An H-bridge drive for a motor has two hidden electrical paths: the switching path that everyone sizes, and the freewheeling path that nobody does. When the H-bridge releases a motor, the back-EMF drives current through the Zener or the protection diodes that absorb the stored energy — and that reverse path, repeated thousands of times, carries current, heat, and wear that a static rating never shows. The failure is not the forward rating; it is the reverse-current duty, the repetition, and the heat on the freewheeling loop. This article names the path, quantifies the stress, explains the Zener-controlled overshoot, gives the rating checklist, and works the small-motor, seat, and steering examples.
The Freewheeling Path Nobody Sizes First
The motor in an H-bridge stores energy when it runs, and when the bridge releases or reverses it, that stored energy has to go somewhere. The freewheeling path is the route the back-EMF takes: through the protection diode or the Zener that clamps the released voltage. In a symmetric bridge, the path appears on every one of the four legs depending on the drive state, and its current is the motor’s stored energy discharging through the protection element.
The sizing error is viewing the protection element as a static clamp and missing that it is a real current-carrying path under the motor’s dynamic load. A Zener or diode chosen on its forward rating alone can fail on the reverse freewheeling duty, because that duty is repetitive and hot. The motor-drive failure article and the flyback concept frame the energy that the protection element must absorb, and this article adds the reverse-path sizing that both assume.
A worked freewheel calculation makes the sizing concrete. A seat motor stores 0.5 J when it runs, and on release the stored energy discharges through the Zener at its breakdown. If the Zener holds 24 V, the event dumps the stored energy as a current of roughly 0.5 J divided by the discharge time; a 5 ms discharge gives an average near 40 mA at 24 V, a modest single-event energy of 0.5 J. The power during the event is about 1 W, and the Zener’s junction has ample headroom for a single event. Now apply the same arithmetic to a steering assist that repeats the event 10 times per second: the average power becomes about 10 W, the Zener must shed that heat continuously, and a part that survived the single event comfortably will overheat on the repetition. The example shows why the repetition term, not the event peak, usually sets the freewheeling Zener size in automotive duty.
The three duty types belong in a table because each dominates a different rating:
| Duty | Event current | Repetition | Dominant term | Zener sized by |
|---|---|---|---|---|
| Cooling fan | Modest | Moderate | Peak and cooldown | Reverse current |
| Power seat | High | Rare | Single-event energy | Peak pule, energy |
| Steering assist | Moderate | Very high | Repetition and heat | Average power, duty |
The table is the freewheeling duty in one view: the fan is an easy peak-limited case, the seat is a high-energy rare case, and the steering assist is a repetition-and-heat case that dominates by average power. A designer who names the duty and its dominating term before sizing the Zener has already avoided the error the whole article is about.
Current, Repetition, and Heat on the Reverse Path
The reverse-path stress has three terms, and each must be sized. The current per event is set by the stored motor energy and the loop impedance — the back-EMF discharging through the protection element. The repetition is set by the drive pattern: a steering assist that reverses thousands of times per hour runs the path far harder than a seat that moves once. The heat is the integral of the current times voltage across the repeated events, and it accumulates in the protection element and its mounting.
The three terms compound. A path that carries a modest current once is trivial; the same path carrying that current a thousand times an hour, in a warm enclosure, heats the protection element past its rating. The electric power steering article is the clearest example of the repetition-driven stress, and the rectifier thermal guide closes the heat loop that the repetition produces.
Adding a Zener to the Loop: Controlled Overshoot
A Zener in the freewheeling loop adds controlled overshoot. A plain diode clamps the released voltage to near zero, discharging the stored energy slowly; a Zener clamps at a controlled level above zero, so the motor releases faster and the overshoot is bounded to the Zener voltage. The designer chooses the Zener voltage as the trade between release speed and spike magnitude.
The controlled-overshoot logic is the same trade the relay flyback article makes with a diode-versus-Zener choice, applied to the motor’s stored energy. A higher Zener voltage releases the motor faster and lets more of the energy pass as a controlled spike; a lower one releases slower with less overshoot. The Zener’s reverse-current rating and its repetition capability now matter exactly as the forward rating would in a power stage, because it is carrying the freewheeling current on every cycle.
Ratings to Check: Forward, Reverse, and the Pulse Shape
The rating checklist for the freewheeling Zener has three parts. The forward rating is the small one — the Zener rarely conducts forward in this role. The reverse rating is the real check: the Zener must carry the freewheeling current at its breakdown voltage, repeated at the drive rate, without exceeding its junction limit. The pulse shape completes the picture: a sharp, high-current back-EMF pulse stresses the junction differently than a long, low one, and the datasheet’s pulse and repetition data are the numbers to compare.
The three checks are the difference between a Zener that clamps the motor and one that dies on the third freewheel. The Zener datasheet guide and the rectifier voltage ratings give the numbers, and the Zener categories list the parts whose reverse and pulse ratings fit the duty.
One design caution completes the picture: a Zener in the freewheeling loop does not make the reverse path optional to monitor. The Zener is carrying the motor’s energy on every release, so its case temperature under the real drive pattern is a health indicator — a steering assist whose Zener runs hotter than the design expected is a sign that the repetition is higher or the thermal path weaker than assumed. An IR probe on the Zener during a drive cycle is the cheapest reliability instrument on the bench, and it confirms the four-term sizing without a single further calculation.
Examples: Small Motor, Power Seat, and Steering Duty
Three worked examples close the sizing. A small cooling fan motor in an H-bridge freewheels every time the fan coasts on a control cycle; the event current is modest, the repetition is moderate, and a small Zener sized for the reverse current handles it in a cool chassis. A power seat motor moves rarely but with high current when it does; the freewheeling event is large but infrequent, so the duty is the peak pulse more than the repetition, and the Zener is sized for the single-event current. A steering assist reverses constantly when the vehicle turns, carrying a moderate current thousands of times an hour in a warm environment — the hardest duty, where the repetition and heat, not the peak, decide the part.
The three examples map the same checklist to three different dominating terms: the fan is an easy peak, the seat is a large peak, and the steering assist is a repetition-and-heat problem. The electric power steering article and the motor-drive failure article are the two field references, and the Zener families supply the parts whose reverse-current ratings are read against the real duty. The freewheeling path, sized for current, repetition, heat, and pulse shape, stops being the path nobody sized.
The pulse-shape check deserves a fuller note because it is the rating most often skipped. A motor back-EMF pulse is not a clean half-sine; it is a decaying discharge whose peak and duration depend on the motor’s inductance and the loop impedance. A sharp, fast pulse stresses the junction by its peak current, while a longer, lower one stresses the thermal path by its energy. The datasheet’s pulse-rating and repetition data must be matched to the pulse the motor actually produces, and the rectifier voltage ratings reading order is the discipline that checks the pulse against the rating rather than against a guess.
The closing rule for the freewheeling path is that it is a real electrical path with a real duty, and it deserves the same sizing attention as the forward switching path that everyone remembers. The reverse current, the repetition, the heat, and the pulse shape are the four terms that decide the Zener; a part sized on forward rating alone will survive the bench and fail in the first week of the real drive pattern. The electric power steering article and the motor-drive failure article are the two field records of that failure, and the Zener families supply the parts whose reverse ratings close the loop.