In short: A Zener diode is an ordinary junction engineered to break down at a sharp, repeatable reverse voltage, either by quantum tunneling (below about 5 V) or by avalanche multiplication (above about 7 V), and it regulates well only when you keep it above its knee current, where its dynamic resistance is small.
Reading the Zener V-I curve
Plot current against voltage for any silicon junction diode and you get three regions. In forward bias the current is negligible until roughly 0.6 to 0.7 V, then rises exponentially. In reverse bias below the breakdown voltage only a tiny leakage current flows, typically nanoamps, set by thermally generated minority carriers. Past the breakdown voltage VZ the reverse current explodes while the voltage across the device barely moves. That near-vertical third branch is the whole reason a Zener exists.
The key difference from a rectifier is intent. A rectifier is rated so that breakdown never happens; a Zener is doped, sized and packaged so breakdown is the normal operating mode, limited only by power dissipation PZ = VZ x IZ. A 5.1 V, 500 mW part such as the BZX55C5V1 can therefore carry about 98 mA continuously at 25 °C before it overheats.
Key fact: the datasheet VZ is quoted at a specific test current IZT (for a 1N4733A, 5.1 V at 49 mA). Run it at 5 mA and the voltage will be measurably lower, by roughly rZ times the current difference.
Two physical mechanisms, one symbol
Engineers say "Zener diode" for every breakdown regulator, but two different physics are at work.
- Zener (tunneling) breakdown. With both sides very heavily doped, the depletion region is only a few nanometres wide. The field exceeds roughly 106 V/cm and valence electrons tunnel straight across the bandgap, with no collisions needed. This dominates below about 5 V.
- Avalanche breakdown. With lighter doping the depletion region is wider. Carriers crossing it gain enough energy between collisions to knock new electron-hole pairs loose by impact ionization; each new carrier does the same, and current multiplies. This dominates above about 7 V.
- Between 5 and 7 V both contribute, which is why parts near 5.6 V are the most thermally stable.
Why the temperature coefficient flips sign
Higher temperature shrinks the bandgap slightly, so tunneling gets easier at a lower voltage: the temperature coefficient is negative, about −2 mV/°C for a 3.3 V part. In avalanche devices, hotter lattice vibrations scatter carriers more often, so they need a higher field to gain ionizing energy: the coefficient is positive, perhaps +3 to +6 mV/°C at 12 V. Around 5 to 6 V the two cancel. That is why precision references combine a roughly 6.2 V buried Zener with a forward diode junction (−2 mV/°C) to build near-zero-drift references, the trick behind classic devices like the LM399.
Key fact: for a 12 V regulator, a 50 °C rise moves VZ by roughly +0.2 V from tempco alone. That is why you should not use a Zener as an absolute reference in an ADC without checking drift.
Knee current and dynamic resistance
Below the knee current IZK the junction is only partly in breakdown, the curve is rounded, and the voltage sags. The 1N4733A lists IZK = 1 mA with an impedance of about 550 Ω there, versus 7 Ω at its 49 mA test point. That change is the whole story of regulation quality.
Dynamic (small-signal) resistance is the slope of the breakdown branch:
rZ = dV / dI (at the bias point)
Delta_VZ = rZ x Delta_IZ
Example: rZ = 7 ohm, supply ripple makes IZ swing 10 mA
Delta_VZ = 7 x 0.010 = 70 mVBias at 1 mA with 550 Ω and the same 10 mA swing would shift the output by 5.5 V, which is useless. Rule of thumb: bias at least 5 to 10 times IZK, and never let the minimum supply and maximum load push IZ near the knee. Low-voltage tunneling parts (3.3 V) have much softer knees than a 6.8 V part; you can see 100 Ω or more at 3.3 V, against 10 to 15 Ω at 6.8 V.
Comparing the regimes
| Property | Zener (tunneling) | Avalanche |
|---|---|---|
| Typical VZ | Below ~5 V | Above ~7 V |
| Doping | Very heavy, thin depletion | Moderate, wide depletion |
| Tempco | Negative | Positive |
| Knee | Soft, rounded | Sharp |
| Dynamic resistance | Higher at low VZ | Lower |
| Noise | Lower | Higher (multiplication noise) |
Practical design check
Take a 5.1 V Zener fed from 12 V through a series resistor, supplying a 100 Ω load (51 mA). Choose IZ(min) = 10 mA, ten times the knee. Then R = (12 − 5.1) / (51 mA + 10 mA) = 113 Ω, so use 110 Ω. Zener current with no load rises to about 63 mA, dissipating 0.32 W, safe for a 500 mW part but with little margin at 50 °C ambient where you should derate. This kind of margin check is exactly what interviewers probe. We cover the circuit side in Zener diode applications, and the junction physics behind these numbers in PN junction depletion region explained.
In simulation, SPICE-based tools such as Cadence Spectre/PSpice, Siemens EDA Questa ADMS and Synopsys HSPICE model breakdown with the BV, IBV and NBV parameters of the diode model. Many default models set BV effectively infinite, so a "Zener" that never breaks down in simulation usually means BV was never set. On-chip, foundry PDKs often provide a Zener or "ESD clamp" device whose breakdown you characterize with a DC sweep, because the same knee behaviour sets your ESD protection and bandgap-reference headroom. Teams at design-services firms such as Avecas routinely verify these corners across process and temperature.
Key fact: always specify minimum Zener current at worst case (lowest supply, highest load) and maximum Zener power at best case (highest supply, no load). Passing one corner and ignoring the other is the most common Zener design bug.
Where to go next
Once you can sketch the curve and defend each number on it, practice by extracting VZ and rZ from a measured or simulated sweep. If you want guided practice with labs and interview questions, see the CourseTron courses, and for instructor-led VLSI career training, visit ChipXpert's training programs.
Ready to master analog fundamentals with simulations and interview-style problems? Explore the analog electronics track in our CourseTron courses and start building real circuit intuition today.
FAQ
Is a Zener diode the same as an avalanche diode?
Electrically they share the symbol and use, but the mechanism differs. Below about 5 V breakdown is tunneling; above about 7 V it is avalanche. Most "Zener" parts above 7 V are really avalanche devices.
What happens if Zener current falls below the knee current?
The voltage drops below VZ and the dynamic resistance climbs sharply, so ripple and load changes pass through to the output and regulation is lost.
Why is the 5.6 V Zener used for references?
Its tunneling (negative) and avalanche (positive) temperature coefficients roughly cancel, so VZ drifts very little with temperature.
How do I calculate dynamic resistance from a datasheet?
Divide the small change in Zener voltage by the small change in current at the bias point. Datasheets list it as ZZT at the test current and ZZK at the knee current.