Every electronics course teaches that “a silicon diode drops about 0.7 V.” Then the first real measurement produces 0.62 V at low current, the lab power supply shows 0.85 V at high current, and the number stops behaving like the constant the textbook promised. The resolution is not that the book was wrong; it is that 0.7 V is a point on a curve, not a law. Forward voltage drop (VF) is set by the physics of the junction, and it moves with current exponentially, with temperature at a few millivolts per degree, and across materials by design. This article derives where the number comes from, shows why the familiar “knee” is really an exponential curve, and gives you the sheet for turning VF into a loss estimate without guessing.
Where 0.7 V Comes From: The Physics in One Equation
The forward voltage of a diode is the sum of two contributions: the built-in potential of the junction and the extra voltage needed to push a useful current through it. In the simplified form used for silicon:
V = Vb + (η · kT/q) · ln(I/Is)
- Vb is the built-in potential, set by the doping levels on either side of the junction — for silicon this is close to the 0.7 V you memorized.
- η is the ideality factor, typically 1–2 for a real junction.
- kT/q is the thermal voltage, about 0.026 V at room temperature.
- Is is the saturation current, a tiny material constant.
- I is the forward current.
The logarithm is the entire story. Because V depends on ln(I), the voltage changes quickly when current is small and slowly once current is large. That is why 0.5 V at 1 mA and 0.8 V at 1 A live on the same curve — a thousand times more current only moves the voltage by about 0.18 V at room temperature, which is the “pleasant surprise” hidden in the math. The built-in potential dominates, which is why every silicon diode looks like it drops “about 0.7 V” regardless of size or rating.
The practical lesson is that VF is a function, not a single number. Datasheets print the value at a stated test current precisely because the number is meaningless without the current. When a design article says “use 0.7 V,” it is using the room-temperature value at a modest current — a fine starting assumption and a wrong substitute for the curve when the operating current is far from the test point.
Plotting VF vs Current: Why the “Knee” Is Not a Switch
The classic VF-current curve looks like a sharp knee: flat near zero, then a steep climb. The dominant mental model, inherited from circuit diagrams, treats a diode like a switch that “opens” at some threshold. The knee is not a threshold — it is the steep part of an exponential, and the difference matters in real designs.
Below the knee, current rises rapidly for tiny voltage changes as the junction leaves the saturation region; this is the region where a meter reads 0.5–0.6 V on a silicon part. Above the knee, the curve flattens because the logarithm compresses: to double the current you pay only the same tens-of-millivolts increment. The visual “knee” is simply where the exponential changes from rising sharply (as a fraction of voltage) to growing slowly, and there is no discontinuous switch at any point.
The misconception shows up in practice as design errors around low current. A circuit that runs a diode at microamps reads a lower VF than the datasheet’s test current suggests, so a “0.7 V” budget drawn from the datasheet overestimates the real drop — usually a safe error. The reverse error is dangerous: a datasheet test at 100 mA is often quoted as “the” forward drop, and a design running 10 A sees a real drop 0.2–0.3 V higher than budgeted, under-sizing the thermal allowance. Reading the curve at the operating current, not the test current, is the fix.
The −2 mV/°C Story: How Junction Temperature Rewrites Your Reading
Forward voltage also moves with temperature, and the sign surprises people: heating a silicon PN diode lowers its forward drop. For ordinary silicon rectifier diodes at moderate current, the coefficient is roughly −2 mV per °C — a useful engineering approximation, not a universal law — so a junction at 100 °C reads about 0.15 V lower than the same junction at 25 °C by that estimate. The value varies with device type, construction, and operating current, so the datasheet’s own curve at the working current is the first reference; Schottky, SiC, and LED devices should not be forced onto the same coefficient.
The mechanism is the thermal voltage term in the equation. As temperature rises, kT/q grows and the saturation current Is climbs steeply, so a given forward current is reached at a smaller applied voltage. The effect is consistent, measurable, and often the largest correction you will make between a bench measurement and a datasheet figure, because datasheets are published at 25 °C and operating junctions rarely are.
