Understanding RS485: 18 Questions and Answers Explained

The RS485 interface consists of a half-duplex network, typically using a two-wire system, often employing shielded twisted pair cables for transmission. This wiring method allows a maximum of 32 nodes to be connected on the same bus in a bus topology. Initially, data was output as an analog signal for simple process quantities, and later, the RS232 interface was introduced for point-to-point communication. However, this method could not achieve networking capabilities, leading to the emergence of RS485, which solved this issue. This article provides a detailed introduction to the RS485 interface in a Q&A format.

Answer: The RS-232-C interface standard was established earlier and has several shortcomings, mainly as follows:

(1) The signal voltage levels of the interface are relatively high, which can damage the interface circuit chips. Additionally, because it is not compatible with TTL levels, a level conversion circuit is needed to connect with TTL circuits.

(2) The transmission speed is relatively low, with a baud rate of 20Kbps in asynchronous transmission.

(3) The interface uses one signal line and one signal return line, forming a common ground transmission method, which is susceptible to common-mode interference, resulting in weak noise immunity.

(4) The transmission distance is limited, with a maximum standard transmission distance of 50 feet, practically only usable for around 50 meters. To address the shortcomings of RS-232-C, new interface standards have emerged, one of which is RS-485, which has the following characteristics:

1) The electrical characteristics of RS-485: Logic “1” is represented by a voltage difference of + (2-6)V between the two wires; Logic “0” is represented by a voltage difference of – (2-6)V between the two wires. The interface signal level is lower than that of RS-232-C, making it less likely to damage the interface circuit chips, and this level is compatible with TTL levels, allowing for easy connection with TTL circuits.

2) The maximum data transmission rate of RS-485 is 10Mbps.

3) The RS-485 interface utilizes a combination of balanced drivers and differential receivers, enhancing common-mode noise immunity, resulting in good noise resistance.

4) The standard maximum transmission distance of the RS-485 interface is 4000 feet, practically reaching 3000 meters. Additionally, while RS-232-C only allows one transceiver to be connected on the bus, RS-485 allows for the connection of up to 128 transceivers, providing multi-station capabilities, enabling users to easily establish a device network using a single RS-485 interface.

5) Due to the excellent noise immunity, long transmission distance, and multi-station capabilities of the RS-485 interface, it has become the preferred serial interface. The half-duplex network formed by the RS485 interface generally requires only two wires, so shielded twisted pair cables are used for transmission. The RS485 connector uses a DB-9 9-pin plug, while the smart terminal RS485 interface uses a DB-9 (socket), and the keyboard connection uses a DB-9 (pin).

1. Balanced Transmission

Unlike RS-232, RS-422 and RS-485 use differential transmission, also known as balanced transmission, which employs a pair of twisted wires, with one wire defined as A and the other as B.

Typically, the positive level between the driver A and B is +2 to +6V, representing one logic state, while the negative level is -2 to 6V, representing another logic state. There is also a signal ground C; in RS-485, there is an ‘enable’ terminal, which is optional in RS-422. The ‘enable’ terminal is used to control the connection and disconnection of the transmitter driver and the transmission line. When the ‘enable’ terminal is activated, the transmitter driver enters a high-impedance state, referred to as the ‘third state’, which is distinct from logic ‘1’ and ‘0’.

The receiver is also defined relative to the transmitter, with the receiving and transmitting terminals connected correspondingly through the balanced twisted pair. When there is a voltage greater than +200mV between AB at the receiving end, a positive logic level is output; when it is less than -200mV, a negative logic level is output. The input voltage range of the receiver on the balanced line typically ranges from 200mV to 6V.

2. RS-422 Electrical Specifications

The full name of the RS-422 standard is ‘Electrical Characteristics of Balanced Voltage Digital Interface Circuits’, which defines the characteristics of the interface circuit. Figure 2 shows a typical RS-422 four-wire interface. In reality, there is also a signal ground line, totaling five wires. Figure 1 defines the pin configuration of its DB9 connector. Because the receiver uses high input impedance and the transmitter driver has stronger driving capability than RS-232, it allows multiple receiving nodes to be connected on the same transmission line, up to 10 nodes. This means one master device and the rest as slave devices, with no communication between slave devices, thus RS-422 supports point-to-multipoint bidirectional communication. The input impedance of the receiver is 4k, so the maximum load capability of the transmitter is 10×4k+100Ω (termination resistance). The RS-422 four-wire interface does not need to control data direction, as any necessary signal exchange between devices can be achieved through software (XON/XOFF handshake) or hardware (a pair of separate twisted wires). The maximum transmission distance of RS-422 is 4000 feet (approximately 1219 meters), with a maximum transmission rate of 10Mb/s. The length of the balanced twisted pair is inversely proportional to the transmission rate, allowing the maximum transmission distance only at rates below 100kb/s. The highest rate can only be achieved over very short distances. Typically, the maximum transmission rate over 100 meters of twisted pair is only 1Mb/s.

