PROFIBUS DP network sizing
A PROFIBUS DP network is a chain of RS-485 segments separated by active elements — repeaters, hubs and optical link modules. Each segment has its own length, spur and station budget, set by the data rate; the network as a whole has an address space, a slot time that grows with every active element and every kilometre of fibre, and a bus cycle. This page shows those checks and the tables behind them.
When to use
When choosing the data rate of a DP network, when deciding where a repeater or ProfiHub is needed, when a station has to hang from a T-piece instead of the trunk, when sizing an optical link between buildings and when setting the TTR and the watchdog in the master configuration.
What gets sized in a DP network
A PROFIBUS DP network is described by its segments. The first one leaves the master and may end at an active element — a repeater, a ProfiHub or an optical link module — from which new segments leave, each with its own terminators. Every segment is checked against the data rate: total length, capacitance of the spur lines and number of stations. The network as a whole is then checked for addresses, slot time, optical links and the bus cycle. Unlike DeviceNet, the RS-485 pair carries no power, so there is no voltage-drop budget; unlike ControlNet, length is set by the data rate rather than by the number of taps.
Length and spurs depend on the data rate
The higher the rate, the shorter the segment: 1200 m at low rates, 100 m from 3 Mbit/s up. Spur lines count in that length and also have their own limit in capacitance, which shrinks with the rate until it reaches zero above 1.5 Mbit/s. The tool also sweeps the rates and reports the highest one at which the whole layout passes — only rates that were actually checked, never one without tables.
Active elements, cascade and slot time
Each active element splits the bus and lets the network grow, but costs time: the request and the reply cross every unit in the path and every kilometre of fibre. The tool adds those delays, compares them with the default slot time of the profile and with the cascade count published by the manufacturer, and, for optical links, also applies the Siemens formula with the 1300 nm module floors. Each optical link has its budget checked — launched power, sensitivity, fibre attenuation, connectors and splices — together with the minimum and maximum reach of the fibre. Hubs, OLMs and the fibre patch panels where the links end usually share a cabinet; their space and power strips are sized in 19” rack sizing.
Addresses, cycle and watchdog
Station places are a local limit (32 per segment); addresses are global (0 to 125, with 126 and 127 reserved), and duplicates fail the network. With the cyclic bytes of each slave — remote I/O, drives, the communication modules of the MCC drawers — the tool calculates the bus cycle, the minimum TTR with a 20 % margin and the two watchdog rules in use. In a redundant optical ring it also checks for a free address below the HSA and at least 3 retries.
What this tool does not cover
It does not configure the master: the bus parameters it calculates are checks to compare with what the engineering tool generates. Two-wire types other than type A are only represented through the declared cable capacitance, and active elements outside the catalogue need their channels and delay entered from the datasheet. The PA segments behind a DP/PA coupler are sized in PROFIBUS PA network sizing, and instrument networks that skip the coupler altogether in Ethernet-APL network sizing.
Formulas and fundamentals
L_total = L_trunk + Σ L_spur ≤ L_seg(rate) The normative criterion; exceeding it fails the segment. The SMAR trunk ceiling (380 m at 500 kbit/s, 193.4 m at 1.5 Mbit/s…) reserves budget for spurs and only raises a warning.
C_spur = (Σ L_spur × C'_cable + n_conn × C_conn) / 1000 ≤ C_max(rate) [nF] Capacitance is the primary axis; the limit in metres is derived from it and changes with the declared cable (30 pF/m by default for type A). A spur only exists where a station hangs from a T-piece — stations chained through the D-sub connector are trunk.
Tslot = maxTSDR + 2 × (Σ delay_unit + Σ L_fibre[km] × delay_fibre) Each active element in the path and each kilometre of fibre delay the reply (fibre: 5 ns/m). With optical link modules the Siemens check Tslot = a + b·L_FO + c·N_OLM is also applied and the larger value governs. Exceeding the default slot time of the profile is a warning — the engineering tool usually raises it; it only fails when a configured Tslot is declared and does not cover the delay.
