Planar Ic Junction Isolation Step 07 Guarantee Reverse Bias Supply Design
# Guarantee the Reverse-Bias Condition Through Substrate-Tie and Power-Supply Design
## 1. Why the Fix Lives in the Power Supply, Not in the Diffusion
Because Step 6 showed that even a brief forward-bias excursion collapses isolation, this architecture cannot leave the reverse-bias condition to chance; it has to be guaranteed by design, and the standard guarantee is simple and specific — tie the substrate to the most extreme supply rail the circuit uses, so that no signal node, however it swings, can ever reach a potential past the substrate's own, by construction rather than by hope. For an NPN-island architecture, this means tying the substrate to the most negative rail the circuit will ever present; every island's n-type region can then only be more positive than the substrate, never less, which keeps the isolation diode's polarity pinned in the reverse direction regardless of what the active circuitry inside any island does:
where $\epsilon$ is a margin the designer chooses deliberately, not left to accident, and the entire guarantee rests on the substrate tie being a low-impedance connection to the rail — a high-impedance or poorly bonded substrate connection can itself drift under load, reopening exactly the failure Step 6 described even with the correct rail chosen.
## 2. Real Diagram: What a Weak Substrate Tie Lets Happen Under Load
A substrate tie with any meaningful impedance of its own will see its effective potential shift under load current, because every island's leakage and switching transients sum into that one connection; a substrate tie strong enough in the datasheet's quiescent condition can still drift into forward bias territory for some islands when the circuit is switching hard, which is why this guarantee is a layout and packaging decision, not only a schematic one.
## 3. Why This Step Closes the Bias-Guarantee Thread This Series Opened at Step 5
Steps 5 and 6 established that isolation is a standing, binary condition and named exactly what its failure looks like; this step answers the question those two steps leave open — how a real circuit actually keeps that condition true across every operating case, not just in the typical one. The answer is specific and load-bearing: a hard substrate tie to the correct extreme rail, implemented with enough physical conductivity that transient currents from switching islands cannot shift the reference the rest of the chip depends on.
Step 7 closes this series' bias-discipline thread by naming the concrete design practice Steps 5 and 6's requirements actually demand of a real power-supply and substrate layout.