Troubleshooting
Troubleshooting S7-1200 PROFINET servo motion control issues
Short answer
Fix an S7-1200 PROFINET servo project in order: compile the motion calls in FB1 first, set drive and TIA Portal to the same telegram — Telegram 3 on this pairing — then prove one axis holds enable and velocity through repeated scans. A first-generation CPU 1214C executes independent single-axis motion only; the flying shear's deterministic master-slave cut belongs in the drives or in a motion-class controller, and no telegram adds it.
On this page
- Why are the MC blocks red in FB1 with instance DB errors?
- Does the CPU 1214C support PROFINET servo control with MC blocks?
- Should the drive and TIA Portal be on Telegram 3 or Telegram 12?
- Why does the motor stop about 5 seconds after MC_SpeedControl starts it?
- Why does MC_Halt not stop the axis?
- What does a 17-bit encoder without a battery change?
- Can a first-generation S7-1200 run the flying shear master-slave cut?
- Questions asked
Work this fault from the top down: make the motion calls compile, prove cyclic telegram exchange, prove one stable axis, and only then face the flying shear. Red MC blocks with instance-DB errors inside FB1 are a compile problem, not a PROFINET fault, and no telegram change repairs a call that cannot compile. The decision at the end matters most: a first-generation CPU 1214C executes independent single-axis motion over PROFINET, but the deterministic master-slave synchronization a flying shear needs is beyond what it runs, so that cut belongs in the drives or in a motion-class controller. The sections below take you to a healthy single axis fast and mark the ceiling exactly.
Why are the MC blocks red in FB1 with instance DB errors? #
Red highlighting on MC_Power, MC_Reset, MC_MoveVelocity, or MC_MoveJog inside FB1 is a compile-time configuration error — the call has invalid instance data or a wrong axis reference — and it will not clear by fixing the network, the telegram, or the drive.
Work in this order:
- Fix every compile error until the project compiles hardware and software together with zero errors.
- Match the drive and the TIA Portal submodule to the same telegram and prove cyclic exchange with the axis idle.
- Prove one axis end to end: continuous enable, one velocity command, and a halt that actually stops it.
- Only then decide the two-axis architecture, which is where the 1214C hits its ceiling.
Open the compiler messages for FB1 and treat the first error as the root of the correction sequence; later errors are usually fallout from the first invalid declaration. Each motion instruction gets its own instance DB created for that call: sharing one between calls, or editing it as ordinary application data, keeps the call red. The axis reference on every call must point at the configured axis object with the expected data type — a similarly named PLC tag is not an axis reference. Compile hardware and software together and expect zero errors tied to the axis, its drive interface, and the motion calls. If compilation still fails, stay on this step: changing Telegram 12 to Telegram 3 cannot repair an invalid instance declaration.
Does the CPU 1214C support PROFINET servo control with MC blocks? #
Yes — the 1214C runs S7-1200 Motion Control against a PROFINET servo axis, and every block this project needs is in its single-axis command set: MC_Power, MC_Reset, MC_Home, MC_MoveJog, MC_MoveVelocity, MC_SpeedControl, MC_Halt.
Keep the three layers separate while you diagnose. The programming layer is the motion call and its instance DB. The axis layer is the technology object that binds engineering units and motion commands to the drive interface. The network layer exchanges cyclic process data through the installed GSDML device description.
The documented behavior you will rely on: MC_Power enables and disables the axis, and MC_MoveVelocity starts on a positive edge at Execute, holds the set velocity, and signals target velocity reached — the function chart labels that indication InVel_1 — until a newer command or an error ends it. That one-command-at-a-time, started-by-an-edge character is the whole capability class of this set. Whether the controller class matches the drive's motion features is a pairing decision that repeats across vendors — the same question drives why a Kinetix 5500 needs a motion-capable Logix controller.
Decision: commission single axes on the 1214C and expect that to work; do not expect firmware or telegram changes to turn it into a synchronized-motion controller.
Should the drive and TIA Portal be on Telegram 3 or Telegram 12? #
Set both ends to the same telegram, and for this drive-and-CPU pairing set both to Telegram 3 together — the drive parameter and the TIA Portal submodule must name the same telegram number or cyclic exchange never starts, whatever the byte counts suggest.
