S7-300 MPI Network: Maximum Stations and CPU 312 Limits

S7-300 MPI Network: Maximum Stations and CPU 312 Connection Limits

Overview

The MPI (Multi-Point Interface) protocol on the S7-300 family imposes two independent constraints that engineers routinely conflate. The first is the physical bus limit: a single MPI segment supports a maximum number of node addresses, and the bus terminates with repeaters or RS-485 terminators. The second is the per-CPU connection resource, which is a finite pool of PG, OP, S7, and S7-standard connections defined by the S7-300 firmware. The CPU 312 is one of the most constrained controllers in the S7-300 line, and any topology that exceeds its connection budget will fail to establish links at runtime even though the bus is wired correctly.

This reference walks through the bus-level limits, the CPU 312 connection breakdown, the configuration steps in HW Config, the highest MPI address parameter, and a worked example of a 13-node network (9 CPUs + 3 OPs + 1 WinCC PC) spread across three independent machines. Every value is keyed back to the CPU 312 manual and the Siemens SIOS FAQ 299124 on maximum OP connections per S7-300 CPU.

MPI Bus-Level Physical Limits

The MPI bus is an RS-485-based, half-duplex, token-passing network that runs at 19.2 kbps, 187.5 kbps, or 1.5 Mbps. Each bus segment has hard electrical and addressing ceilings:

The 32-node cap applies to addresses in a single electrical segment, but the "Highest MPI address" (sometimes called "HSA") network parameter is selectable at 15, 31, 63, or 126. The HSA is not the physical cap; it is the highest address the token will visit and is normally set to the highest configured node so the token rotation stays efficient. For a 13-node network, the HSA should be set to 13 (or 15, the next standard option) so the token does not waste time probing unused addresses up to 31.

CPU 312 Connection Resources

The S7-300 CPU connection budget is split into four configurable connection types in the CPU properties dialog of HW Config (or TIA Portal device configuration). The CPU 312 manual lists the following ceiling:

The "6 connections" figure on page 308 of the S7-300 CPU 31xC and CPU 31x manual is the total. Page 309 of the same manual breaks these into the four categories. The 5-OP maximum is a hard firmware ceiling on the CPU 312: even if you raise the OP slot in HW Config to a number greater than 5, the firmware will reject the configuration at download. CPU 312C, 312 IFM, and 313 variants have identical or smaller budgets; only the CPU 315 family (and the 317/319 lines) raise the ceiling to 8, 12, or 16 connections respectively.

Connection Cost Per Node Type

Count connection cost from the perspective of the CPU whose budget is being consumed. The table below shows typical costs for a WinCC Runtime + S7-300 HMI architecture:

Worked Example: 9 CPUs + 3 OPs + 1 WinCC PC

The source scenario is three independent machines. Each machine has three S7-300 CPUs (one per subsystem) and one OP for local operation. The three machines are not currently networked to each other; the engineering team wants to add a single WinCC PC that monitors all three machines. The MPI topology is a single segment spanning all three machines.

Step 1: count the nodes on the bus.

• CPUs: 3 machines x 3 CPUs = 9 nodes
• OPs: 3 machines x 1 OP = 3 nodes
• WinCC PC: 1 node
• PG (occasional, docked for service): 1 node
• Total: 14 nodes

Bus-level cap of 31 is satisfied. The HSA should be set to 14 (round up to 15 to keep token rotation on a standard boundary).

Step 2: count the connections consumed on each CPU. The WinCC PC polls all 9 CPUs, and each local OP also polls its three local CPUs.

Each CPU 312 consumes 3 of its 6 available connections. The 6-connection budget is large enough to support this topology. The OP budget of 5 per CPU is more than sufficient; each CPU only uses 1 OP slot. The original concern that the CPU 312 would not work is unfounded as long as the connection count per CPU remains under 6 total and under 5 OP.

CPU 312 vs Larger S7-300 CPUs

If the connection budget ever becomes a concern, upgrade the CPU rather than try to multiplex OP links. The relevant ceilings from the S7-300 family manual:

The CPU 315-2 DP at 16 total connections is the typical upgrade target when a project grows beyond the 312 budget. Note that the PG connection is implicit and does not consume a configured slot on most CPU 312/314 firmware versions; verify on your specific firmware with the online diagnostics buffer if you see "No resource available" alarms during online attach.