RS-422 requires a termination resistance, which should be approximately equal to the characteristic impedance of the transmission cable. Termination resistors are generally unnecessary for transmission distances less than 300 meters. The termination resistor is connected at the far end of the transmission cable.

3. RS-485 Electrical Specifications

Since RS-485 is developed based on RS-422, many of its electrical specifications are similar to those of RS-422. Both use balanced transmission methods and require termination resistors on the transmission line. RS-485 can use both two-wire and four-wire methods; the two-wire system can achieve true multipoint bidirectional communication.

In contrast, when using a four-wire connection, like RS-422, it can only achieve point-to-multipoint communication, meaning there can only be one master device and the rest as slave devices. However, RS-485 has improved upon this, allowing up to 32 devices to be connected on the bus regardless of whether it is a four-wire or two-wire connection.

Another difference between RS-485 and RS-422 is their common-mode output voltage; RS-485 ranges from -7V to +12V, while RS-422 ranges from -7V to +7V. The minimum input impedance for RS-485 receivers is 12k, while for RS-422 it is 4k. RS-485 meets all RS-422 specifications, allowing RS-485 drivers to be used in RS-422 networks.

Like RS-422, RS-485 has a maximum transmission distance of approximately 1219 meters and a maximum transmission rate of 10Mb/s. The length of the balanced twisted pair is inversely proportional to the transmission rate, allowing the maximum cable length only at rates below 100kb/s. The highest transmission rate can only be achieved over very short distances; typically, the maximum transmission rate over 100 meters of twisted pair is only 1Mb/s.

RS-485 requires two termination resistors, with resistance values equal to the characteristic impedance of the transmission cable. Termination resistors are generally unnecessary for transmission distances below 300 meters. The termination resistors are connected at both ends of the transmission bus.

RS-422 supports 10 nodes, while RS-485 supports 32 nodes, allowing for multi-node network configuration. The network topology generally adopts a terminal-matched bus structure and does not support ring or star networks. When constructing the network, the following points should be noted:

1. Use a twisted pair cable as the bus, connecting each node in series. The lead length from the bus to each node should be kept as short as possible to minimize the impact of reflected signals in the leads on the bus signal. The illustration shows some common incorrect connection methods (a, c, e) and correct connection methods (b, d, f) in practical applications. Although the three incorrect network connections (a, c, e) may work normally at short distances and low speeds, their adverse effects will become more severe as communication distance increases or speed rises. This is primarily because signals reflecting at the ends of each branch will superimpose with the original signal, degrading signal quality.

2. Ensure continuity of the bus characteristic impedance; signal reflections will occur at points of impedance discontinuity. The following situations are likely to create such discontinuities: different cables are used for different segments of the bus, or too many transceivers are installed too close together on one segment, or excessively long branch lines are drawn to the bus.

In summary, a single, continuous signal path should be provided as the bus.

Termination resistors are generally required for matching in RS-422 and RS-485 bus networks. However, at short distances and low speeds, termination matching may not need to be considered. Under what circumstances can matching be disregarded? Theoretically, when sampling the data signal at the midpoint of each receiver, as long as the reflected signal decays to a sufficiently low level before sampling begins, matching can be ignored. However, this is difficult to manage in practice. An article by Maxim in the US mentions an empirical rule that can be used to determine when matching is necessary based on the data rate and cable length: when the signal transition time (rise or fall time) exceeds three times the one-way propagation time of the signal along the bus, matching can be omitted. For instance, the RS-485 interface MAX483 has a minimum rise or fall time of 250ns, with the typical signal transmission rate over twisted pair cable being about 0.2m/ns (for 24AWG PVC cable). Therefore, if the data rate is below 250kb/s and the cable length does not exceed 16 meters, termination matching can be omitted when using MAX483 as the RS-485 interface.

Typically, termination matching employs termination resistors, as previously mentioned. In RS-422, the termination resistors are connected at the far end of the bus cable, while in RS-485, termination resistors should be connected at both the start and end of the bus cable. Termination resistors are typically 100Ω in RS-422 networks and 120Ω in RS-485 networks, corresponding to the cable’s characteristic impedance, as most twisted pair cables have characteristic impedances around 100–120Ω. This matching method is simple and effective but has the drawback of consuming considerable power, making it unsuitable for systems with strict power consumption limits.

