Resolving TI545-1101 Communications Error TI-Soft RS232

David Krause13 min read
PLC HardwareSiemensTroubleshooting
Licensed PE Working through this on a live machine? A Maine-licensed engineer can take it from here — included with IMD hardware, by the hour for everything else. Book an engineer

Problem Description: TI545-1101 Stuck in Communications Error State

The SIMATIC TI545-1101 CPU module is a member of the legacy Texas Instruments / Siemens 545 family of PLCs, typically deployed in OEM industrial machinery such as curing ovens, conveyor lines, and process skids. After an extended power-down period (commonly a seasonal shutdown or "winter sleep"), the controller may boot to a healthy indicator state — BATT GOOD lit, PLC GOOD lit — but refuse to enter RUN mode. When the field engineer presses F4 from the TI-Soft batch file under Windows 95 / Windows 98 to upload or download the program, the software returns a generic "communications error" and aborts the session.

The symptom set is consistent across multiple field reports:

  • Power-up self-test appears normal (BATT GOOD, PLC GOOD LEDs steady).
  • CPU does not transition to RUN; the mode key or soft-key command to start is ignored or returns an error.
  • TI-Soft on the engineering workstation cannot establish a session on Port 1.
  • An HMI / SCADA visualization on Port 2 may still be partially responsive but cannot upload or modify logic.
  • Repeated attempts to change baud rate, toggle battery switch SW9, or replace the lithium battery do not restore programming-port communications.

Because the original programming vendor is often out of business when these legacy systems fail, the field engineer is left to recover the controller with only the documentation shipped in the original equipment cabinet. This article documents the full diagnostic ladder and the cable-level root cause that resolves the issue in the majority of cases.

Hardware Reference: TI545-1101 CPU Module

The TI545-1101 is the standard CPU for the SIMATIC TI545 rack. It executes the TI-Soft (TISOFT) instruction set and supports up to 1 MB of RAM backed by a lithium battery. Two serial ports are provided on the front faceplate:

Port Type Typical Use Default Baud (auto-negotiated by TI-Soft)
Port 1 (COMM 1) RS-232C, DB-25 female TI-Soft programming, modem, point-to-point HMI 300 / 1200 / 9600 / 19200
Port 2 (COMM 2) RS-232C / RS-422 selectable HMI / SCADA peer, third-party device 300 / 1200 / 9600 / 19200

Both ports use the Siemens/TI proprietary session protocol layered on top of standard asynchronous serial framing (8-N-1 by default). TI-Soft automatically scans the supported baud rates and locks onto the one the CPU is currently configured for, so a manual baud rate change in the software is normally not required. This auto-negotiation is one of the most important facts the field engineer should internalize: a "communications error" that does not respond to a baud change is almost never a baud-rate mismatch.

Reference: The TI545 / TI555 System Manual (Siemens attachment 17518465) documents the CPU serial interface, the Remote Base Controller (RBC) communication behavior, and the configurable "outputs on communication error" mode that defines what the controller drives when peer comms drop. Always confirm the exact port assignments against the as-built drawings before swapping cables.

Root Cause Analysis

The TI545-1101 has three failure modes that present identically to the operator but require different remedies. Before swapping hardware, isolate the layer at which the fault exists.

Layer Suspected Cause Evidence
1. Program loss Battery discharged during long storage; user program corrupted or wiped CPU enters RUN, then immediately faults, or refuses RUN; variables are all zero
2. Port / cable Wrong cable pinout, broken shield, swapped TX/RX, RTS/CTS loopback missing TI-Soft reports "communications error" instantly, no echo of any baud; HMI on Port 2 also drops out if cable swapped
3. Hardware CPU power supply or base failure during power-up surge PLC GOOD LED flickers, redundant PS indicators fault, or base generates a parity / watchdog code

In the documented case, the HMI on Port 2 was partially responsive. That single observation is decisive: it proves the CPU is alive, the power supply rails are healthy, and the lithium-backed RAM is intact enough to maintain at least the port-2 driver. The fault is therefore confined to Layer 2 — the physical and logical connection between the programming workstation and Port 1.

Why a "BATT GOOD" LED Does Not Guarantee Program Integrity

The BATT LED monitors only the battery's terminal voltage under a small test load. It does not verify that the user program checksum matches the stored image, nor that the RAM refresh is happening within specification after months of unpowered storage. A battery that is just above the BATT-LED threshold can still drop out for a few microseconds during a power-up inrush, scrambling one or more program words. The CPU then boots with a partially valid program, refuses RUN, and accepts no download because the session handshake cannot complete against the corrupted memory map.

Diagnostic Procedure (Field-Proven Sequence)

Execute these steps in order. Each step either isolates a layer or restores a known-good state.

