Siemens 810D PF/PS LED Faults: CCU Replacement and Recovery

Siemens 810D PF/PS LED Faults: CCU Replacement and Recovery

Field-proven troubleshooting procedure for the SINUMERIK 810D when the Compact Control Unit (CCU) module exhibits a steady red PF LED combined with a blinking PS LED. This reference covers root-cause isolation between HMI, power supply, and CCU; module hot-swap technique; memory-card structure; NCK and PLC reload from archive; battery-backed SRAM preservation; SHOPMILL/ISO mode startup; and final commissioning verification. It is written for the maintenance technician who has a faulted machine on the floor and needs an ordered, defensible procedure to follow before reaching for a replacement card.

1. Problem Overview

The 810D is a mid-range SINUMERIK CNC platform produced from approximately 1999 through the mid-2000s. It integrates the numerical control kernel (NCK), a SIMATIC S7-300 class PLC, the HMI runtime, and Profibus-DP fieldbus on a single backplane populated by a CCU, a PCU (HMI), and a PSU. The symptom most often reported by a customer calling for support on a 2002-era 810D mill is the combination of a steady red PF LED on the CCU and a blinking PS LED on the CCU or PSU. The HMI may be in a "reboot holding pattern," a SHOPMILL "would not load" banner, or a black screen with cursor only.

This combination is one of the more common CCU-fatal patterns. It is almost never a single component; it is the surface symptom of a corrupted or aged CCU that has lost its NCK/PLC image in SRAM. The next-level question is: Is the CCU the failure, or is the failure downstream (power supply, Profibus slave fault taking the bus down, HMI boot failure looping the controller)? The procedure below walks through isolation, replacement, and reload in the order that produces the fastest mean-time-to-recovery.

2. SINUMERIK 810D Architecture Refresher

Before touching the cabinet, internalize the three-module topology of the 810D, because every LED on the front of the CCU can be explained by which of these subsystems is healthy.

The CCU is a single module that hosts: (a) the NCK with the part program interpreter and axis interpolators, (b) the SIMATIC S7-300 PLC runtime, (c) the Profibus-DP master, (d) the SRAM with battery backup that holds machine data, tool offsets, and the active part program, and (e) the boot logic that hands off to the PCU once the HMI is up. There is no separate PLC CPU. When the CCU fails, all of these subsystems die together, which is why a single bad module can take down the entire cell.

3. CCU LED Meanings

Front-panel LED behavior is the single fastest diagnostic available. Map it before you crack any screws.

The classic "PF steady, PS blinking" pattern is what Siemens diagnostic flowcharts call a boot fault: the CCU is alive enough to run its POST and toggle the PS LED, but the POST is failing one of its gates and latching PF. The most common gates that fail on a 2002-vintage 810D are, in descending order of frequency:

1. SRAM/Flash filesystem integrity (file system corruption, NVRAM CRC failure, or ungraceful power-down that wrote a half-block).
2. Battery discharged — the CCU has booted, detected that SRAM contents are not valid, and latched PF as a safety.
3. PLC boot block on the MMC is corrupt or missing the signature S7 expects.
4. Profibus master cannot complete slave enumeration within the timeout (a missing or dead ET200 slave that wasn't there before).
5. Internal regulator out of tolerance (rare on field-aged units; common on units that have cooked for a decade in a 40 °C cabinet).

4. Power Supply Verification Before Module Swap

Resist the temptation to swap the CCU on first suspicion. The cheapest and fastest fix is sometimes the PSU, and an uncontrolled CCU swap can wipe SRAM if the battery is also flat. Perform the following checks in order, with the cabinet energized and locked.

