Troubleshooting the TP170A 'Memory Card (Save Memory)' Alarm

The SIMATIC TP170A is a compact 5.7-inch operator panel from the SIMATIC TP170/270/370 family, configured with ProTool or ProTool Lite and widely deployed on S7-200, S7-300, and LOGO! controlled machines. The panel has been discontinued for many years but remains in service on legacy lines, packaging machines, and small process skids. A recurring field failure is the runtime alarm "Alarm memory card (save memory)" appearing on the display, which halts the connected PLC program at the second step of the operator sequence and forces a power-cycle to recover.

This reference walks through the alarm mechanism, the documented difference between the TP170A and TP170B memory card subsystems, the internal flash failure mode that mimics a missing card, and the verified recovery procedure with a configuration patch and a hardware-replacement fallback.

1. Affected Hardware and Firmware Identification

Before changing anything, confirm the exact panel variant. The 170-series housing is shared across multiple models, and the memory card slot is physically present on the TP170A even though its firmware does not fully implement the recipe/log card feature set.

2. Alarm Text and Behavior on the TP170A

The on-screen message is rendered by the TP170A runtime in one of the following localized forms:

• English: Alarm memory card (save memory)
• German: Alarm Speicherkarte (Speicher sichern)
• French: Alarme carte mémoire (sauvegarder la mémoire)

Observed runtime symptoms on field units:

1. The alarm appears during the second operation step, typically the first time the project reads a tag area or a recipe record.
2. The panel continues to display the active screen, but writes from the HMI to the PLC are blocked.
3. Outputs driven by HMI commands (start, jog, setpoint) stop responding.
4. A power cycle restores operation for a variable interval, then the alarm recurs.
5. The PC card slot is empty; no CF or PCMCIA card is fitted.

The alarm text is generated by the runtime when it evaluates the memory-management task and detects that a card-handle pointer or a save-buffer descriptor is invalid. On a TP170A where the configuration never references a card, this should not happen unless either the internal flash is failing or a corrupted configuration is loaded.

3. Root Cause Analysis

Three root causes are documented in the Siemens TP170A knowledge base, listed in order of field frequency.

3.1 Configuration Cross-Reference from a TP170B Project

Plants that previously operated a TP170B sometimes migrate the original ProTool .pdb project to a TP170A without re-targeting the device. ProTool stores the destination device in the SystemSettings\Target section of the project. A TP170B-targeted project loaded onto a TP170A activates recipe and log areas that the TP170A runtime tries to write to the card slot. When no card is present, the runtime emits the alarm and freezes any HMI-to-PLC write that depends on the recipe or log buffer.

Verification: open the project in ProTool/Pro Configuration Tool, select System > Target, and confirm the target is set to TP170A, not TP170B. The OP/Target dialog box displays the panel resolution and color depth; mismatches are common after a project has been cloned across machines.

3.2 Internal Flash Wear-Out

The TP170A uses an on-board NOR flash device for persistent storage of the project, runtime parameters, and a small log buffer. The flash is rated for approximately 100,000 erase cycles per sector. Units in continuous service for 8+ years, especially those on three-shift operations, reach end-of-life. The runtime interprets the bad-block table as a missing save destination and raises the card alarm as a generic save-memory error.

Indicators that flash, not configuration, is at fault:

• The alarm appears within 30 to 120 seconds of a cold start.
• Recompiling and reloading the project through ProTool does not suppress the alarm.
• The boot screen shows an extended initialization sequence (longer than 20 s) with the progress bar stalling.
• Powering the panel from a known-good 24 V supply with ripple under 50 mV pk-pk does not change behavior.

3.3 Corrupted Boot Image

A failed flash write (brown-out, EMC event, or interrupted transfer during a ProTool download) can leave a partially valid boot image. The runtime boots, but the memory card driver module is loaded with invalid pointers. The alarm then triggers on the first memory-management call. The standard recovery is a full image re-flash using ProTool and the serial transfer cable (6XV1 440-2K... or USB-to-RS232/PPI multimaster cable on a host PC).