  1. Confirm indicator state. BATT GOOD, PLC GOOD must be steady. If PLC GOOD blinks, suspect a base / power-supply problem and stop — swap the PS first.
  2. Power down, pull the CPU module from the rack (handle the module by the card rails; do not touch the gold edge connector).
  3. Wait 10 seconds to allow all internal charge to dissipate. This clears the watchdog, the serial port FIFOs, and any latched fault state.
  4. Set battery switch SW9 to OFF before reinsertion. With the battery disabled, the CPU will power up into a known uninitialized state, which paradoxically is more tolerant of a fresh program download than a corrupted image.
  5. Reinsert the CPU, apply power. Observe PLC GOOD. It should illuminate within 3–5 seconds without flicker.
  6. Attempt a programming session on Port 1 first. If the HMI is on Port 2, leave its cable connected — the CPU supports simultaneous sessions on the two ports when both are configured for different peer addresses.
  7. If Port 1 still errors, swap the cable for a known-good TI545 programming cable. The DB-25 pinout is not a straight-through RS-232 extension; it is a null-modem style crossover with handshaking. Pin-for-pin reference is given below.
  8. Try both ports with the programming cable. If the HMI is using Port 2 with a custom cable, disconnect it temporarily so the port is free for TI-Soft.
  9. Use a different engineering workstation if available. A failed UART on the P5-133 MHz PC's COM1 (common after 20+ years) will produce the exact same "communications error" message.
  10. Once a session is established, download the program from the saved .bat archive. Verify the checksum reported by TI-Soft against the source archive.
  11. With the program verified, turn SW9 back ON to re-enable battery backup. The PLC will now hold the program through the next power cycle.

TI545 Programming Cable Pinout (DB-25 to DB-9 / DB-25)

CPU Port 1 (DB-25F) Signal PC COM (DB-9M) PC COM (DB-25F)
2 TXD (out from CPU) 2 3
3 RXD (in to CPU) 3 2
4 RTS (out from CPU) 7 4
5 CTS (in to CPU) 8 5
6 DSR (in to CPU) 6 6
7 Signal Ground 5 7
20 DTR (out from CPU) 4 20

Critical points:

  • Pin 2 and Pin 3 must be crossed between the CPU and the PC. A straight-through cable will produce zero echo and an instant communications error.
  • RTS (pin 4 on the CPU) must loop back to CTS (pin 5) on the CPU side if the PC does not assert RTS, or to DSR (pin 6) on the PC side if the application software requires local handshaking.
  • Shield must be bonded at the PC chassis end only; floating shields pick up the inverter / SCR noise typical of an industrial oven cabinet and corrupt characters.
Electromagnetic interference caveat: Industrial ovens use SCR or triac phase-angle firing on the heating elements. The high dv/dt on the AC mains injects common-mode noise onto every cable in the cabinet. A programming cable that worked perfectly in the office will fail intermittently on the shop floor if the shield is not terminated and the cable is run parallel to the heater feeders. Route the programming cable at right angles to the heater cables and bond the shield at one end only. Reference: Siemens TIA Portal — Fault Types, Causes, and Corrective Measures.

Resolution: Cable + Cold-Start Reload

In the documented case, the original P5-133 MHz workstation was retained and a separate vintage laptop with a fresh TI-Soft install was brought on site. The correct TI545 programming cable was identified and used. With SW9 off, the CPU was cold-started, the saved .bat archive was downloaded, and the controller returned to RUN with all I/O and HMI sessions restored.

The two specific changes that mattered:

  1. Correct cable. The original cable had the TX/RX pair straight-through rather than crossed. Without the cross, the CPU never saw the TI-Soft session request and the workstation never saw the CPU's response — the link was effectively one-way in both directions, which is the textbook symptom of a null-modem pinout error.
  2. Cold-start with SW9 OFF. Removing the battery during reinsertion forced the CPU into a clean uninitialized state, eliminating the corrupted memory image that was preventing the session handshake from completing even after the cable was corrected.

Verification Checklist

Confirm the following before signing off the job:

  • TI-Soft Online > CPU Status reports the CPU model (TI545-1101) and firmware revision matching the nameplate.
  • The program checksum displayed by TI-Soft matches the checksum of the saved .bat archive.
  • CPU is in RUN, the RUN LED is steady, and the mode key is in the REMOTE position.
  • HMI / SCADA on Port 2 resumes normal polling and writes to setpoints are echoed back in the operator screen.
  • Toggle SW9 ON. Power down the cabinet for 30 seconds. Power back up. Confirm the program is retained and the controller returns to RUN within 10 seconds (cold start) or 3 seconds (warm start).
  • Document the cable part number, port assignment, baud rate, and checksum in the cabinet drawing package for the next outage.

Spare-Parts Strategy for Aging TI545 Installations

The TI545 line has been superseded by the SIMATIC S7-300, S7-1200, and S7-1500 families. New TI545-1101 CPUs, lithium batteries (catalog 2587746-8001 or equivalent), and base rack power supplies are increasingly difficult to source. A pragmatic spare-parts posture is:

Item Criticality Recommended Stock
TI545-1101 CPU module High 1 spare per 5 deployed, tested and sealed in ESD bag
Lithium battery (3.6 V, RAM backup) High 2 spares per cabinet, rotate on a 3-year cycle
Base rack power supply Medium 1 spare per site
TI545 programming cable (DB-25) High 2 spares, labeled and zip-tied to the inside of the cabinet door
TI-Soft installation media (floppy or CD) and a known-good laptop with RS-232 High One per site, stored in the engineering office

Migration Path: From TI545 to a Current SIMATIC Platform

When the customer is ready to retire the TI545, the migration is a structured project rather than a drop-in swap. The relevant destination families are:

  • SIMATIC S7-1200 / S7-1500 — for greenfield-style replacements where the I/O count is moderate (≤ 256 points) and EtherNet/IP or PROFINET connectivity is desired.
  • SIMATIC S7-300 — for drop-in retrofits that preserve the existing 19-inch rack footprint and many of the I/O module catalog numbers.
  • LOGO! 8 — for very small standalone oven controllers with fewer than 24 I/O and a single setpoint loop.

