Serial Communication Explained: From UART to Electrical Interfaces
This article clarifies the layered concepts behind "serial port" communication, covering how UART frames bytes into asynchronous serial data, how TTL/CMOS logic levels and RS-232/RS-422/RS-485 electrical standards differ, why USB-to-serial adapters create COM ports, and how to wire, troubleshoot, and select the right interface for your application.
1. First, Clarify the Layers: “Serial Port” Is Not a Single Fixed Interface
When discussing serial ports, the most common mistake is comparing concepts that belong to different layers. Questions like “What’s the difference between UART and RS-485?” or “Is a USB port a COM port?” cannot be answered with a simple yes or no.
To understand these terms, break a communication session into five questions:
- What object does software use to access the device? This involves system port names such as COM ports.
- How is data organized into a continuous stream of bits? This involves UART’s asynchronous transceiver mechanism and frame format.
- How do bits become electrical signals on a wire? This involves logic levels, RS-232, RS-422, RS-485, and so on.
- What connector and cable are used? This involves pin definitions, connectors, and wiring.
- What does the received data mean? This involves application protocols, device addresses, commands, lengths, and checksum rules.
Therefore, determining whether two devices can interoperate cannot rely solely on whether both are labeled “serial port”—you must verify each layer individually.
2. What Do the Common Terms Actually Mean?
| Name | Layer or Category | Core Role |
|---|---|---|
| Serial port | Generic term | An interface that transmits data bit by bit over time; in engineering practice it usually refers specifically to asynchronous serial interfaces |
| UART | Hardware peripheral or functional module | Performs asynchronous serial transmission and reception |
| USART | Hardware peripheral or functional module | Supports specific synchronous communication modes in addition to asynchronous communication |
| COM port | Port identifier in the operating system | Provides applications with an access entry point to serial devices |
| USB | Bus architecture | Defines connection, power delivery, enumeration, and data transfer mechanisms |
| TTL, CMOS | Logic circuit and level-related concepts | Determine how digital inputs and outputs recognize and generate logic states |
| RS-232 | Serial interface standard | Defines single-ended signal electrical characteristics and related interface requirements |
| RS-422 | Differential electrical interface standard | Uses balanced differential transmission, suitable for corresponding point-to-point or multi-receiver scenarios |
| RS-485 | Differential electrical interface standard | Supports multi-drop bus connections, suitable for industrial field communication |
Notably, COM is simply a naming convention for communication ports; it should not be interpreted as referring to any specific connector or voltage standard. It is first and foremost a name from the software access perspective.
Also keep in mind that serial communication in the broad sense includes not only UART but also other bit-by-bit buses. It’s just that in everyday engineering conversation, “serial port” usually carries the narrower meaning.
3. What UART Does: Convert Bytes into Asynchronous Serial Frames
UART stands for Universal Asynchronous Receiver/Transmitter. It is typically integrated into microcontrollers, processors, or bridge devices, and can also exist as a standalone functional module.
When transmitting, the UART takes bytes handed over by software and outputs the bits sequentially according to a preset format; when receiving, it reconstructs bytes from changes in the input signal.
3.1 Why Asynchronous Communication Needs No Clock Line
A typical UART connection includes a transmit line (TX), a receive line (RX), and a ground reference (GND). The two sides do not transfer a separate clock; instead, they agree in advance on the baud rate and frame format.
After detecting the start bit, the receiver samples at expected positions based on its local clock. Therefore, the clock error between the two sides cannot grow without bound. The tolerable error depends on frame length, sampling mechanism, and device implementation—you cannot assume every UART tolerates the same percentage of deviation.
3.2 What a Frame Contains
A typical UART frame consists of:
- Start bit: marks the beginning of a frame;
- Data bits: carry the payload;
- Optional parity bit: provides limited error detection;
- Stop bit(s): marks the end of the frame and provides an interval before the next frame.
The common “8N1” notation means 8 data bits, no parity, and 1 stop bit. For the common non-inverted logic-level UART, the idle state is high and the start bit is low.
UART is a transceiver function and its common communication method—it is not a complete communication protocol covering all layers. It does not define business commands, device addresses, or the meaning of data fields.
3.3 Baud Rate Does Not Equal Effective Byte Rate
For the common two-level UART, each symbol carries 1 bit, so the baud rate value generally equals the line bit rate.
With the 8N1 format, transmitting 1 byte requires:
1 start bit + 8 data bits + 1 stop bit = 10 bits
Assuming a line rate of 115200 bit/s, and ignoring inter-frame gaps, the theoretical effective data rate is:
115200 ÷ 10 = 11520 bytes/second
If upper layers add overhead such as addresses, lengths, checksums, and acknowledgments, the actual payload throughput drops further.