Margin = (P_tx − sensitivity) − (α × L + n_conn × a_conn + n_splice × a_splice) Below 3 dB the link is flagged (ageing reserve); below 0 dB it fails. A link shorter than the minimum of the fibre overdrives the receiver.
T_msg = 300 + 11 × bytes ; T_cycle = T_token + T_GAP + Σ T_msg [tBit] Each byte travels as 11 bits; bytes are input plus output of the slave. T_token = 216 tBit and T_GAP = 99 tBit + TSL. The estimate (380 + n × 300 + bytes × 11) × tBit + 75 µs is shown as a cross-check.
TTR_min = 1.20 × [n_masters × (T_token + T_msg,max) + Σ T_msg] ; WD = 6 × T_cycle or 10 × TTR × tBit 20 % margin for retries. The two watchdog rules in circulation are shown side by side; neither is normative.
Standards & methods
- IEC 61158-2 / EN 50170 — RS-485 segment length by data rate
- IEC 61158-2 Table 105 — spur capacitance
- SMAR, "Profibus-DP e Repetidores" — Tables 1 and 2 (segment, trunk, spur)
- PROCENTEC ProfiHub B2FO2+R manual — unit delay, maxTSDR, cascade
- Siemens SIMATIC NET OLM Operating Instructions 06/2022 — Tslot with fibre, fibre reach and power
- M. Felser, PROFIBUS Manual — bus parameters, cycle time, TTR
Typical reference values
| Quantity | Typical range | Note |
|---|---|---|
| Segment length (type A cable) | 1200 m ≤ 93.75 kbit/s · 1000 m at 187.5 · 400 m at 500 · 200 m at 1500 · 100 m at 3–12 Mbit/s | — |
| Total spur capacitance | 15 nF (9.6–19.2) · 3 nF (93.75) · 1 nF (187.5) · 0.6 nF (500) · 0.2 nF (1500) · 0 above | with 30 pF/m → 500 / 100 / 33.3 / 20 / 6.7 m of spur |
| Recommended trunk (SMAR Table 2) | 500 m (9.6–19.2) · 900 m (93.75) · 967 m (187.5) · 380 m (500) · 193.4 m (1500) · 100 m above | — |
| Stations per segment | 32 | an active element takes a place on both sides |
| Addresses | 0–125 usable · 126 factory default · 127 broadcast | — |
| Default slot time (DP) | 100 tBit ≤ 187.5 · 200 at 500 · 300 at 1500 · 400 at 3 M · 600 at 6 M · 1000 at 12 Mbit/s | — |
| maxTSDR | 60 tBit ≤ 187.5 · 100 at 500 · 150 at 1500 · 250 at 3 M · 450 at 6 M · 800 at 12 Mbit/s | — |
| Cyclic data per slave | 244 bytes in and 244 out | — |
| Optical margin | ≥ 3 dB | — |
Worked example
Backbone with a ProfiHub and two channels at 1.5 Mbit/s
Inputs
- Data rate / profile
- 1500 / DP kbit/s
- Cable
- type A, 30 pF/m
- Backbone — trunk
- 150 m
- Backbone — stations
- PLC-01 (class 1, addr 2), HMI-01 (class 2, addr 1), RIO-01/02 (32 in / 16 out), FT-01 (10 in) —
- FT-01 spur through a T-piece
- 3 m
- Active element at the end of the backbone
- ProfiHub B4+ (HUB-01) —
- Channel 1 — VFD-01…04 (4 in / 4 out)
- 120 m
- Channel 2 — MCC-01…03 (8 in / 4 out)
- 80 m
Results
- Backbone length (150 + 3)
- 153 m of 200
- Backbone trunk
- 150 m of 193.4 recommended
- Spur capacitance
- 0.09 nF of 0.2 (≈ 6.7 m)
- Stations backbone / channel 1 / channel 2
- 6 / 5 / 4 of 32
- Slot time needed (150 + 2 × 4)
- 158 tBit (default 300)
- Bus cycle (10 slaves, 174 bytes)
- 3.69 ms (estimate 3.60)
- Minimum TTR with 20 % margin
- 8402 tBit (5.60 ms)
- Suggested watchdog
- 22.1 to 56.0 ms
- Highest viable data rate
- 1500 kbit/s
The network passes at 1.5 Mbit/s with every segment inside its budget: the backbone uses 153 of 200 m, the 3 m spur takes 0.09 of the 0.2 nF allowed and one hub adds only 8 tBit to a slot time that has 300. The tool also finds that 1.5 Mbit/s is the ceiling for this layout: at 12 Mbit/s the backbone fails (153 m against 100 m) and the FT-01 spur becomes forbidden. Moving to 500 kbit/s would also pass, but the cycle and the watchdog roughly triple (65 to 168 ms).