1. 24 VDC input to the PSU. Measure at the PSU input terminals with a calibrated DMM. Acceptable range: 20.4 V to 28.8 V. Many plant supplies sag under load when the cabinet fan starts; if you see <21 V on the input, the PSU cannot regulate and you will chase ghost faults all afternoon.
2. 5 VDC backplane rail. Measure at the backplane test points (consult the wiring diagram for your specific frame; common test points are TP1/TP2 on the backplane). Acceptable range: 4.85 V to 5.20 V. Anything outside 4.75 V to 5.25 V will trigger the CCU's undervoltage monitor and produce a PF latch.
3. 15 V and -15 V analog rails. The analog I/O and spindle encoder interfaces need ±15 V. Measure both rails. Common failures: electrolytic capacitor aging in the PSU, producing ripple on the 5 V rail that the DMM in DC mode won't catch (use AC mode and look for <50 mV ripple).
4. 24 VDC I/O bus. Many 810D cells also feed the 24 V Profibus termination and the ET200 power through the same PSU. A shorted ET200 can drag the 24 V down. Isolate by pulling Profibus connectors one at a time and re-observing PS and PF.

If the rails are clean, the failure is downstream of the PSU and you are authorized to suspect the CCU. If the rails are dirty, fix the PSU or its source first, then re-evaluate. Skipping this step is the most common cause of "I swapped the CCU and now I have two faulted units."

5. CCU Module Replacement Procedure

The good news is that CCU replacement on the 810D is hot-swap-friendly in the sense that the form factor and connectors are uniform; the bad news is that the battery-backed SRAM lives on the CCU, so a CCU swap without planning will cost you machine data and tool offsets. The procedure below preserves as much state as possible.

1. Verify the mem card is present and read it on a PC first. The mem card (MMC 100.2 / 6FC5 447-0AAxx-0AA0 family) lives in the PCU, not the CCU. Power down, eject it, and image it on a Windows machine with a PCMCIA-to-USB adapter. File names to look for:

2. Copy the entire MMC contents to a backup directory with timestamps. You will want this for the warranty claim against your replacement CCU, and it is your safety net if the replacement boots with a different firmware version and rejects the original archive format.
3. Document all LED states and any error code displayed on the HMI (if it will boot that far). Take photos.
4. Power down, lock out, and verify zero energy. Use an ESD wrist strap when handling the CCU. The CCU is a class-1 ESD-sensitive device; walking across a concrete floor and then picking up the module is enough to destroy it.
5. Remove the CCU. Release the top and bottom latches (some frames use a single side lever). Pull straight out; do not rock. Place the module on a static-safe mat.
6. Install the replacement CCU of the same catalog family. Mismatched CCU part numbers (for example, swapping a 6FC5 210-0DF22-2AA0 for a -0DF31) will not boot without a firmware match, and they will reject the existing mem card archive. The rule of thumb: match the last three characters of the catalog number, or call Siemens with the original serial number before swapping.
7. Re-seat the mem card in the PCU if you removed it.
8. Power up. Watch the LED sequence: PS should blink for 10-30 s, then go solid. PF should remain off. STOP should go off after the PLC user program loads. The HMI should reach either the SHOPMILL home screen or the ISO channel ready screen within 60-90 s.

6. Memory Card Structure and Series Startup

The 810D stores its identity in two places: the CCU's battery-backed SRAM (volatile across power-down if the battery is dead) and the mem card (non-volatile). After a CCU swap, the SRAM is empty; the mem card is your source of truth. The procedure to bring a "clean" CCU to the same identity as the previous one is the Siemens "Series Startup" or "Inbetriebnahme-Serie" flow.

On a 2002-era 810D with SHOPMILL loaded, the HMI will, on first boot with a foreign CCU, prompt: "Series startup with archive?". Answer Yes, navigate to the archive that matches the machine serial number, and confirm. The NCK will then:

1. Erase SRAM.
2. Reload the NCK firmware from the MMC.
3. Apply the machine data (MD) from the selected archive.
4. Reset to NCK cold-start state.
5. Prompt for PLC user program load.

7. NCK and PLC Reload Procedure

If the mem card is good and the CCU is a clean replacement, follow this ordered procedure. The order is not arbitrary: NCK must be alive before the PLC will accept certain user blocks.