4. Diagnostic Procedure

Use the following ordered checklist before changing hardware. Each step is non-destructive and adds evidence to the root-cause decision tree.

1. Capture the project file. Connect a PC running ProTool/Pro V6.0 SP3 or later over the panel's RS-232/PPI port. Use File > Transfer > Transfer from Panel > Backup Project to read the current runtime project from the TP170A. Save the .pdb to a known folder. If ProTool cannot establish a connection, the flash is likely compromised; jump to step 5.
2. Inspect the target device. Open the backup in ProTool Configuration Tool. Navigate to System > Target. Confirm the panel type. If the value reads TP170B or TP170B color, the configuration mismatch is confirmed.
3. Check the recipe and log area pointers. Select Recipes > Recipe Container in the project tree. On a TP170A these containers should be empty or absent. If a recipe container is defined, right-click and note the storage location: a value of Memory Card on a TP170A is the trigger for the alarm.
4. Compare with a known-good project on a sister machine. Many plants have several TP170A panels running the same machine. If only one panel raises the alarm and the project backup on that panel differs from a working sister panel, the project file is the issue.
5. Test the panel on the bench. Disconnect the panel from the PLC and supply 24 VDC. Create a minimal ProTool project containing a single screen with one indicator tag, transfer it, and run it. If the alarm recurs on a minimal project, the flash is degraded and a panel replacement is the only durable fix.
6. Inspect the card slot pins. With power removed, use a magnifier and inspect the PC card slot pins for bent contacts, contamination, or solder joint cracks. A bent pin can short the card detect line and cause the runtime to believe a card is in the process of being written, triggering the save-memory alarm.

5. Solution Path A: Configuration Correction

When the diagnostic in step 2 of section 4 confirms a TP170B project loaded on a TP170A, the fix is a project rebuild and re-transfer. The procedure is as follows.

1. Launch ProTool/Pro Configuration Tool on the engineering PC.
2. Open the .pdb backup retrieved from the panel.
3. Select System > Target and change the target to TP170A. ProTool will prompt to scale the project to the TP170A resolution and color depth. Accept the conversion.
4. Open Recipes. Delete any recipe container whose storage location is set to Memory Card. Change the storage location to Internal Flash for any recipe that must be retained, or remove recipes entirely if they are not used on this machine.
5. Open Messages > Logs. Set the log destination to Internal Buffer. The TP170A internal buffer is volatile across power-cycle; if persistent logs are required, recommend a TP177 or KTP400 Basic as a modern replacement rather than relying on a TP170A flash for log storage.
6. Compile the project (File > Compile > All) and check the Output window for warnings referencing the card subsystem.
7. Connect to the panel, place it in transfer mode by holding the lower-right soft key area during power-up (or selecting Transfer from the panel's service menu), and download the rebuilt project.
8. Watch the boot sequence. The runtime should not raise the save-memory alarm.

For projects that must remain on a TP170A but still require save-memory semantics, configure the Save Memory function in ProTool under System > Save Memory and set the destination to Internal Flash. This disables the card path and satisfies the runtime's expectation of a valid save target without requiring a card.

6. Solution Path B: Hardware Replacement

When the diagnostic in step 5 of section 4 confirms a flash failure, the field-proven replacement options, in increasing order of effort, are:

1. Replace the panel with a like-for-like TP170A spare. The 6AV6 545-0BA15-0AX0 and 6AV6 545-0BB15-0AX0 are mechanically and electrically compatible. Transfer the corrected ProTool project from the engineering PC, attach to the PLC, and verify the alarm is gone.
2. Replace the panel with a TP177 micro or KTP400 Basic. These panels are still in distribution and support modern TIA Portal projects. They use the same MPI/PPI connector pinout for legacy S7-200/LOGO! networks. A migration project in WinCC flexible (TIA Portal) can import the original ProTool project using the Migration Tool; verify area pointers for tags, recipes, and alarms after migration. Note that absolute address references (e.g., DB1.DBD0) require re-mapping in the new project because the tag database is rebuilt.
3. Replace the on-board flash (advanced, repair-only). The flash IC is a TSOP-48 NOR device. This is a depot-level repair. After replacing the part, the panel must be re-flashed with a known-good image using the Siemens ProTool Service Tool. This path is rarely economical versus panel replacement.

