Bicmos Logic 1980s Concede the Light Load Case Pure CMOS Still Wins
# BiCMOS Logic (Late 1980s): Step 11 — Concede the Light-Load Case: Pure CMOS Still Wins
## 1. Taking the Crossover Point Seriously in Both Directions
Step 10's crossover point is only useful if it is honored on both sides of the line, and this step exists specifically to state the side that does not favor BiCMOS: for the large majority of gates in a typical digital design — internal logic gates driving only a handful of adjacent gate inputs, well below the crossover capacitance — a BiCMOS output stage is strictly worse than a pure CMOS gate. It adds the fixed base-drive and junction-capacitance overhead from Step 10 while gaining nothing, because the load was never large enough to reach the region where bipolar's shallower delay slope pays that overhead back.
## 2. Why This Keeps BiCMOS From Being a Universal Replacement
This is the direct, unavoidable consequence of a merged process: adding a bipolar output stage everywhere would make the typical, lightly loaded gate slower and more power-hungry than leaving it as pure CMOS, which is precisely why BiCMOS designs of this era used bipolar output stages selectively — at the specific nodes identified in Steps 12 through 14 — rather than converting every gate in the design. The right question for a BiCMOS designer was never 'is bipolar faster,' but 'is this specific node's load above or below the crossover,' and for most internal logic, the honest answer was below it.
## 3. Companion Diagram One: Quantitative View
## 4. Companion Diagram Two: Structural View