Configuring Connections in HW Config

Connections are not automatic. They are explicitly enabled in the CPU properties before download. The procedure is the same in Step 7 V5.x and TIA Portal (with cosmetic differences in the dialog labels):

1. Open the S7 project in Step 7 Manager (or TIA Portal).
2. Open HW Config and select the CPU 312 in the station.
3. Right-click the CPU and select Object Properties (Step 7) or Properties > Communication (TIA Portal).
4. Open the Communication tab.
5. Set the number of reserved connections in each row: PG, OP, S7 basic, S7 standard.
6. Confirm that the sum does not exceed the CPU total (6 for CPU 312).
7. Compile and download the hardware configuration to the CPU.

MPI Address and Highest Station Address

The "Highest MPI address" parameter is set per CPU and per CP/OP in HW Config. The four selectable values are 15, 31, 63, and 126. According to the PTC Kepware S7 MPI driver reference, the value must be at least as large as the highest configured MPI node on the bus. If the HSA is 15 but a node has address 22, that node will be invisible to the token-passing protocol and the CPU will appear offline.

For the 13-node example, set HSA = 15 on every CPU. Do not mix HSA values across nodes on the same bus; mismatched HSA causes intermittent connection drops and token timeouts.

MPI Address Assignment Scheme

Unique MPI addresses are mandatory. If two devices share an address, the bus will lock up with token-passing errors and the SF/BF LEDs on the CPUs will blink. Recommended scheme for the 13-node example:

Step-by-Step: Building a 13-Node MPI Network in Step 7

1. Create a new Step 7 project named with the plant identifier.
2. Insert three SIMATIC 300 stations, one per machine, using Insert > Station > SIMATIC 300 Station.
3. Open each station, insert the rail, power supply, CPU 312 (with order number 6ES7312-1AE14-0AB0 or the project-specific variant), and signal modules.
4. Configure the MPI interface on each CPU: open Properties > MPI Interface > Parameters, set the address per the table above, and set the subnet transmission rate to 187.5 kbps.
5. Set the Highest MPI Address to 15 on every CPU.
6. Open CPU Properties > Communication and raise the OP connection count to 2 (1 for the local OP, 1 for the WinCC PC).
7. Insert a fourth PC station. Add a WinCC Application and a CP 5611 / CP 5621 (or 5613 A2) configured for MPI at address 14.
8. Open NetPro and assign every CPU and the PC station to the same MPI subnet (typically auto-created on the first CPU).
9. Compile and save. Step 7 will warn about MPI address conflicts; resolve any duplicates before download.
10. Download the hardware configuration to each CPU. Cycle power if the BF LED remains on after download.
11. From the WinCC PC, open the MPI driver configuration and verify that all 9 CPUs respond to a connection test (e.g., by browsing online tags).

Verification

After download, perform these checks in order:

1. Bus LED scan: On every CPU, the SF LED must be off and the BF LED must be off. A steady BF LED indicates a bus fault (cable, termination, or address conflict).
2. Online reachability: From the PG, select PLC > Display Accessible Nodes. The list must show all 9 CPUs, all 3 OPs, and the WinCC PC with their assigned addresses.
3. Connection diagnostics: Open PLC > Diagnostics > Connection Diagnostics on each CPU. The number of active connections must equal the configured count, and the partner endpoint must match the expected HMI/PG.
4. WinCC tag test: In WinCC Explorer, run a tag simulation. All 9 CPUs must respond within one polling cycle (typically 1 s for WinCC on MPI).
5. OP test: On each local OP, change an operator value and confirm that the corresponding CPU accepts the write and the WinCC PC sees the new state.

Troubleshooting Matrix

When to Migrate Off MPI

MPI is acceptable for the 13-node scenario above, but a migration to PROFINET or PROFIBUS-DP is recommended when any of the following become true:

• Number of nodes exceeds 31 (or 14 with a single WinCC PC, depending on CP limits).
• Connection budget on a CPU approaches its ceiling (5 of 5 OP for CPU 312).
• Bus length between machines exceeds 50 m and the cost of repeaters and fiber-optic couplers approaches that of a PROFINET switch.
• Redundant HMI panels are required (the second OP pushes a CPU 312 over its 5-OP limit).
• Cybersecurity or network segmentation is required: MPI is not routable and cannot pass through firewalls.

The CPU 315-2 PN/DP (6ES7315-2EH14-0AB0) and CPU 317-2 PN/DP are the typical S7-300 upgrade targets when PROFINET is required, with 16 and 32 total connections respectively.

FAQ