Another, more power-efficient matching method is RC matching, which uses a capacitor C to block the DC component, saving most of the power. However, selecting the value of capacitor C is challenging, requiring a trade-off between power consumption and matching quality.

Another method involves using diodes for matching. Although this method does not achieve true ‘matching’, it uses the clamping effect of diodes to quickly attenuate reflected signals, improving signal quality. The energy-saving effect is significant.

Proper grounding of electronic systems is crucial but often overlooked. Inadequate grounding can lead to unstable operation of electronic systems or even jeopardize system safety. Grounding in RS-422 and RS-485 transmission networks is equally important, as improper grounding can affect the stability of the entire network, especially in harsh working environments and long transmission distances, where grounding requirements are more stringent. Otherwise, the interface failure rate may be high. In many cases, when connecting RS-422 and RS-485 communication links, a simple twisted pair is used to connect the ‘A’ and ‘B’ terminals of each interface, neglecting the connection of the signal ground. This method may work normally in many situations but poses significant risks for the following two reasons:

1. Common-mode interference: As previously mentioned, both RS-422 and RS-485 interfaces use differential methods for signal transmission, which do not require detection of signals relative to a reference point; the system only needs to detect the potential difference between the two wires. However, it is often overlooked that the transceiver has a certain common-mode voltage range, with the common-mode voltage range for RS-422 being -7 to +7V and for RS-485 being -7 to +12V. Only when these conditions are met can the entire network function correctly. When the common-mode voltage in the network exceeds this range, it can affect communication stability and reliability, or even damage the interface. For example, when the transmitting driver A sends data to the receiver B, the common-mode voltage output from transmitter A is VOS. Since both systems have their own independent grounding systems, there is a ground potential difference VGPD. Thus, the common-mode voltage at the receiver input can reach VCM = VOS + VGPD. Both RS-422 and RS-485 standards stipulate that VOS ≤ 3V; however, VGPD can vary widely (tens of volts or even more) and may be accompanied by strong interference signals, causing the receiver’s common-mode input VCM to exceed the normal range and generate interference current on the transmission line, which can lead to communication errors or damage to the communication interface circuit.

2. EMI issues: The common-mode portion of the output signal from the transmitting driver requires a return path; if there is no low-resistance return path (signal ground), it will return to the source end in the form of radiation, causing the entire bus to act like a large antenna radiating electromagnetic waves.

For the above reasons, even though RS-422 and RS-485 use differential balanced transmission methods, the entire RS-422 or RS-485 network must have a low-resistance signal ground. A low-resistance signal ground connects the working grounds of both interfaces, short-circuiting the common-mode interference voltage VGPD.

This signal ground can be an additional line (non-shielded twisted pair) or the shielding layer of a shielded twisted pair. This is the most common grounding method.

It should be noted that this method is only effective for high-resistance common-mode interference; since the interference source has high internal resistance, short-circuiting will not create significant ground loop current, which does not greatly affect communication. When the internal resistance of the common-mode interference source is low, significant loop current can form on the grounding line, affecting normal communication. The author believes that the following three measures can be taken:

(1) If the internal resistance of the interference source is not very low, a current-limiting resistor can be added to the grounding line to restrict the interference current. Increasing the grounding resistance may raise the common-mode voltage, but as long as it is controlled within an appropriate range, it will not affect normal communication.

(2) Use floating ground technology to break the grounding loop. This is a common and effective method. When the common-mode interference internal resistance is very low, the above methods may not be effective, and at this point, consideration can be given to floating the node introducing interference (for example, field devices located in harsh working environments), thereby isolating the grounding loop and preventing significant loop current from forming.

(3) Use isolation interfaces. In some cases, for safety or other considerations, the circuit ground must be connected to the chassis or ground and cannot be floated. In this case, isolation interfaces can be used to break the grounding loop, but there should still be a ground line connecting the common terminal on the isolated side with the working ground of other interfaces.

Both RS-422 and RS-485 standards specify a receiver threshold of ±200mV. This specification provides relatively high noise suppression capabilities. As mentioned earlier, when the level of receiver A is more than +200mV higher than that of receiver B, the output is positive logic; conversely, it is negative logic. However, the existence of a third state presents a problem: after the host finishes sending a piece of information, the bus is set to the third state, meaning that when the bus is idle, no signal drives the bus, causing the voltage between A and B to fall between -200mV and +200mV, eventually approaching 0V. This brings about an issue of uncertainty in the receiver output state. If the receiver output is 0V, the slave devices in the network will interpret this as a new start bit and attempt to read subsequent bytes, leading to a frame error result, with no devices requesting the bus, causing the network to become paralyzed. In addition to the bus being idle leading to a voltage difference between the two lines dropping below 200mV, open circuits or short circuits can also cause this situation. Therefore, measures should be taken to prevent the receiver from being in an uncertain state.