7. Memory Card Subsystem Reference

The shared alarm text between the two panels is one of the primary sources of confusion. The TP170A raises the same alarm string for any save-memory subsystem failure, including internal flash, because the alarm is generated by a common module. Engineers reading the alarm text and concluding that a card is required frequently insert a card and find that the alarm persists. The card slot is not the underlying problem on a TP170A.

8. Verification

After applying Solution Path A or B, run the following verification sequence on the machine before returning the line to production.

1. Power-cycle the panel three times in succession. The boot time should be consistent (under 25 s) and no alarm should appear.
2. Run the operator sequence that originally triggered the alarm. Confirm that the second-step tag reads and writes complete without an alarm popup.
3. Trigger a save-memory action from the panel's service menu (System > Save Memory if exposed, otherwise simulate the event with a tag-driven recipe save). The action should complete within 5 s with no alarm.
4. Leave the panel powered continuously for 24 hours. Re-check for the alarm. On a flash-degraded panel, alarms often recur within this window; a clean 24-hour run is a strong indicator of resolution.
5. Document the panel order number, firmware version, project file name, and date in the maintenance log. The TP170A is at end of life; capturing the asset state supports future migration to a TP177 or KTP panel.

9. Preventive Measures

For sites that continue to operate fleets of TP170A panels, the following measures reduce the probability of recurrence.

• Project governance. Maintain a master .pdb file per machine. After every project edit, the target device in System > Target must be verified. A simple Excel index of Panel MLFB, ProTool Version, Project Name, Date Compiled, Target Device catches mismatches at the engineering stage.
• Spare parts pool. Keep at least one tested TP170A spare in the parts crib. The spare should be loaded with the most common machine project and labeled clearly. Spare units older than 5 years should be re-flashed annually to refresh the internal flash and verify boot integrity.
• Power-quality audit. Many TP170A panels are powered from the same 24 VDC bus as solenoids and VFDs. Brown-outs below 19 V for more than 20 ms are a documented trigger of flash write interruption. Add a DC UPS or a 24 V buffer module (e.g., SITOP) to protect the panel supply.
• Temperature management. The TP170A is rated to 50 °C ambient. Cabinet temperatures above 40 °C accelerate flash wear. Verify cabinet ventilation or add filtered fans if the cabinet houses a VFD or a transformer.
• Migration roadmap. Build a multi-year migration plan to replace TP170A panels with TP177 micro or KTP400 Basic panels. The newer panels support the same MPI/PPI communication and reduce the support burden for legacy parts.

10. Field Commissioning Notes

When commissioning a replacement panel, observe the following wiring and configuration conventions to keep the panel stable in service.

• Use a shielded MPI/PPI cable with the shield grounded at one end only. The shield ground is typically at the panel end; confirm with the cabinet grounding diagram.
• Set the panel's MPI address to a value that does not collide with the PLC, the drive, or any other HMI. Default is 1; many legacy systems use 1 for the PLC, so 2 is a common safe address for a TP170A.
• Set the baud rate to 187.5 kbps for MPI, 9.6 kbps or 19.2 kbps for PPI to an S7-200. Mismatched baud rates do not raise the save-memory alarm, but they do prevent tag updates and are a common commissioning trap.
• Set the panel's time-of-day source to either the PLC or an internal RTC. A failed RTC battery on the TP170A can cause the runtime to repeatedly attempt a save of the buffered log, which in turn can trigger the save-memory alarm on a flash-degraded unit.

11. Quick Reference Fault Code Matrix

12. Related Documentation

For the canonical Siemens guidance on memory card handling, refer to the TIA Portal documentation for Unified Comfort Panels, which codifies the same handling rules that apply to the TP170A. The relevant page covers the principles of inserting and removing memory cards, the associated alarm text conventions, and the prohibition against removing a card while data is being accessed. Engineers supporting TP170A panels should treat those rules as a baseline for any panel that exposes a memory card slot to the runtime.