A concrete VF-at-temperature example ties the pieces together. A datasheet states VF = 0.72 V at 1 A and 25 °C. The design runs 1 A continuously in a 70 °C enclosure, and the junction settles near 110 °C. The temperature shift is (110 − 25) × (−2 mV) = −0.17 V, so the real VF at operating temperature is about 0.55 V — 0.17 V lower than the headline. The conduction loss drops from 0.72 W to 0.55 W, a 24% reduction that changes both the efficiency estimate and the thermal loop. Designs that skip the correction build in a hidden error of that size, which is why the temperature rule is not a fine point but a first-order correction in any hot design.
For design, the correction cuts both ways. In thermal calculations, a warmer junction means lower VF and therefore lower conduction loss — which is why Schottky parts can enter a mild self-stabilizing loop instead of runaway at moderate temperatures (the runaway story belongs to leakage, not to VF). But the temperature coefficient also means that measuring VF on a hot part and comparing it to a 25 °C datasheet curve produces an apparent shortfall that is pure physics, not a fault. The measurement trap is real enough to deserve its own rule: record junction temperature alongside every VF reading, and always compare at the same temperature.
The same curve-reading discipline applies to comparing parts. Two diodes with identical 25 °C VF at 100 mA can differ at 10 A by more than 0.1 V because their ideality factors and package temperatures diverge; a comparison table built from headline VF alone cannot see that. The honest method is to read both curves at the operating current and temperature, which is why the datasheet comparison guide treats test conditions as part of the specification rather than as fine print.
Low-VF Families: Schottky, Trench, and SiC SBD at a Glance
The same physics produces different numbers across materials because the built-in potential differs. A Schottky junction uses a metal-semiconductor barrier with a lower built-in potential, so its VF lands below 0.5 V at moderate current; the identical exponential structure applies, but the whole curve sits lower. A trench Schottky modifies the geometry to reduce reverse leakage while keeping the low VF, and a SiC Schottky barrier diode (SBD) extends the low-drop behavior to high voltage and high temperature where a silicon Schottky would not survive.
These low-VF families matter because conduction loss scales as VF × I. On a high-current rail, shaving 0.4 V off the forward drop saves real watts and a smaller thermal spread. The comparison between silicon, Schottky, and germanium for low-voltage rails covers when the trade is worth making, and the wider forward-voltage picture feeds directly into the loss calculations used by selection articles.
The caution is the flip side: lower VF usually trades against something. Schottky and SiC SBD parts carry higher leakage and lower reverse-voltage headroom than a silicon PN part of the same size, so the low-drop advantage is only a win when the rest of the duty — reverse blocking, standby leakage, temperature — does not cancel it. Choosing between the families is a five-dimension comparison, not a single-number race.
One-Page Sheet for Estimating Conduction Loss
The same physics are explored from the datasheet and application side in the forward voltage drop article, which pairs with this page’s derivation as the practical companion. For design work, the forward voltage becomes a loss number through the simplest equation in power electronics:
P = VF × I
where P is the conduction loss, VF is the forward drop at the operating current and temperature, and I is the average forward current. A 10 A rectifier at 0.8 V dissipates 8 W of conduction loss; at 0.5 V the same current dissipates 5 W. The equation is the start of every thermal and efficiency calculation, and it is the term that low-VF parts are bought to reduce.
The correct VF to use is the curve value at your operating current, corrected for your junction temperature, not the datasheet headline. The correction sequence: find VF at the operating current on the curve, apply the silicon −2 mV/°C approximation for the difference between operating temperature and 25 °C (or better, the datasheet’s coefficient), then multiply by the average current. The output of this reading feeds the 250 V loss worksheet, which applies the identical VF-and-temperature method to a real output stage, and the thermal design guide takes the number from there. The error bounds come from the same sources the datasheet comparison guide points to — test conditions, curve resolution, and thermal uncertainty — so a hand calculation is best treated as a bracket: compute with low and high VF and design the thermal margin to the high number. The loss figure this sheet produces is exactly the input the conduction-loss worksheet and thermal design guides expect, which is why the same method anchors the broader loss-calculation hub.