4. TTL and CMOS: “3.3V Serial” Is Not Universally Compatible
UART describes the transceiver function, but the voltage on a chip’s pins is determined by its I/O circuitry.
TTL stands for transistor-transistor logic; CMOS stands for complementary metal-oxide-semiconductor logic. Both are primarily logic-circuit concepts, not just labels for “how many volts a high level is.”
Today, most devices use CMOS processes, yet engineers still habitually call the low-voltage serial port on chip pins a “TTL serial port.” This naming is convenient for communication but is not sufficient as a basis for wiring.
4.1 Check Input/Output Thresholds, Not Just Supply Voltage
To determine whether two pins can be connected directly, you must at least verify:
- The transmitter’s guaranteed minimum output high level;
- The transmitter’s guaranteed maximum output low level;
- The receiver’s guaranteed input lower bound recognized as high;
- The receiver’s guaranteed input upper bound recognized as low;
- The receiver’s allowable input voltage range.
A classic 5V TTL interface example: input high no lower than 2.0V, input low no higher than 0.8V; guaranteed output high no lower than 2.4V, guaranteed output low no higher than 0.4V. These values come with specified test conditions and cannot be generalized to all low-voltage logic interfaces.
CMOS thresholds are often related to the supply voltage, but the exact ratios must be taken from the device datasheet—do not treat any empirical ratio as a universal standard.
4.2 “Recognizable” and “Won’t Be Damaged” Are Two Different Questions
A 3.3V output may be recognized as a high level by a 5V input, but that does not mean the reverse connection is equally safe. Connecting a 5V output to a 3.3V input without appropriate voltage tolerance can cause abnormal current flow or even damage the device.
Therefore, use a proper level-shifting scheme when connecting across voltage domains. Also consider power-down states: if one side is powered off while the other continues driving high, reverse powering may occur through input protection paths.
5. RS-232: Cannot Connect Directly to Logic-Level UART
RS-232 typically uses single-ended transmission relative to a signal ground, and its voltage ranges and logic polarity differ from common low-voltage UART pins.
On RS-232 data lines, a negative voltage corresponds to logic 1 and a positive voltage corresponds to logic 0. Receivers typically recognize regions below −3V and above +3V as valid logic states, while the region in between should not be relied upon for a valid determination.
This means:
Low-voltage UART TX and RX pins must not be connected directly to RS-232 signal lines.
The correct link should be:
UART logic side ⇄ RS-232 transceiver ⇄ cable ⇄ RS-232 transceiver ⇄ UART logic side
The transceivers handle voltage conversion, polarity inversion, and line driving.
RS-232 is mostly used for point-to-point connections, and many devices also use hardware flow control signals. When connecting, do not judge pin functions by connector shape alone, and do not mechanically assume that “same-numbered pins connect together” is correct. Verify the transmit, receive, signal ground, and flow control definitions based on the role of each end.
6. RS-485 and RS-422: Differential Transmission Does Not Mean Automatically Reliable
Differential transmission conveys logic states through the voltage difference between a pair of wires. Compared with observing a single signal wire’s voltage relative to ground, it generally offers better rejection of interference superimposed on both wires.
However, this capability is limited by the transceiver’s allowable common-mode range, cable balance, routing, and grounding conditions—it is not immune to field conditions.
6.1 Why RS-485 Suits Multi-Drop Buses
RS-485 transceivers typically have a controllable driver enable, allowing multiple nodes to share one bus. The common two-wire configuration uses half-duplex communication: only one node may drive the bus at any given time.
The transmit sequence is typically:
- Confirm the current node has permission to transmit;
- Enable the transmit driver;
- Output the complete data frame;
- Wait until the last stop bit has truly left the UART;
- Disable the transmit driver and release the bus.
Here you must distinguish between “transmit buffer empty” and “transmission complete on the line.” If the driver is disabled too early, the last few bits may be truncated.
RS-485 itself does not handle node addressing, collision arbitration, or retransmission. When multiple devices are allowed to transmit must still be defined by an upper-layer protocol.
6.2 RS-422 vs. RS-485 Cannot Be Distinguished by Wire Count Alone
RS-422 typically targets a differential link with one driver connecting to one or more receivers. RS-485 supports multi-driver buses subject to the relevant constraints.
Two-wire half-duplex RS-485 is common in practice, and corresponding four-wire connections can also be built using two differential pairs—but the specific topology, node roles, and bidirectional communication method must be designed separately. Do not simply equate “two wires” with RS-485 and “four wires” with RS-422.
6.3 Distance and Speed Must Be Considered Together
A differential interface does not maintain its maximum speed at arbitrary distances. Communication capability is simultaneously affected by cable attenuation, propagation delay, signal edges, stub length, node loading, and the noise environment.
Therefore, “maximum distance” and “maximum speed” must not be quoted together without conditions. Field selection should be validated against the cable, topology, and target bit error rate.
