The PROFIBUS DP Bus Cycle and Token Passing
PROFIBUS DP is deterministic because access to the bus is strictly controlled: masters pass a token to decide who may talk, and each master polls its slaves in a fixed order. Understanding this cycle explains why DP timing is repeatable and where a slow slave hurts. This page walks through token passing and the master-slave polling that make up a DP bus cycle.
PROFIBUS DP Bus Cycle in one line: The PROFIBUS DP bus cycle combines token passing and polling. A logical token circulates among the masters, and only the master holding the token may initiate traffic. While it holds the token, a master polls each of its slaves in turn with a request-response exchange, reading inputs and writing outputs. When done it passes the token on, giving a deterministic, repeatable scan of the whole bus.
Token Passing Decides Who May Talk
PROFIBUS uses a hybrid access method. Among the active stations - the masters - a logical token circulates in a defined order, and only the station currently holding the token is allowed to initiate communication on the bus. This prevents two masters from talking at once and makes access orderly rather than contention-based. When a master finishes its work it passes the token to the next master in the logical ring, and eventually the token comes back around, so each master gets a predictable turn.
This matters even on a single-master bus, which is the common case for a DP line with one PLC: there the token simply returns to the same master each rotation, and the master spends its turn polling slaves. Where a second master exists - often a diagnostic or programming device - token passing is what lets it share the bus without colliding with the control master. The token, and the timing budget it enforces, is a big part of why DP timing is predictable enough to call deterministic.
Polling the Slaves in a Fixed Order
The passive stations - the DP slaves such as remote I/O and drives - never initiate; they only answer. While a master holds the token it runs through its list of slaves, sending each a request and receiving that slave's response in one exchange. That exchange carries the outputs the master is writing and returns the inputs the slave is reporting, so one poll refreshes both directions of data for that slave. The master polls every configured slave once per cycle, in order, then repeats.
Because the cycle is a fixed sweep, its duration is the sum of the individual slave exchanges plus overhead, which is what makes the DP scan time predictable and repeatable. Add more slaves or slow ones and the cycle lengthens uniformly; that is why a single unresponsive slave, timing out and being retried, visibly stretches the cycle for everyone. This request-response, master-writes-and-reads pattern is the same basic model as a HART master polling a device, just applied cyclically across a whole bus. The data DP gathers is then handed up to the PLC, and from there to any historian or SCADA that trends it.
The Timing Parameters Behind the Cycle
Several configured times shape how the cycle behaves, and they explain most odd timing symptoms. The target rotation time (TTR) is the budget for one full token rotation: how long, at most, a rotation should take with all cyclic traffic plus any acyclic extras. Each master compares the actual rotation time against TTR to decide how much low-priority traffic it may add on its turn. The slot time bounds how long a master waits for a reply before treating the exchange as failed and retrying, and the retry limit caps how many attempts a silent slave gets before being marked absent.
On the slave side, a watchdog timer supervises the master: if a slave stops being polled for longer than its watchdog time, it assumes control is lost and drives its outputs to their configured fail-safe state. This is the parameter that decides what a remote I/O rack's valves do when the cable is cut, so its value and the fail-safe output states deserve deliberate review with the people who own the process, not defaults accepted at download time.
Baud Rate, Segment Length, and Station Limits
DP runs over RS-485, and the standard defines baud rates from 9.6 kbit/s up to 12 Mbit/s, with slaves auto-detecting the rate in use. The physics trade is simple: the faster the bus, the shorter the allowed segment, with the length-per-baud-rate table fixed by the PROFIBUS cabling standard. The RS-485 rules also cap a segment at 32 stations before a repeater is required, while the protocol's address space allows up to 126 stations on the whole network; repeaters extend both reach and station count at the cost of added delay per hop.
Bus speed and cycle time connect directly: every poll exchange takes time proportional to its data length at the configured baud rate, so a long list of slaves at low baud produces a slow cycle. Before raising the baud rate to fix a slow cycle, verify that the installed cable, connectors, and total segment lengths actually support the higher rate, because a marginal physical layer at high speed produces exactly the intermittent retries that stretch the cycle worse than the low baud rate did.
Commissioning Checks That Prevent Cycle Problems
A short discipline at commissioning prevents most DP field calls:
- Terminate both physical ends of each segment, and only the ends - the DP terminator is powered, so it must sit in a connector plugged into a station that stays energized.
- Verify every station address is unique and matches the configuration, and that the master's highest-station-address setting is not below any real address.
- Confirm the GSD-based configuration for each slave matches the physical modules installed, since a mismatch leaves the slave refusing data exchange.
- Check the master's diagnostic or live list to see every expected slave present with no retries accumulating.
- Record the healthy cycle time as a baseline for later comparison.
That baseline matters because a stretching cycle is DP's early-warning signal: corrosion in a connector or water in a cable shows up as retries to one station long before anything fails outright. The bus statistics live in the master, and once the PLC's data reaches a historian the symptoms can be trended alongside process data - the plumbing for that is the same as for any controller, as described in connecting a Siemens S7 PLC to a cloud SCADA.
Frequently Asked Questions
How does token passing work in PROFIBUS DP?
A logical token circulates among the masters in a defined order, and only the master holding the token may initiate traffic. When it finishes, it passes the token to the next master. On a single-master bus the token simply returns to the same master each rotation. Token passing keeps bus access orderly and prevents two masters from talking at once.
Do PROFIBUS DP slaves ever initiate communication?
No. DP slaves are passive stations that only respond to a master's poll. While a master holds the token it sends each slave a request and receives a response, which carries outputs to the slave and returns the slave's inputs in one exchange. Slaves never speak unprompted, which is what makes the polling cycle deterministic.
Why does one slow slave stretch the whole DP cycle?
The bus cycle is a fixed sweep in which the master polls every slave once per rotation. The cycle time is the sum of all the individual exchanges plus overhead. A slave that responds slowly, or times out and gets retried, lengthens its part of the sweep, which pushes out the whole cycle and delays every other slave's refresh.
What does the PROFIBUS watchdog on a slave actually do?
It supervises the master from the slave's point of view. If polls stop arriving for longer than the configured watchdog time - a cut cable, a dead master, a broken segment - the slave assumes control is lost and drives its outputs to their configured fail-safe states. Reviewing those fail-safe states with the process owners is as important as the watchdog value itself, because they define what the field equipment does during a bus failure.
Why must a PROFIBUS DP terminator stay powered?
The DP termination network is active: it uses the interface's supply to bias the bus to a defined idle state as well as to match the line. If the station carrying a termination plug is powered off, the segment loses both its bias and its matching at that end, and the whole segment can start throwing retries or fail outright even though every other station is healthy. That is why terminators belong on stations that are always energized.
Sources and verification
This page references the vendor products and their official documentation published by the organizations below. Editions, product capabilities, and documentation change over time - confirm current requirements and specifications directly with the source.
Merobix is not affiliated with, endorsed by, or sponsored by these organizations; their names are used only to identify the standards and products discussed.
Automation services
Need help turning this into a working system?
Merobix integrates SCADA, programs Allen-Bradley and Siemens PLCs, and designs and fabricates industrial control panels.
Meeting requests are reviewed before confirmation.