EM ANALYSIS

Hi master_picengineer,

I dont think so Spacin wil reduce EM Violation.. We wil be using Spacin for Eliminating Noise(Crosstalknoise and Crosstalk delay)....Can u tel me,How Spacin wil reduce,EM Problem?

Regards,
Asha

 

[master_picengineer]

EM ANALYSIS

Yes this is when we talk about signals, spacing help eliminating Crosstalk. When we talk about copper the more wires are close the more capcitance is high so Electrostatic and magnetics field/flow which is the cause of the EM.

 

[Asha_eda]

Re: EM ANALYSIS

Hi,
Worthable Information...
Ok fine..I wil try this to eliminate EM Violation...If i get any prob,Come back to u...

Thanks,
Asha...

 

[ericyuan]

Re: EM ANALYSIS

what's the difference between RMS and average and peak?
any example?

 

[Asha_eda]

Re: EM ANALYSIS

Hi,

While we are doin EM,We have to consider 3 types of Current Density of Wire:
1.Average 2.Peak and 3.RMS...
Average: Average is associated wit Long term reliability of wires.Currents in Wires are purely AC currents (no Dc components),.. therefore required a specila definition of Average.. While using Slope current calculation method,absolute average definition is used.. If not,using commands u have to define...
2.Peak: Related to potential near term failure of wires...
3.RMS: Related to joule self heatin of wires... wire deterioration due to temperature effect.. Incre in Temperature,approx equals to square of RMS current...

If we solve Average and Peak then,No need to worry about RMS Current density violations...

Regards,
Asha...