Questions regarding delay in wire

[venkatramanan]

if width of the metal increases means resistance ll be reduced?what ab capacitance?it ll increase or decrese?

 

[Gagan23]

Re: regarding delay in wire

Hello Venkat,

R= resistivity*Length/Cross section area.

In width increases, cross sxn area increases and resistance, therefore, decreases.

C= Epsilon*Cross sectional area of plates/distance between plates.

I am not sure I am visualizing it right but going by this formula, distance between plates is increased. Capacitance is bound to decrease.

delay = R*C. So, it will decrease as a square of the width.

 

[godfreyl]

Re: regarding delay in wire

C= Epsilon*Cross sectional area of plates/distance between plates.

I am not sure I am visualizing it right but going by this formula, distance between plates is increased. Capacitance is bound to decrease.

I would expect capacitance to increase as wider metal means greater area. :-?

 

[Gagan23]

Re: regarding delay in wire



seeing this formula...

as width, that means radius of the wire increases, the capacitance would decrease.

 

[venkatramanan]

Re: regarding delay in wire

thank uguys...
one question .consider one net.if width of the metal =2micron length of the metal =5micron now i'm increasing metal width=4 so wat ab net delay?increse or decrese.resistence ll increase wat about capacicance?

 

[godfreyl]

Re: regarding delay in wire

View attachment 78296

seeing this formula...

as width, that means radius of the wire increases, the capacitance would decrease.

How do you figure that out?
From that formula, for a given length: increase in the radius a => decrease in l/a => decrease in ln(l/a) => increase in capacitance.

 

[FvM]

Re: regarding delay in wire

As in most cases, the question suffers from an incomplete problem specification. Wire capacitance and inductance are unknown without defining the full geometry.

To refer to a well-defined case with known electrical properties, you can assume that the wire is part of a transmission line, either coaxial or parallel wires. Then increasing the diameter will increase capacitance per length unit and decrease inductance. As a result the characteristic impedance Z = Z0*√L/C will decrease, but the propagation speed (~ √LC) stays constant.

To give a first order estimation, I would expect a similar behaviour for arbitrary wire geometries, e.g. bond wires.

 

[birdy123]

Re: regarding delay in wire

Hi All,

Why cann't you read this..

Interconnect RC corners

 

[Gagan23]

Re: regarding delay in wire

yes, you're right.

Going by it, Capacitance would increase.

Venkat, C will increase. I am sorry about previous statement.

- - - Updated - - -

Venkat,

R will decrease by 4 times the previous value.

ln(5/2)= ln 2.5= 0.916.
1/ln(2.5)= 1.0913

C will roughly be the same or rather increase by a factor of 1.0913 aprrox.
Note: I am ignoring higher order terms.

 

[jeet_asic]

Re: regarding delay in wire

One doubt , correct me if i am wrong.

To avoid electromigration , one of the way is to choose wire width according to drive strength. So if i choose wider width wire for higher drive strength gate , it will have more capacitance. It will thus increase parasitic.

Am i correct ? if yes , then what factor we should focus more , resistance or capacitance ?

Pls correct me if i am wrong.

 

[Somashekhar]

Re: regarding delay in wire

Electromigration depends on current density.
Its better to put wider metals for higher drive strength.

 

[Gagan23]

Re: regarding delay in wire

Jeet,

In the quick first look, C will increase as well as parasitic capacitaces also.

 

[pavanks]

Re: regarding delay in wire

The cap will increase with the inc in area of the metal.
So if u inc the width or the length the cap will inc.