How does the effective-capacitance (k-factor) model correct for what Elmore delay gets wrong, and why must k be computed iteratively?

Short Answer

Given driver resistance R_d, wire resistance R_w, and loads C1/C2, wire delay is D_w = R_d(C1+C2) + R_wC2. If R_d >> R_w, driver delay is accurately a function of total load (C1+C2). If R_w is comparable to or exceeds R_d, driver delay decreases and part of C2 is shielded, so delay is characterized as a function of (C1 + k*C2), where k (the effective capacitance factor) ranges 0 to 1. Because k itself depends on driver resistance -- which depends on the delay being computed -- it has to be computed iteratively, independent of any specific driver model, rather than solved in one closed-form step.

Technical Explanation

• Given driver resistance R_d, wire resistance R_w, and loads C1/C2: wire delay D_w = R_d(C1+C2) + R_wC2.
• If R_d >> R_w, driver delay is accurately a function of the TOTAL load (C1+C2) -- Elmore-style summation works fine here.
• If R_w is comparable to or exceeds R_d, driver delay decreases and part of C2 is effectively shielded from the driver -- delay is then characterized as a function of (C1 + k*C2), where k ranges 0 to 1.
• k (the effective capacitance / k-factor) depends on driver resistance -- but driver resistance itself depends on the delay being computed, creating a circular dependency that has to be resolved ITERATIVELY, not in one closed-form calculation.
• This iterative computation is done independent of any specific driver model -- it's a general correction mechanism, not tied to one particular cell's characterization.

Formula Or Decision Rule

D_w = R_d(C1+C2) + R_wC2 in general; when R_w is significant, effective load becomes (C1 + k*C2) with 0 <= k <= 1, computed iteratively since k depends on R_d which depends on the delay itself.

What To Check

• Warning sign: delay calculation for a net with significant wire resistance relative to driver resistance converges slowly or shows instability during extraction/timing.
• Inspect: confirm the delay-calc engine is genuinely iterating to solve for k rather than using a single-pass approximation on a net where R_w is significant.
• Correct: this is expected behavior for resistance-significant nets -- the iteration is inherent to getting an accurate k-factor, not a sign of a problem to eliminate.

Healthy, Suspicious & Hard-stop Results

• Expected: nets with wire resistance comparable to driver resistance show delay-calc results consistent with iterative k-factor correction, not a naive full-load Elmore assumption.
• Investigate: delay-calc results for a resistance-significant net look identical to what a naive full-load (C1+C2) Elmore calculation would predict -- worth confirming the k-factor correction is actually being applied.
• Stop: a hand-verification of a critical net's delay assumes k=1 (no shielding) on a net where wire resistance is clearly significant relative to driver resistance -- that assumption discards the entire point of the k-factor correction.

What The Interviewer Is Testing

Whether you understand WHY k-factor computation must be iterative (the circular dependency between k and driver resistance), not just that k exists as a correction between 0 and 1.