Typically, a bias is added to the bus; when the bus is idle or open, bias resistors are used to bias the bus to a definite state (differential voltage ≥ -200mV). As shown in Figure 1, A is pulled to ground, and B is pulled down to 5V, with the typical value of the resistors being 1kΩ, with specific values varying according to the capacitance of the cable.

The above method is a classic approach, but it does not solve the problem of short circuits on the bus. Some manufacturers have shifted the receiver threshold to -200mV/-50mV to resolve this issue.

The grounding measures mentioned earlier only protect against low-frequency common-mode interference and are ineffective against high-frequency transient disturbances. Since transmission lines act like inductors for high-frequency signals, the grounding line is effectively an open circuit for high-frequency transient disturbances. Although such transient disturbances last only a short time, they can produce hundreds to thousands of volts.

In practical applications, there is still a possibility of high-frequency transient disturbances. Generally, switching large inductive loads such as motors, transformers, relays, or during lightning events can generate high-amplitude transient disturbances, which can damage RS-422 or RS-485 communication interfaces if not adequately protected. To protect against such transient disturbances, isolation or bypass methods can be employed.

1. Isolation protection method. This scheme effectively transfers the transient high voltage to the isolation layer of the isolation interface. Due to the high insulation resistance of the isolation layer, no damaging surge current will occur, thus protecting the interface. High-frequency transformers, optocouplers, and other components are typically used to achieve electrical isolation of the interface, and some device manufacturers have integrated all these components into a single IC for ease of use. The advantage of this scheme is that it can withstand high voltages and long-duration transient disturbances, and it is relatively easy to implement, but it is more expensive.

2. Bypass protection method. This scheme uses transient suppression components (such as TVS, MOV, gas discharge tubes, etc.) to divert harmful transient energy to ground. The advantages are low cost, but the protection capability is limited, only able to protect against transient disturbances of certain energy levels and durations, and it requires a good connection to ground, which can be challenging to achieve. In practical applications, these two methods are often combined flexibly, as shown in Figure 1. In this method, the isolation interface isolates high-amplitude transient disturbances, while bypass components protect the isolation interface from excessive transient voltage breakdown.

When using RS485 interfaces, the maximum allowable cable length for a specific wire gauge from the generator to the load is a function of the data signal transmission rate, primarily limited by signal distortion and noise. The relationship between maximum cable length and signal rate is based on the use of 24AWG copper twisted telephone cable (wire diameter of 0.511mm), with a line-to-line bypass capacitance of 52.5PF/M and a terminal load resistance of 100 ohms. When the data signal rate drops below 90Kbit/S, assuming a maximum allowable signal loss of 6dBV, the cable length is limited to 1200M. In practice, it is entirely possible to achieve greater cable lengths than this.

On an RS-485 bus, only one transmitter can send at any time. In half-duplex mode, only one master can send. In full-duplex mode, the master station can always send, while the slave stations can only send one at a time.

In long-line signal transmission, termination matching resistors are generally required at the receiving end to avoid signal reflection and echo. The value of the termination matching resistor depends on the impedance characteristics of the cable and is independent of the cable length.

RS-485/RS-422 generally uses twisted pair cables (shielded or unshielded), with termination resistors typically ranging from 100 to 140Ω, with a typical value of 120Ω. In practical configuration, a termination resistor is connected at each of the two terminal nodes of the cable, that is, at the nearest and farthest ends, while nodes in the middle should not connect termination resistors, as this would cause communication errors.

This situation arises when users set up the RS-485 network without following the principle that the wiring from the nodes to the bus should be as short as possible. If the bus wiring adheres to this principle, the issue of not knowing which node is the farthest would not arise. Additionally, it should be noted that such wiring will lead to poor system performance.