7. USB and COM Ports: One Is a Bus, the Other Is a Software Entry Point
USB is not simply UART’s TX and RX with a different plug. It has its own device discovery, configuration, and data transfer mechanisms, and cannot be connected directly to ordinary UART pins.
A so-called “USB-to-serial” adapter actually converts between the two communication mechanisms via a bridge function.
Common link configurations include:
USB ⇄ bridge ⇄ low-voltage UARTUSB ⇄ bridge ⇄ RS-232 transceiverUSB ⇄ bridge ⇄ RS-485 transceiver
All three may appear in the operating system as an accessible serial port, but the electrical characteristics on the line side are completely different.
Therefore, seeing a COM port only tells you that software has obtained a serial access entry point—it does not tell you whether the external interface is 3.3V, 5V, RS-232, or RS-485. Some virtual serial ports have no actual UART line at all, and their baud rate settings may not correspond to any physical line rate.
8. Practical Wiring: Handle Each Interface Type Separately
8.1 Low-Voltage UART Connection
Typically wired as:
Device A TX ─── Device B RXDevice A RX ─── Device B TXDevice A GND ─── Device B GND
Before connecting, you must confirm voltage compatibility between the two ends. A power pin on a module is not a required communication signal—do not connect the power rails of both sides directly without understanding the power direction.
8.2 RS-232 Connection
Verify the connector pinout, device roles, and flow control configuration, and ensure signals pass through matching transceivers. Never feed its voltages directly into low-voltage logic pins.
8.3 RS-485 Connection
Connect the differential pair according to the polarity definitions in the documentation at both ends. Naming conventions for the A/B labels may be inconsistent across documents, so do not judge by the letters alone.
Buses typically use a trunk structure, with stubs kept as short as possible. Whether to add termination resistors, and what value to choose, should be determined by transmission line characteristics, signal edges, and topology. A common practice is to terminate at the two physical ends of the bus with the cable’s characteristic impedance—not to add termination at every node.
Idle biasing also needs unified design to avoid extra loading from bias at multiple points. Receivers with internal failsafe capability can improve output determinism under specific input conditions, but they cannot replace a complete bus design.
Differential communication also requires attention to reference potential. When facing large ground potential differences, evaluate isolation, reference connection, and protection schemes—do not simply assume “two wires means ground doesn’t matter.”
9. From Garbled Output to No Response: An Effective Troubleshooting Sequence
Step 1: Check Electrical Safety
First verify the interface type, voltage range, power state, and grounding method. Signals with negative or higher voltages must not be connected to ordinary digital inputs or measurement equipment without confirmation.
Step 2: Check Link Direction
Confirm TX connects to the peer’s RX, RS-485 polarity matches, and connector pins agree with the documentation.
Step 3: Unify UART Parameters
Check baud rate, data bits, parity, stop bits, and flow control item by item. Do not assume the configuration matches just because the baud rates are the same.
Step 4: Inspect the Actual Waveform
Use measurement equipment suitable for the voltage range under test, and observe the idle level, start bit, bit width, and edge quality. Garbled output in software only proves the decoded result is abnormal—it does not directly indicate which layer is faulty.
Step 5: Verify the Upper-Layer Protocol
If the physical waveform and UART frames are both normal but the device does not respond, check the device address, command bytes, length, byte order, checksum fields, and response timeout.
A serial port delivers a data stream; the amount of data an application reads in one call does not necessarily equal one frame. Programs should buffer input, parse it according to frame headers, lengths, or delimiter rules, and be able to resynchronize when checksums fail.
Step 6: Validate Under Real Operating Conditions
Success with short cables, a single node, and low speed does not prove the system is reliable. Gradually increase cable length, node count, communication frequency, and field interference, while checking error counts, timeout rates, and recovery mechanisms.
10. How to Quickly Choose the Right Communication Method
- Within the same board or between adjacent boards: Consider low-voltage UART first, but confirm level compatibility, ground reference, and signal integrity.
- Connecting to existing RS-232 equipment: Use a matching RS-232 transceiver link and verify pinout and flow control.
- Longer distances or multiple field nodes: Consider RS-485, while designing the bus topology, termination, biasing, protection, and access rules.
- Connecting a computer to an external serial device: First determine the device’s line-side interface, then choose the corresponding USB bridge type—do not decide based only on whether a COM port appears.
In the end, UART handles asynchronous transmission and reception, the COM port is a software access entry point, USB is a separate bus mechanism, TTL and CMOS relate to logic circuits and levels, while RS-232, RS-422, and RS-485 solve the line-side electrical transmission problem. Only by separating these layers and matching them item by item can you truly answer “Can these be connected?”, “How should they be wired?”, and “Why is no data being received?”
