Definition
OCP (Over Current Protection) refers to the power module automatically limiting current or cutting off output when the output current exceeds the rated value, preventing the power supply from overheating and the load from short circuiting and enlarging the fault. Industrial-grade DC-DC power supplies are usually used together with OVP (overvoltage protection), short-circuit protection, and reverse connection protection, meeting the power supply reliability requirements for automotive, outdoor, industrial, and other scenarios.
Implementation Methods
OCP can detect output current through current-sensing resistors, Hall sensors, etc., and trigger current limiting or shutdown after comparing with the threshold. Current-limiting type: reduces output voltage when current exceeds the limit to maintain constant current; Shutdown type: cuts off output when exceeding the limit, requiring restart or fault removal to recover. Industrial power supplies often use hiccup mode: periodically attempt to restart during a short circuit, automatically recovering after the fault is cleared. Combined with OVP, it can fully protect the power supply and load, meeting EMC and reliability certification requirements.
Use Cases
Power supply for industrial gateways, RTU, DTU and other devices; harsh environments such as vehicles, outdoors, and underground; each channel OCP for multi-output power supplies. When selecting, it is necessary to confirm the OCP threshold, recovery method, and compatibility with load characteristics.
Current Knee and Recovery
After the sample is compared, limit or shut off
Overcurrent protection watches output current, not a fuse on the input. A sense resistor or a Hall element turns current into voltage, and a comparator tells the control loop when the knee is passed. A limiting type pulls the output voltage down so current stays near the knee. During a short the supply is still heating, but the heat is bounded. A shutdown type opens the switch. Heating stops, and the load loses power. A hiccup type waits a short time after shutdown and tries again. If the fault is still there it shuts off again. If the fault is gone it returns to regulation on its own. The three actions mean different things to the load. The selection has to name which one, not only "overcurrent protection included."
The knee is usually set above rated current, leaving margin for steady state, and it must stay below what the printed trace and the connector can carry for a long time. A threshold too close to the rating trips on small swings of a normal load. A threshold set too high lets the cable fail before the supply. Temperature drift of the sense resistor walks the knee with temperature. Recheck a supply in a closed cabinet at the highest ambient temperature.
Inrush needs its own margin
Capacitive loads, disk-style storage, radio modules, and relay coils draw far more current in the instant of power-up than their average a few milliseconds later. If protection has no separate startup window, every cold start is treated as a short. The right use is a steady-state knee with margin over the long-term load, a separately specified peak and duration for startup, and confirmation that downstream equipment can accept the voltage sag while current is limited. With several outputs, each channel's knee and the total power limit have to hold at the same time. Total current alone is not enough.
Outdoor and vehicle sites also have to consider input current when input voltage is low. Output power unchanged, a sagging input voltage raises current on the front end, and input-side protection may trip first. The symptom is "the output is not shorted, but the whole unit restarts." Include cable loss, the fuse, and the transmit pulse of a module such as a DTU in that sum. Do not look only at the output rating on the supply nameplate.
Hiccup mistaken for latch-off
In maintenance it is easy to read hiccup as a failed supply. The output kicks every so often on a scope because the supply is testing whether the short is still there. Replacing the supply then leaves the fault in the load or the wiring. Latch-off protection, by contrast, stays off after one overcurrent event until someone removes power and applies it again. With nobody on site, one transient leaves the station dark. Write the recovery policy as an operating requirement. An unattended station more often needs to come back by itself once the fault is gone. A critical control loop often should not be hammered repeatedly while a short is still present.
Another trap is a protection event that is never reported. While the supply limits current, the downstream gateway is undervoltage, communication drops, and the center only sees a device offline. It does not see overcurrent. If the supply offers a fault contact or a status pin, bring it into a discrete input on the RTU or the gateway, and separate "protection has tripped" from "the link has failed." Do not raise the fuse rating in place of adjusting the knee. A fuse is slow and will not protect the semiconductor switch.
When several devices share one supply, leave the knee for simultaneous startup, not for average current. If a radio transmit, a relay pull-in, and capacitor charging land in the same millisecond, the stacked current crosses a threshold set from the average and the whole cabinet restarts. Staggering the high-inrush loads, or giving that branch its own supply, is safer than lifting the knee without a ceiling. Raising the knee spends the margin of the cable and the connector. On a real short, protection arrives too late and the fault spreads farther.