Migration deliverables that should be on the SOW:

  1. Reverse-engineering of the TI-Soft ladder / SFC into the equivalent S7 program (manual or via Siemens migration services).
  2. I/O assignment cross-reference table (TI545 slot / channel → S7 slot / channel).
  3. HMI re-development: WinCC flexible or TIA Portal WinCC replacing the legacy Win95 visualization.
  4. Communication migration: replace the RS-232 programming cable with EtherNet/PROFINET for programming and HMI.
  5. Parallel-run period: typically 4 to 8 weeks, with the legacy TI545 in standby, hot-standby, or active-backup mode.

Field Engineering Notes & Edge Cases

  • Old batteries, new controllers: A fresh-from-the-box TI545-1101 may also fail to communicate on first power-up if the battery shipped in a discharged state. Always measure the battery terminal voltage at the CPU connector; replace if below 3.4 V.
  • COM port virtualization on modern PCs: USB-to-RS-232 adapters (FTDI, CH340) do not reliably support the hardware handshaking the TI545 expects. Use a native COM port or a high-quality adapter with full RTS/CTS/DTR/DSR breakout.
  • TI-Soft version drift: TI-Soft versions 5.x and 6.x have slightly different default session timeouts. If the session drops after 30 seconds of inactivity, increase the session timeout in the OPTIONS dialog to 600 seconds.
  • Port 2 RS-422 mode: If Port 2 is configured for RS-422 (used by some HMI packages), it will not respond to an RS-232 probe from TI-Soft. Move TI-Soft to Port 1.
  • Watchdog timer: If the original program is unrecoverable, a fresh program written from scratch will run, but the RBC output-failsafe settings (defined in the TI545 / TI555 System Manual) must be re-entered. Outputs controlled by the RBC will revert to their configured fault state (last value, zero, or one) on the next communications loss.

Troubleshooting Matrix (Quick Reference)

Symptom Likely Layer First Action
"Communications error" immediately on F4 Cable / port Swap programming cable, confirm null-modem pinout
TI-Soft cannot find CPU on any baud PC COM port / cable Loopback test the PC port; try a different workstation
CPU in RUN, HMI on Port 2 works, Port 1 dead Port 1 driver / cable Try Port 1 on a different PC; inspect connector pins
CPU will not enter RUN, PLC GOOD blinks Base / power supply Measure PS rails (5 V, ±12 V, 24 V); replace PS
Program loads but values are scrambled RAM / battery Replace battery, cold-start with SW9 OFF, reload
Intermittent errors, worse when heater is on EMI / shield Re-route programming cable, bond shield at PC end only

FAQ

What does the "communications error" message in TI-Soft actually mean?

It is a generic failure of the session handshake between the TI-Soft driver and the CPU serial port. The error does not distinguish between cable, baud, port, or program-state causes. Treat it as "I could not complete a session" and walk the diagnostic ladder — port, cable, cold-start — before swapping the CPU.

Does the TI545-1101 auto-detect the baud rate?

Yes. TI-Soft scans the supported baud rates (300, 1200, 9600, 19200 bps) when the F4 key is pressed and locks onto the rate the CPU is using. Manually changing the baud in TI-Soft is not necessary and rarely helps. The exception is when a non-TI-Soft device is using the port — that requires matching the configured rate from the CPU's port configuration.

Why does a "BATT GOOD" LED not guarantee the program is intact?

The BATT LED monitors terminal voltage under a small test load; it does not verify RAM contents or program checksum. A marginal battery can drop out for microseconds during a power-up inrush and corrupt a few program words, leaving the CPU unable to start a TI-Soft session. Always cold-start the CPU with SW9 OFF and reload the program from a verified archive before assuming the CPU is healthy.

Can I use a USB-to-RS232 adapter to program a TI545?

It is not recommended. TI-Soft expects full hardware handshaking on RTS, CTS, DTR, and DSR. Many low-cost USB adapters (especially CH340-based units) do not drive all four handshake lines. Use a native COM port on a vintage PC or laptop, or a high-quality industrial-grade adapter with full breakout of all modem-control signals.

Where can I find the official TI545-1101 manual and pinout reference?

The official SIMATIC TI545 / TI555 System Manual is published by Siemens as attachment 17518465 on the Siemens Industry Online Support portal: SIMATIC TI545/TI555 System Manual (PDF). For EMI and fault-cause guidance applicable to any SIMATIC serial port, see the Siemens TIA Portal fault-types-and-corrective-measures reference.

Back to blog