A telegram is the PROFIdrive profile's fixed cyclic contract: which control words the controller writes and which status and actual-value words the drive returns, every cycle. The number lives in the device description — the GSDML you installed — on the TIA Portal side and in the drive's own parameter set on the other side, and one vendor's telegram numbering is not a universal layout: similar byte counts do not make two telegrams interchangeable. The mismatch failure mode is absolute. A drive left on Telegram 12 under a Telegram 3 submodule is a configuration mismatch, cyclic I/O stays down, and the axis has no command or actual-value path at all.
So make it one coordinated change: select Telegram 3 in the drive, replace or reconfigure the matching submodule in TIA Portal, reconnect the technology object's drive interface, compile, download, and restart the affected equipment by the product procedure. Either endpoint left on Telegram 12 invalidates the test.
Acceptance, with no motion commanded, is a list:
- Actual-value data updates every scan.
- Device diagnostics report the drive reachable with no cyclic-I/O or configuration fault.
- The axis interface stays available continuously, not for a few seconds.
The cyclic contract itself is walked end to end in controlling a PROFINET servo drive with cyclic telegram data.
Why does the motor stop about 5 seconds after MC_SpeedControl starts it? #
A stop that arrives about 5 seconds after start is a state change somewhere — enable dropped, command error, drive fault, or lost cyclic exchange — and the diagnosis is to capture which indication changes first, because that first transition picks the branch.
Stable here means continuous enable and command ownership through repeated scans, not several seconds of rotation. Call MC_Power cyclically for the same axis and expect its enabled status to stay true while the application requests power. Then issue one velocity command and watch four things in the same watch table: the enable chain, the motion block's outputs, the drive fault state, and PROFINET device diagnostics. The 5 s value itself is a symptom; the first indication to change is the fault:
| First indication to change | What it means | Next action |
|---|---|---|
| Enable or power status drops | the PLC sequencing removed authority | trace the logic writing the enable and remove one-scan or timer-dependent control |
| Motion block reports an error | the axis rejected or aborted the command | read the error ID and the axis diagnostics before any reset |
| Drive fault bit sets | the drive stopped motion on its own | read the drive fault record, correct the condition, then call MC_Reset |
| PROFINET device drops offline | cyclic exchange was interrupted | correct device naming, addressing, wiring, or the configuration fault shown online |
| All states healthy, speed goes to 0 | another command or application write took ownership | cross-reference every motion call for that axis and their execution conditions |
A stop a few seconds in, with every status still healthy, usually lands in the last row: a timer or a second sequence in the application, not the network.
Why does MC_Halt not stop the axis? #
MC_Halt stops the axis only when its Execute sees a positive edge on a scan where the call runs, it references the moving axis, and no newer motion command wins arbitration — a halt inside a branch later scans skip, or retriggered speed commands, leave the shaft turning.
Calling MC_Halt once inside a branch that later scans skip makes diagnosis ambiguous, and continuously retriggering MC_SpeedControl or MC_MoveJog lets a newer command replace the stop. Check four things:
- Watch the halt request and the halt block's outputs in the same watch table; expect the block to become active when the stop is commanded.
- Confirm the halt call and MC_SpeedControl reference the identical axis object; exactly one sequence owns motion at any instant.
- Suppress new jog, velocity, and speed-control triggers while the halt is active; expect commanded speed to ramp to zero on the configured deceleration.
- If the halt reports an error, record the error ID before calling MC_Reset — resetting first destroys the branch evidence.
If the halt is active and commanded speed is 0 while the shaft still turns, check feedback scaling and the drive interface. If the halt never becomes active, fix the PLC sequencing; no telegram touches this. While any stop command is in question, the stop you trust is the hardwired one: keep the drive's enable or safe-torque-off input outside the PROFINET path so torque drops without the PLC.
What does a 17-bit encoder without a battery change? #
A 17-bit absolute magnetic encoder gives 131,072 increments per revolution (2 to the 17th), but running it without a battery makes it incremental in practice: position survives only while powered, so plan a reference move at every power-up instead of trusting a remembered position.