Due to RS-485/RS-422 requiring that all transmit enable control signals be turned off and the receive enable signal remain active after data transmission is complete, the bus driver enters a high-impedance state, allowing the receiver to monitor the bus for new communication data. However, because the bus is in a passive driving state (if there are termination resistors on the bus, the differential level between A and B will be 0, and the receiver output will be uncertain; if there are no termination resistors, the bus will be in a high-impedance state, and the receiver output will also be uncertain), it can easily be affected by external noise. When the noise voltage exceeds the input signal threshold (typically ±200mV), the receiver will output data, leading to invalid data being received by the corresponding UART, resulting in errors in subsequent normal communications; another situation may occur during the moment of enabling/disabling the transmission control, causing the receiver to output signals, which will also lead to UART receiving erroneous data.

Solutions:

1) Use a pull-up on the non-inverting input of the communication bus (A line) and a pull-down on the inverting input (B line) to clamp the bus, ensuring the receiver output is fixed to a ‘1’ level;

2) Replace the interface circuit with MAX308x series products that have built-in fault tolerance;

3) Eliminate through software by adding 2-5 start synchronization bytes within the communication data packet, ensuring that real data communication only begins after meeting the synchronization header.

1. Signal Reflection in Communication Cables

During communication, two types of signal reflection can occur due to impedance discontinuities and mismatches. Impedance discontinuity occurs when the signal suddenly encounters a very small or no impedance at the end of the transmission line, causing reflection at that point, similar to how light reflects when entering a different medium.

To eliminate such reflections, a termination resistor equal to the characteristic impedance of the cable should be bridged at the end of the cable to ensure impedance continuity. Since signals on the cable are transmitted bidirectionally, a termination resistor of the same size should be bridged at the other end of the communication cable. Theoretically, as long as a termination resistor matching the cable’s characteristic impedance is connected at the end of the transmission cable, signal reflection should not occur. However, in practical applications, due to the characteristic impedance of the transmission cable and the communication baud rate, complete equality between the characteristic impedance and termination resistance is unlikely, so some level of signal reflection will still exist.

The other cause of signal reflection is the impedance mismatch between the transceiver and the transmission cable. This type of reflection mainly manifests when the communication line is in idle mode, leading to data chaos across the network.

The impact of signal reflection on data transmission ultimately arises from reflected signals triggering the comparator at the receiver’s input, causing the receiver to receive erroneous signals, leading to CRC check errors or entire data frame errors.

In signal analysis, the parameter used to measure the strength of reflected signals is RAF (Reflection Attenuation Factor). Its calculation formula is as follows:

RAF=20lg（Vref/Vinc）（1）

Where: Vref – the voltage level of the reflected signal; Vinc – the voltage level of the incident signal at the connection point between the cable and the transceiver or termination resistor.

The specific measurement method is illustrated in Figure 3. For example, if the peak-to-peak value of an incident signal sine wave at 2.5MHz is +5V, and the peak-to-peak value of the reflected signal is +0.297V, then the reflection attenuation factor for this communication cable at a frequency of 2.5MHz is: RAF=20lg（0.297/2.5）=-24.52dB.

To mitigate the impact of reflected signals on communication lines, noise suppression and bias resistor methods are typically employed. In practical applications, for relatively small reflected signals, the bias resistor method is often used for simplicity and convenience. The principle of how adding bias resistors can improve communication reliability will be detailed later.

The second factor affecting signal transmission is signal attenuation during transmission through the cable. A transmission cable can be viewed as an equivalent circuit composed of distributed capacitance, distributed inductance, and resistance.

The distributed capacitance C of the cable is mainly generated by the two parallel conductors of the twisted pair. The resistance of the conductors has minimal influence on the signal and can be ignored. Signal loss is primarily due to the LC low-pass filter formed by the distributed capacitance and inductance of the cable. The attenuation coefficients of standard LAN-type two-core inductance used in PROFIBUS (the standard cable selected for Siemens DP bus) vary with different baud rates.

The third factor affecting communication performance is the magnitude of the pure resistive load (also known as DC load). This refers primarily to the termination resistance, bias resistance, and RS-485 transceiver.

In discussing the RS-485 specification, it was mentioned that the RS-485 driver can output a differential voltage of at least 1.5V when connected with 32 nodes and configured with 150Ω termination resistance. The input resistance of a receiver is 12kΩ. The equivalent circuit of the entire network is shown in Figure 5. Based on this calculation, the load capability of the RS-485 driver is: RL=32 parallel input resistances || 2 termination resistors=（（12000/32）×（150/2））/（12000/32）+（150/2））≈51.7Ω.

Currently, commonly used RS-485 drivers include MAX485, DS3695, MAX1488/1489, and the SN75176A/D used by Holley, among others. Some RS-485 drivers can achieve load capacities of up to 20Ω. Without considering many other factors, the maximum number of nodes that a driver can support will far exceed 32.