1. Confirm HMI is in Setup mode (keyswitch to position 3, or softkey password 11111, on machines with that protection scheme).
2. NCK reset: Setup → Commissioning → NCK → Reset (PoR). Confirm the prompt "NCK will reset, all data lost". The CCU will reboot the NCK only; PLC continues to run.
3. Series startup: Setup → Commissioning → Series startup. The HMI will list archives on the MMC. Select the one whose timestamp matches the last known good. Confirm.
4. Verify axis references appear in the channel display. If "Wait for axis reference" is missing, the MD did not apply.
5. PLC stop and reload: Switch the PLC mode selector on the CCU to STOP (or use the HMI softkey under Diagnostics → PLC → Stop). Connect a Siemens programming cable (MPI/PC adapter, 6ES7 972-0CA23-0XA0 or equivalent) and use STEP 7 / SIMATIC Manager to download the S7 project. Use the project file with the same timestamp as the NCK archive.
6. PLC run: Switch the PLC mode selector to RUN-P or use the HMI softkey to Start. Watch for SF to clear within 5 s.
7. Profibus check: Diagnostics → Profibus. All configured slaves should appear as "OK". BF1/BF2 LEDs should be off.

8. Battery Replacement and SRAM Preservation

The CCU holds a single 3.6 V lithium thionyl chloride cell, type 6FC5 247-0AA18-0AA0 (or functionally equivalent, capacity ~2.6 Ah). Battery life at room temperature is rated for roughly 3-5 years; in a hot cabinet (35 °C and up) it can drop to 18 months. A discharged battery has two failure modes that map to the symptoms in this article:

1. Slow discharge: the BAT LED comes on; the controller still works; if you replace the battery within a few weeks, no data loss.
2. Sudden deep discharge: the controller boots with default machine data; tool offsets, R parameters, and the active part program are gone; PF may latch on the next boot because the SRAM CRC fails.

Replacement procedure:

1. Power on. The CCU must be energized to preserve SRAM during the swap. Confirm the 5 V LED is solid and the PS LED is solid before opening the cabinet.
2. Locate the battery holder on the front of the CCU, bottom edge. The cell is keyed.
3. Pull the old cell straight out using the pull tab. You have 30 s before the controller notices the loss; do not dawdle.
4. Insert the new cell in the same orientation. Confirm the BAT LED goes off within 60 s.
5. If the BAT LED stays on, the new cell may be reversed or the contact is dirty. Power down, clean with isopropyl, retry.

9. SHOPMILL, ISO Mode, and HMI Boot Recovery

One of the more common follow-on symptoms after a 810D boots in a degraded state is the HMI banner: "SHOPMILL would not load". This is a red herring in the sense that the HMI is not actually broken — it is telling you that a supporting component on the MMC has not been read or has failed its signature check. The recovery sequence is:

1. Exit the error banner (press INPUT or CANCEL twice).
2. Open Diagnostics → HMI → Version. Confirm SHOPMILL version is present. If the version field is blank, the SHOPMILL app package on the MMC is missing or corrupt.
3. Re-image the SHOPMILL package from your backup to \mmc2\apps. The 810D SHOPMILL package is typically 30-60 MB; the directory must contain shopmill.ini and a \bin folder with the SHOPMILL DLLs.
4. Reboot the HMI: Setup → HMI → Reboot HMI. The HMI restarts; CCU does not.
5. Confirm SHOPMILL home screen appears.

For ISO mode, the active interpreter is in the channel configuration. If the HMI reverts to ISO and the part program will not run, the channel MD has reverted to default. Re-apply from your archive: Setup → Commissioning → Archive → Restore → Select ISO channel config → Confirm.

10. Tool Offset and Active Program Recovery

The two pieces of state that are most painful to lose on a customer machine are the tool table and the active part program. A CCU swap will lose both unless one of the following is true:

• The mem card contains a recent TOA archive that was saved before the failure.
• The tool offsets are also saved as part of the NCK machine data (rare; depends on the integrator's setup).
• The customer has a printed tool sheet from setup.

Active part program recovery paths, in order of preference:

1. Shop floor network or DNC server: many 2002-era cells had the part program on a server. Pull the file from the network; copy to MMC.
2. Mem card: if the operator ran the program in the last session, the most recent program may still be in \mmc2\user\mpf or \spf. Check both.
3. Shop's CAM system: if the part was programmed offline, regenerate from the post processor.
4. Manual re-entry: slow, error-prone, but a valid last resort. The Siemens manual Programming Manual / ShopMill for the 810D contains the program structure for every supported cycle and the G-code equivalents.

