Uncovering the physical layer of PROFIBUS-DP

Author : Christopher Gobok, Linear Technology

27 June 2016

Figure 1 -  Cable, termination & pin differences in RS485 & PROFIBUS-DP networks

The PROFIBUS-DP (Distributed Peripherals) fieldbus standard has been around for over two decades and, yet, physical layer requirements can still be unclear, which often leads to confusion in transceiver definitions. 

(Click here to view article in digi-issue)

However, any ambiguity clearly didn’t stop PROFIBUS from becoming a success - perhaps the most successful fieldbus solution, with over 50 million devices installed worldwide.

As new systems are deployed, it’s important to know you are using transceivers designed to the most up to date and accurate interpretation of the PROFIBUS-DP standard.  For example, Linear Technology’s LTC2877 PROFIBUS-DP transceiver stays true to the latest IEC 61158-2 PROFIBUS-DP standard and adds protection mechanisms, increasing the compatibility and reliability of new master and slave devices.

The faster, simpler PROFIBUS-DP standard was born in 1993 from the slower, more complex PROFIBUS FMS (Fieldbus Message Specification) parent standard.  PROFIBUS-DP also has a younger sibling or derivative standard, PROFIBUS-PA (Process Automation), which uses Manchester Bus Powered (MBP) transmission, adding power over the bus and making it well suited for intrinsic safety applications in hazardous environments.  Otherwise, PROFIBUS-DP is the most widely used version of PROFIBUS today, probably because its plug-and-play nature, flexibility, and cost effectiveness are attractive in the majority of fieldbus applications.  From the management of sensors and actuators in an industrial plant, to communication with flowmeters out in the rail yard, PROFIBUS-DP decentralises I/O cards (masters) from controllers and brings them closer to sensors and actuators (slaves), resulting in numerous installation and operational benefits.  

PROFIBUS-DP can communicate over a variety of media, including copper wires, fiber optics, and even air in an infrared communicator.  By far, the most commonly used media for bit transmission (layer 1 of the ISO/OSI model) by PROFIBUS-DP masters and slaves is a twisted pair of wires, connecting devices that communicate using TIA/EIA-485-A (RS485) transceivers.  This isn’t surprising considering RS485’s high speed differential signaling and robust communication between multiple devices over long distances in noisy environments such as factory applications.  Multiple masters, like PLCs (programmable logic controllers) using RS485 transceivers, can connect to 30 slaves per segment in a linear topology, where networks can be expanded to 124 slaves with the use of hubs (parallel segments) or repeaters (serial segments).  Each segment must be terminated at both ends using active termination.  All slaves can be hot swapped into the bus and their position does not matter since each slave is assigned a unique network address.

95% RS485, 5% confusion

PROFIBUS-DP adopted much of the TIA/EIA-485-A  (RS485) standard, but did make a few changes that can be accidentally overlooked due to larger system issues.  As a result, contrary to popular belief, not all RS485 transceivers and cables are suitable for PROFIBUS-DP networks and vice versa.  Differences in cabling, termination, signal names, and driver requirements do exist; being too quick to dismiss these differences could easily cost you the performance or worse, the certification, of your master or slave device.

Figure 2 - Testing the LTC2877 differential output voltage (VOD) using a PROFIBUS-DP load

While the RS485 standard does not specify any specific cabling requirements, 120? shielded twisted pair has become the normal recommendation.  PROFIBUS-DP, however, recommends 150? shielded twisted pair.  Unfortunately, 120? cannot be approximated as 150? and this small difference in cable impedance actually necessitates the use of different and, in most cases, new cables.  PROFIBUS-DP also specifies a maximum cable length that depends on which one of ten baud rate “steps” is used, ranging from 1,200m at 9.6kbps to 100m at 12Mbps.  

Of course, with different cable impedance requirements, comes different termination requirements.  To minimise signal reflections, RS485 installations typically use a single 120? termination resistor at both ends of the bus, whereas PROFIBUS-DP recommends a 171? termination network at both ends of the bus.  Wait, was that a typo?… PROFIBUS-DP recommends a 171?  termination network, thereby not matching the 150? characteristic impedance of the cable it also recommends?  Absolutely!  Figure 1 shows how the cable and termination network used for PROFIBUS-DP is different from RS485.  You can see that two 390? bus biasing resistors are used in conjunction with the 220? termination resistor for PROFIBUS-DP; the effective differential resistance of this termination network is 171?, which is obviously not a perfect match for the 150? cable, resulting in a slightly underdamped network.  Don’t worry though, as this reveals itself as only a small bump or increase in signal voltage at the receiving end of the cable, lasting twice as long as the cable propagation delay.

If the cable/termination mismatch were not enough, then the naming of the bus pins on PROFIBUS transceivers should further break your expectations.  You may have noticed the opposite pin names used in Figure 1.  In most general purpose RS485 transceivers, pin A is the non-inverting receiver input (and non-inverting driver output) and pin B is the inverting receiver input (and inverting driver output) relative to the receiver output and driver input.  However, the PROFIBUS standard describes the bus polarity in such a way that pins B and A are exactly the opposite of this.  Why the inconsistency?  The original TIA/EIA-485-A standard is not explicit in its definition of the bus polarity relative to the logic signal function, so RS485 IC designers have almost always interpreted the specification one way while others interpreted it the other way.  What that means for you, especially if you have both RS485 and PROFIBUS-DP projects, is to pay close attention when mapping transceiver bus pins to connectors.

