When to Use a Zener Diode vs an LDO for Voltage Regulation

Data drawn from EE StackExchange "Zener vs linear regulator" discussions, Nexperia BZX84/BZT52H datasheets (Rev. 7, 2023), and TI LP5912 datasheet (SNVSA77).

The Shunt vs Series Distinction

A Zener is a shunt regulator. It sits in parallel with the load, burning off the current the load doesn't consume. The series resistor sets the total current: I_total = (V_IN − V_Z) / R. The Zener takes whatever the load doesn't. At no load, the Zener burns it all.

An LDO is a series regulator. It sits between the input and the load, acting as a variable resistor controlled by a feedback loop. It only passes the current the load actually draws. At no load, it burns almost nothing - just its quiescent current.

This is the fundamental difference. A Zener always runs at full power. An LDO runs at whatever power the load demands. For a 5mA reference, the difference is negligible. For a 500mA rail, it's the difference between cool-running and a space heater.

Per an EE StackExchange discussion, the community consensus was blunt: Zener shunt regulators are "only suitable for loads which draw a low and constant current. A linear voltage regulator would be a much better fit" for anything variable or above ~20mA.

When a Zener Is the Right Choice

Zeners shine where the load is small, constant, and noise-tolerant. Classic cases:

• MOSFET gate clamping. A 16V Zener gate-to-source protects V_GS from transient overvoltage. It only conducts during the transient - no sustained power draw. The Zener's noise doesn't matter because it's not in the signal path.
• Voltage references for comparators. A UVLO circuit comparing a divided rail against a Zener reference. Load current is microamps. The Zener's ±5% tolerance is fine for a trip point that doesn't need 1% accuracy.
• Level shifting in analog circuits. Dropping a signal by a fixed 2.4V before it enters an ADC. The soft knee of a low-voltage Zener can actually be useful here - graduated clamping instead of hard clipping.
• Pre-regulation for high-voltage inputs. An LDO rated for 26V max input can't handle a 36V rail. A Zener pre-drop from 36V to 16V keeps the LDO alive. The Zener handles the voltage, the LDO handles the regulation.

The common thread: load current under ~5mA, noise not critical, or the Zener is doing a job (clamping, level-shifting) that an LDO isn't designed for.

When an LDO Is the Right Choice

LDOs take over where Zeners hit their limits. The thresholds:

• Load current above 5mA. A Zener shunt regulator wastes (V_IN − V_Z) × I_total as heat 100% of the time. At 12V in, 5V out, 100mA load, that's 0.7W continuous in the series resistor and Zener - even when the load is idle. An LDO at the same conditions dissipates (12−5) × 0.1 = 0.7W only under load. At idle it drops to ~30µA × 7V = 0.2mW.
• Noise-sensitive analog. A Zener reference with a 100nF bypass cap typically measures 150–250µVrms in bench testing. A jellybean AMS1117 runs ~100µVrms (estimated - the datasheet doesn't spec noise). A low-noise LDO like the LP5912 delivers 12µVrms per the TI datasheet (SNVSA77). For a 16-bit ADC where each LSB is 50µV, that's the difference between 0.25 LSB of supply noise and 4+ LSB. When your ADC costs $15, the $0.30 LDO is cheap insurance.
• Variable loads. A Zener shunt must be sized for the minimum load current - meaning at maximum load it may not have enough current to stay in breakdown. An LDO handles load steps from zero to full rated current without changing its regulation behavior.
• Low dropout. Need 3.3V from a 3.7V Li-Ion battery? A Zener at 3.3V can't regulate from 3.7V - it needs several volts of headroom. An LDO like the LP5912 regulates 3.3V from 3.5V input with 95mV dropout.

The Hybrid: Zener Pre-Regulator + LDO

Sometimes neither part alone fits. Example: a 36V industrial rail needs to power a 3.3V MCU pulling 200mA. An LDO rated for 36V input exists, but (36−3.3) × 0.2 = 6.5W of heat in a single package is a thermal nightmare. A switching regulator adds EMI and complexity.

The hybrid: a 16V Zener pre-drops 36V to ~20V, an LDO regulates 20V to 3.3V. Zener takes voltage stress, LDO handles regulation. Power splits: 3.2W in the Zener, 3.3W in the LDO. Still hot, but spread across two packages instead of 6.5W in one.

This isn't elegant. A buck converter would be 85%+ efficient and dissipate under 1W total. But if you can't use a switcher - EMI constraints, BOM restrictions, certification requirements - a Zener + LDO cascade is a legitimate, production-tested approach.

Quick Decision Table

The Bottom Line

Zeners aren't obsolete. They're the right tool when the job is clamping, crude referencing, or pre-regulation - anywhere precision doesn't matter and cost does. LDOs took over the regulation job because they're better at it: lower noise, higher efficiency at varying loads, tighter accuracy.

But the real skill isn't picking one over the other. It's knowing when to use both. A Zener gate clamp protecting a MOSFET, an LDO powering the ADC that's reading the current through that MOSFET - two $0.02 parts, two completely different jobs, one working circuit.

Related Products

• BZX84-C2V4,215 - 2.4V Zener, SOT-23, soft knee for analog compensation
• BZT52H-C16,115 - 16V Zener, SOD-123F, gate clamping and transient protection
• LP5912-3.3DRVR - 500mA low-noise LDO, 12µVrms, post-switcher cleanup
• Zener Diode Selection Guide - BZX84 vs BZT52H vs BZX55C