11. Field-Troubleshooting Decision Matrix

Use this matrix when the symptom is a CCU that will not come up cleanly. Work top-down: verify the cheapest, most-isolated layer first.

12. Verification and Commissioning Checks After Repair

Do not return the machine to production without a defined set of no-load and loaded checks. The 810D is old enough that a successful HMI display does not guarantee a working machine.

1. No-load motion test. From MDI, issue a safe rapid such as G0 X0 Y0 Z0. Verify the axes move, return to position, and the position display agrees with the actual position. Run at 5% rapid override to start.
2. Spindle test. M3 S500; verify direction and speed via the spindle display. Stop with M5.
3. Tool changer test. M6 T1; verify the changer cycles, the tool number in the display matches the magazine pocket, and the offset is applied.
4. Coolant and chip conveyor. M8 / M9, then verify the conveyor forward/reverse.
5. Limit switch test. Hand-jog each axis into its hardware limit. Verify the E-stop trips and the axis stops. Reset and back off.
6. Probe test (if equipped). Trigger the probe; verify the PLC picks up the input and the cycle counter increments.
7. Run a known part. Cut a coupon from a part with a known cycle time. Compare dimensions to the historical record.
8. Archive the new state. Setup → Commissioning → Archive → Save. Copy to a separate SD card or network share. Label with date and machine serial.

13. Common Pitfalls and Field-Proven Notes

• Don't trust the HMI to know which archive is the right one. Some integrators used a single archive across an entire family of similar machines, and the HMI will happily load whichever one you point it at. Always cross-check the machine serial and the integrator's name in the archive header against the machine's nameplate.
• The "wait 2-3 seconds, lights on, then 3 seconds, lights off" pattern on the CCU is a sign that the PLC has been cleared and the user program is missing. The CCU is not faulted; the PLC is simply waiting for a download. Some shops call this the "PLCs SOS" pattern.
• MMC contamination. Industrial MMC cards are PCMCIA, not SD. They are mechanical, they have 30+ pins, and they lose contact in the slot if the machine has been vibrated for a decade. If the CCU boots intermittently, eject and reseat the MMC before any further diagnosis.
• Time synchronization. A CCU swap with a dead battery will reset the clock. If the shop runs scheduled jobs, time-based part counters, or audit logging, re-set the clock from the HMI or from STEP 7 before resuming production.
• After-sales support path. Siemens service for legacy 810D hardware is time-limited. Document the local Siemens service center contact and the local third-party CNC repair vendor on the inside of the cabinet door. When the unit fault is beyond in-house reach, the fastest path is usually the third-party vendor with a stock of exchange CCUs; the second-fastest is the Siemens service center; the slowest is a fresh OEM unit on a long lead.

14. Reference Equipment and Documentation

Keep the following on hand or on the network before you need them. The list is short and the items are not expensive.

• Siemens 810D Operator's Manual and Programming Manual (current revision on the Siemens Industry Online Support portal). The functional description and the parameter manual are the two documents you will reach for first.
• STEP 7 / SIMATIC Manager, current service pack, installed on a Windows laptop with a working MPI/PC adapter (6ES7 972-0CA23-0XA0 or equivalent USB adapter).
• PCMCIA-to-USB adapter for mem card imaging. A quality unit (the kind with a real chipset, not a $5 eBay special) reads the MMC 100.2 reliably; a cheap one can corrupt the card.
• A bench PSU set to 24 VDC, current-limited to 1 A, for safe power-up of a CCU outside the cabinet for firmware flashing.
• A calibrated DMM with at least 0.5% DC accuracy. Fluke 87V or equivalent.
• A static-safe mat and a wrist strap. Treat every CCU as if it will go back into service in 30 minutes, because it will.

For controller identification, the nameplate on the front of the CCU carries the catalog number, the firmware version, and the serial number. Photograph it before you swap the module. The Siemens Industry Online Support portal accepts the catalog number directly and will return the matching manual set and the matching firmware download.