Based on the number of ill-defined transceivers available off-the-shelf, differential driver output voltage (VOD) is perhaps the most misinterpreted (or most intentionally ignored) specification in PROFIBUS-DP’s physical layer.  RS485 specifies that the VOD between the A and B lines shall be 1.5V to 5V, peak differential, measured at the driver terminals with a 54? resistor between A and B.  However, PROFIBUS-DP specifies that the VOD shall be 4V to 7V, peak-to-peak differential, measured at the far end of the cable, with termination at each end.  Clearly, these requirements are quite different.

A common misunderstanding is that if an RS485 driver simply develops more than 2.1V across a 54? load, then it will meet PROFIBUS-DP’s requirements when used with a PROFIBUS-DP termination network.  However, this is not always true.  The strength of an RS485 driver can be too high and exceed the 7VPP PROFIBUS-DP limit.  In other words, be wary of the all too common “PROFIBUS” compatible RS485 transceiver that only specifies a VOD minimum value (i.e. 2.1V) without a maximum value.  The best way to ensure PROFIBUS-DP VOD compliance is to test the transceiver with a PROFIBUS load.  Figure 2 shows how the LTC2877 Rugged PROFIBUS RS485 Transceiver is tested with a PROFIBUS-DP load and some series resistance to simulate cable losses, where the VOD (blue curve) is generated from measurements taken at the “end of the cable” (A’ and B’) to ensure the PROFIBUS-DP specification is truly met; the LTC2877 is also fully tested with RS485 loads to ensure VOD compatibility with both standards.

Protecting PROFIBUS-DP

Figure 3 - LTC2877 providing multiple levels of protection

The TIA/EIA-485-A standard specifies very little when it comes to combating noise, faults, electrostatic discharge (ESD), electrical fast transients (EFTs), or surge, leaving it open for transceiver manufacturers and system designers to implement their own means of protection.  Despite the flexibility, people are all too familiar with the hazards posed by the harsh environments cables must often snake through and have come to expect a certain minimum level of built-in protection in PROFIBUS-DP transceivers.  While protection requirements differ from application to application, the LTC2877, shown in figure 3, single-handedly blankets market requirements with high levels of electrical protection. 

The TIE/EIA-485-A standard specifies that ground shifts between two devices on a network can be as large as -7V to +12V during operation.  However, many PROFIBUS-DP installations can easily encounter voltages much greater than this, which can cause critical damage to the ordinary PROFIBUS-DP transceiver if these bus voltages are exceeded by even a couple of volts.  PROFIBUS is often used in 24V systems, where shorting a “standard” RS485 device to 24V can be fatal.  The receiver in the LTC2877 features an extended common mode range of -25V to +25V, allowing the LTC2877 to survive large common mode voltages and continue to transmit and receive data without disruption.  A transceiver with limited overvoltage tolerance makes implementation of effective external protection networks difficult without interfering with proper data network performance.  Replacing the usual PROFIBUS-DP transceiver with the ±60V protected LTC2877 can easily eliminate field failures due to overvoltage faults without using costly external protection devices. 

It goes without saying that PROFIBUS-DP transceivers are literally a system’s first line of defense and need to be able to protect themselves against various levels of electrical overstress, especially the most encountered form, ESD strikes.  While some PROFIBUS transceivers are protected to 15kV ESD protection on their bus pins while unpowered, the bus pins on the LTC2877 provide ±26kV HBM protection with respect to ground or either of the supplies without latchup or damage, unpowered or powered, and in any mode of operation.  Furthermore, the bus pins are protected against a whopping ±52kV to ground when unpowered.  This impressive level of ESD protection is just a testament to how rugged the LTC2877 is.

Another form of electrical overstress is EFTs which, according to the IEC 61000-4-4 EFT standard, are bursts of high voltage spikes, lasting 60 seconds.  This type of overstress is usually a result of arcing contacts in switches and relays, common in industrial environments where electromechanical switches are used to connect and disconnect inductive loads.  The LTC2877 meets the highest level of IEC 61000-4-4 severity, which is level 4, equating to an open circuit voltage of 2kV on the bus pins.

Perhaps the most severe form of electrical overstress is the surge Mother Nature delivers in the form of lightning, where a single bolt can carry up to 5 billion Joules of energy. It’s no surprise, then, that tiny transceiver ICs, like the LTC2877, do not possess inherent protection against electrical surges of this magnitude. Instead, external surge protection components, including MOVs (metal oxide varistors), TVS (transient voltage suppression) diodes, TSPDs (thyristor surge protection devices) and GDTs (gas discharge tubes), are typically used in PROFIBUS-DP systems that are exposed to the elements. And, while the LTC2877 won’t fare well against lightning strikes solo, its high ±60V pin rating makes it easy to find external protection components capable of providing this level of protection.  


It’s probably safe to say that PROFIBUS-DP is here to stay a while as the king of fieldbus communications because of its many outstanding qualities. Designing with Linear Technology’s LTC2877 accounts for all of the subtle, but important, differences between RS485 and PROFIBUS-DP, including the strength of the driver. Furthermore, the LTC2877 implements multiple protection mechanisms, including robust ESD cells, to guard against all sorts of threats. Proofing a PROFIBUS-DP prototype has never been easier.

Contact Details and Archive...

Print this page | E-mail this page