An open drain is a type of digital output found on microchips where the output pin can only pull the signal to ground or disconnect entirely. Unlike a standard output that can actively drive both high and low voltage levels, an open drain relies on an external resistor to produce a high signal. This design choice enables several capabilities that standard outputs can’t match, including connecting multiple devices to the same wire, mixing different voltage levels, and safely sharing a communication bus.
How an Open-Drain Output Works
Inside an open-drain output, there’s a single transistor (specifically an N-channel MOSFET) with its drain terminal exposed as the output pin. This transistor has two states. When it turns on, it creates a direct path between the output pin and ground, pulling the voltage low. When it turns off, the output pin is completely disconnected from everything inside the chip, a state engineers call “high impedance” or “Hi-Z.” In this disconnected state, the pin is essentially floating, neither high nor low.
That floating state is the key detail. Because the chip can’t actively drive the output high, you need an external pull-up resistor connected between the output pin and your desired high-voltage rail. When the transistor is off, the resistor gently pulls the voltage up to its high level. When the transistor turns on, it overpowers the resistor and yanks the line down to ground. So the two logic states become: transistor on equals low, resistor pulling up equals high.
Open Drain vs. Open Collector
You’ll often see the terms “open drain” and “open collector” used interchangeably, but they refer to slightly different technologies doing the same job. Open drain describes circuits built with MOSFETs, where the exposed terminal is the transistor’s drain. Open collector describes the same concept built with older bipolar transistors, where the exposed terminal is the collector. The behavior is identical: the output can sink current to ground or go high-impedance. The terminology just depends on which type of transistor is inside the chip.
How It Differs From Push-Pull
The most common alternative is a push-pull output, which has two transistors: one that connects the output to the positive supply and one that connects it to ground. A push-pull output can actively drive the line both high and low without any external resistor. This makes it faster, since both transitions are actively driven. Push-pull outputs generally switch faster than open-drain configurations because there’s no RC delay from a pull-up resistor charging line capacitance.
The trade-off is flexibility. Push-pull outputs cannot safely be wired together on a shared line. If one chip tries to drive the line high while another drives it low, you get a direct short from the power supply to ground through both transistors. This can draw enormous current and potentially damage the chips. Open-drain outputs avoid this entirely: when multiple open-drain outputs share a wire, any device can safely pull the line low while the others simply float.
Why Open Drain Needs a Pull-Up Resistor
Without a pull-up resistor, an open-drain output has no way to produce a defined high state. The pin just floats, picking up noise and creating unpredictable readings for whatever device is listening. The pull-up resistor is what gives the circuit its “high” logic level.
Choosing the right resistor value involves balancing a few competing goals. A smaller resistance pulls the line high faster, giving you cleaner signals at higher speeds. But when the transistor turns on and pulls the line low, all the current flowing through that resistor gets dumped to ground, increasing power consumption. A larger resistance saves power but creates slower rise times and makes the line more vulnerable to noise interference. For most applications, values between 1kΩ and 10kΩ work well, but the exact choice depends on the speed of your signal and how many devices share the line.
There’s also a maximum resistor value to watch for. If the resistor is too large, the voltage at the output when floating won’t reach a level that the next chip recognizes as a valid logic high. That threshold varies by chip, and exceeding it means your signal quietly stops working.
Voltage Level Shifting
One of the most practical advantages of open drain is the ability to bridge different voltage domains. Because the output is just a transistor pulling to ground, and the pull-up resistor sets the high voltage, you can connect the pull-up to a completely different voltage than the chip runs on internally. A 3.3V chip can communicate with a 5V chip simply by connecting the pull-up resistor to the 5V supply. The open-drain output swings between ground and 5V at the output, even though the chip’s internal logic runs at 3.3V.
This works because the output transistor only ever connects to ground. It doesn’t care what voltage the pull-up resistor is connected to, as long as it’s within the maximum voltage rating listed on the chip’s datasheet. This makes open drain a natural fit for mixed-voltage systems without needing a dedicated level-shifting chip.
Wired-AND Logic
When you connect multiple open-drain outputs to the same wire with a single shared pull-up resistor, you get a useful behavior called wired-AND logic. The shared line is high only when every device lets go of it (all transistors off, all high-impedance). If any single device pulls the line low, the whole line goes low. This creates an AND function without any additional logic gates:
- All devices release the line: pull-up resistor brings it high
- Any one device pulls low: the entire line goes low
This behavior is the foundation of shared communication buses and interrupt lines, where multiple chips need to signal on the same wire without conflicting with each other.
I2C and Other Bus Protocols
The most well-known use of open-drain outputs is the I2C communication protocol, used in countless sensors, displays, and peripherals. Both of I2C’s signal lines (the data line and the clock line) use open-drain connections for every device on the bus. This design is what makes I2C possible: dozens of devices can share just two wires.
Because all connections are open-drain, any device that pulls a line low simply wins. There’s no destructive conflict. If two devices try to communicate at the same time, the one sending a low signal naturally takes priority, and the other device can detect this and back off. With push-pull outputs, the same situation would create a short circuit, potentially drawing enough current through the transistors to damage the chips. Open drain turns what would be a hardware failure into a graceful conflict-resolution mechanism.
Beyond I2C, open-drain outputs are common for interrupt lines (where multiple chips share a single interrupt pin to the processor), power-good indicators (where a chip signals whether its output voltage is stable), and any scenario where multiple devices need to share a single signal line safely.
Performance Trade-Offs to Know
Open drain’s flexibility comes with real costs. The biggest is speed. Because the pull-up resistor and the capacitance of the wire form a low-pass filter, the rising edge of the signal is always slower than the falling edge. The transistor can yank the line to ground quickly, but the resistor has to charge up the line capacitance gradually. In high-speed designs, this asymmetry limits how fast you can toggle the signal.
Power consumption can also be higher in some operating states. Whenever the line is held low, current flows continuously through the pull-up resistor to ground. If the line spends significant time in the low state, that static current adds up. Push-pull outputs don’t have this problem because their transistors block current flow in both states.
Noise immunity depends directly on your pull-up resistor choice. A weaker (higher value) pull-up means a weaker high signal, making the line easier for external interference to corrupt. In electrically noisy environments, you may need a stronger pull-up, which circles back to higher power consumption. These trade-offs are the reason open drain is preferred for slower, shared buses like I2C rather than high-speed point-to-point connections where push-pull outputs excel.