Common mistakes
- Counting only the trunk. The segment limit is trunk plus spurs: at 12 Mbit/s a 150 m trunk with a 3 m spur fails twice — 153 m against 100 m, and a spur where none is allowed.
- Hanging a station on a spur at 3 to 12 Mbit/s. Above 1.5 Mbit/s the allowed spur capacitance is zero; the station goes straight on the trunk through the connector, or on a hub channel, which is a segment and not a spur.
- Branching the bus at a passive point. Only an active element can have more than one trunk child; a branch at a T-piece or station fails the network.
- Forgetting that the repeater counts in both segments. It takes a station place on its input and on each channel it opens.
- Cascading hubs without looking at time. Each unit adds delay in both directions; past the default slot time the master gives up waiting for the reply, and past the manufacturer's published count the cascade is not guaranteed.
- Assuming a network without a class 1 master is fine because the physical layer passes. With no master scanning the slaves there is no cyclic exchange, and the network fails.
Frequently asked questions
How long can a PROFIBUS DP cable be?
On type A cable, 1200 m per segment up to 93.75 kbit/s, 1000 m at 187.5 kbit/s, 400 m at 500 kbit/s, 200 m at 1.5 Mbit/s and 100 m from 3 to 12 Mbit/s — trunk plus spurs. Longer networks are built by adding segments with repeaters or hubs, or by crossing the distance on fibre.
Are spur lines allowed in PROFIBUS DP?
Up to 1.5 Mbit/s, limited by total capacitance (0.2 nF at 1.5 Mbit/s, about 6.7 m of type A cable for the whole segment). Above that they are forbidden. A hub channel is not a spur: it is a terminated segment of its own, which is why a 12 Mbit/s design with a hub closes.
How many repeaters or hubs can be cascaded?
What limits the cascade is time. The tool adds the delay of each unit in the path and of the fibre, doubles it (request and reply) and adds maxTSDR to get the slot time needed. It also applies the count published by the hub manufacturer for default bus parameters; failing either one fails the network, unless the designer has raised the slot time, in which case the formula governs.
How does the tool calculate the DP cycle time?
Per slave, a request and a reply of 300 tBit of structure plus 11 bits per cyclic byte; the cycle adds the token and the GAP maintenance with the full slot time. The published estimate formula is shown alongside and the difference between the two is reported.
What happens at 45.45 kbit/s?
That rate has no published spur, trunk or slot-time table in the sources used. Spurs are checked against the slower neighbour as an upper bound; what cannot be checked comes out as not assessed, never approved, and the rate is never suggested as "highest viable".
Glossary
- Segment
- Stretch of RS-485 bus between two terminators, with its own length, spur and station budget.
- Active element
- Repeater, hub or optical link module that refreshes the signal and opens new segments; it adds delay to the slot time.
- Spur line
- Branch from a T-piece to a station; stations chained through the D-sub connector are trunk, not spur.
- Slot time (TSL)
- Time the master waits for a slave reply before declaring it missing; it must cover maxTSDR plus the round-trip delay of the path.
- TTR
- Target token rotation time — the time budget for one token round across all masters.
- HSA
- Highest station address — the master only searches for new stations up to this address.