Measuring Diode Operating Points: What Forward Drops Reveal in a Real Circuit

In our electronics lab we built a diode network in Proteus, ran it, and recorded the operating points. The numbers in our report: D1 and D2 sit at 0.34 V each, D3 at 0.68 V, D5 at 0.72 V, and D4 is completely off, carrying nothing. Branch currents came out in the milliamp range (3.1 mA, 3.7 mA, 9.3 mA). Those numbers are more instructive than the schematic, because every one of them is the diode equation arguing with the rest of the circuit and winning by a specific amount.

Two diodes in series are not two drops of the same size

Why does D3 sit at 0.68 V while D1 manages only 0.34 V? Current. The branch carrying D3 pushes roughly three times the current of the D1 branch (9.3 mA against 3.1 mA), and a diode's forward voltage grows logarithmically with current: each decade of current buys about 60 mV at room temperature. Ten times the current should mean roughly 60 mV more, and 0.68 against 0.34 is bigger than that because the 0.34 V reading belongs to a diode barely nudged into conduction, where the exponential is still in its soft knee. The diagram-level model ("a diode is 0.7 V") is a bookkeeping convention for a strongly-on diode. The measurements show the real curve underneath it.

The diode that isn't there

D4 off is the most instructive reading of all. Reverse-biased or starved, it carries no current, and the circuit quietly re-routes around it. That is the binary behavior people learn first and then misuse: real circuits are full of diodes that are off, and analysis means determining which ones conducted before computing anything. The habitual check from that lab (assumed state, verify against current direction and magnitude, revise) is exactly the iteration you later do with transistor biasing, where the same discipline scales from two states to a load line.

Watching clipping on a scope

The same lab's clipping-and-clamping experiment put a sine wave through a diode limiter and photographed the oscilloscope: 5 V per division, 10 ms per division, and the tops of the sine visibly shaved flat at the clamp level. The flat top is the whole story of a clipper. While the source swings below the diode's conduction threshold, the output tracks it faithfully; the instant it exceeds the threshold plus the diode drop, the diode conducts and the output is pinned. Transfer a sine through that input-output relationship and you get a waveform with its peaks surgically removed.

Clamping is the complement: a capacitor and diode shift the entire waveform by a DC offset instead of removing anything. Same two components, opposite jobs, and the scope screenshots make the difference impossible to confuse.

The debugging story worth keeping

Our group's first simulation runs failed for the least glamorous reason possible: a missing ground. The circuit was correct component-wise, the solver refused to produce sensible node voltages, and the fix was one reference node. Every SPICE-class tool since has taught me the same thing in fancier language: node voltages are differences, and a simulation without a defined reference is not a hard circuit problem, it is an undefined one. The report's discussion section records that failure on purpose.

The takeaway

Diode labs look trivial: one nonlinear part with a two-line model. But the measured numbers (0.34, 0.68, 0.72, one off) are the diode equation, the load line, and the rest of the circuit negotiating in public. Learning to read that negotiation, expected state, then actual operating point, then the difference, is the core skill of every electronics lab that came after it.