The encoder configuration on the axis side must match the drive's reported increments-per-revolution figure, or every actual position the axis reports is scaled wrong. Because nothing retains the count through power-off, the axis has no position at start-up: put an active homing or the drive's reference move into the startup sequence, and design now for the case where the machine cannot home immediately because tooling sits in the travel path.
For the shear goal this matters twice. The slave needs a valid reference before it can hold any position relationship, and the master measurement must be continuous — an encoder the application reads occasionally cannot serve as a shear master.
Can a first-generation S7-1200 run the flying shear master-slave cut? #
No — treat the synchronized cut as out of scope for this CPU: the deterministic master-slave follow a flying shear needs is not added by any telegram, so put the synchronization in the drives with the master measurement wired to them, or move to a motion-class controller.
A flying shear is a continuously coupled position relationship: the slave's position is a defined function of the master's, evaluated every drive cycle through the cut window. The command blocks this CPU generation runs against a PROFINET axis are single-axis blocks — one command at a time, started by an Execute edge, arbitrated per block. Telegram 3 fixes cyclic exchange and stops there; no telegram number couples two axes.
Worked example, with inputs you set to your machine's numbers: web at 300 mm/s, cut length 600 mm, a knife window of 150 mm in which the blade must match web speed, and axis dynamics of 3,000 mm/s² both ways. The 10.0 acceleration and deceleration printed in the MC_MoveVelocity function chart are illustration values, not inputs to commission with.
- Cut period: 600 ÷ 300 mm/s = 2.0 s between cuts.
- Match window: 150 ÷ 300 mm/s = 0.5 s at web speed.
- Reaching web speed at 3,000 mm/s²: 300 ÷ 3,000 = 0.1 s per ramp, each ramp covering 0.5 × 3,000 × (0.1 s)² = 15 mm.
- Return stroke: 150 + 15 + 15 = 180 mm to recover within 2.0 − 0.5 − 0.2 = 1.3 s, a mean return speed of 180 ÷ 1.3 ≈ 138 mm/s.
The speeds are easy for any servo; the coupling is not. The slave must track the master inside that 150 mm window at drive rate, and that coupling is the capability the 1214C does not have. Two architectures carry it instead:
- Put the synchronization in the drives. Wire the master measurement straight into the slave drive so the follow runs at drive rate instead of network rate — that is what an auxiliary encoder input for the master position buys you. The drive holds the relationship itself, as in setting up e-cam for a flying shear cut, while the PLC keeps sequencing, start commands, and supervision.
- Or move the shear into a controller class built for synchronized motion, and run sizing motion hardware for a cut-to-length flying shear as the capacity check before you choose.
Keep the 1214C on Telegram 3 doing single-axis work either way, and do not spend commissioning days trying to turn a telegram number into a coupling feature — that number moves words every cycle; it never synchronizes axes.
Questions asked
Does every motion control block in FB1 need its own instance DB?
Each motion control instruction gets a separate instance DB created for that call and matched to the instruction type. Two calls cannot share one, and the DB is not ordinary application data: editing or retyping it is what keeps the call red with an instance error. If a call stays red after wiring, regenerate the instance DB from the call itself instead of repairing it by hand.
When does MC_Reset actually clear an axis fault?
MC_Reset clears the error state of the axis technology object, and only once the underlying drive condition is really gone. A reset issued while the drive still holds its fault gets rejected, which is why the drive fault record comes first: read it, correct the condition, then edge the reset and confirm the axis returns to a ready state.
Should the shear master value reach the slave drive over PROFINET or through a wired input?
Through a wired input at the drive for a shear. A master value routed through the controller arrives on PLC scan and network cycles, and each of those cycles adds delay between master position and slave correction. A pulse or encoder input at the slave drive is corrected at drive rate, which is the only rate tight enough for a cut window measured in millimetres.
Is a PLC stop command a safe stop for this axis?
No. MC_Halt and every other network-commanded stop is an operational stop that depends on the scan, the telegram, and the drive staying healthy. The safe stop is the drive's enable or safe-torque-off input wired outside the PROFINET path so torque drops without the PLC. Test that hardwired stop with the network cable pulled, and treat a stop that only sometimes works as broken.