SINUMERIK NCU 7-Segment 3346 Error: Troubleshooting Boot Failure

Problem Statement

On a SINUMERIK 840D sl control (NCU 730.3 PN) installed in a DMG MORI 5-axis machining center, the following symptoms appear simultaneously after a normal shutdown and subsequent power-up:

• All NCU status LEDs illuminate steady ORANGE (no flashing pattern).
• The integrated 7-segment display flashes 3-3-4-6 in sequence and repeats.
• The connected operator panel reports "Warning: Connection established - no DHCP server available".
• The thin client / HMI is stuck in the TCU boot loader while attempting to obtain an IP address.
• A Bihler/Wiedemann Profibus gateway on the same machine reports "Profibus error - no connection".
• Power cycling the NCU does not clear the fault.

The combination of a stuck boot phase, a static 7-segment code, and a missing DHCP offer indicates the NCU is not completing its Linux-based service stack and therefore cannot hand an IP configuration to the TCU (Thin Client Unit) or expose the Profibus stack to the gateway. Power cycling fails because the failure is upstream of the user-mode initialization stage, not a transient runtime error.

SINUMERIK NCU LED Boot State Reference

The Siemens SINUMERIK 840D sl documentation defines the NCU front-panel LED behavior during the boot sequence. The states relevant to this fault are summarized in the table below. See the official SINUMERIK 840D sl Operating System NCU (IM7) - ID: 109481527 entry in the Siemens Industry Online Support portal for the complete boot phase diagram.

Decoding 7-Segment Code 3346

The 7-segment display on the NCU is a multi-digit state indicator. Each digit represents a milestone the boot sequence has reached. The 3346 pattern is interpreted as follows:

The repetition of the entire 3346 pattern indicates the NCU is in a retry loop at the server-connection stage. The 3-3 portion completes successfully every cycle, but the 4-6 portion never advances. This is consistent with a network-layer failure (no DHCP lease) or a corrupted system image that has lost the configured network parameters.

Symptom-to-Phase Mapping

Each observed symptom ties back to a specific NCU subsystem. Use this matrix to avoid chasing the wrong root cause.

Root Cause Hypothesis

Based on the boot-state reference and the 3346 decoding, the leading hypotheses are listed in order of probability. The first three account for the majority of field incidents on NCU 720 / 730.3 PN hardware.

1. Corrupted or uninitialised system CF card. The NCU 730.3 PN boots from a CompactFlash card on the front of the unit. If the card has lost its partition table, file-system superblock, or the /etc/hosts hostname binding that corresponds to digit 3 of the 3346 code, the system will reach the 3-3 milestone and then loop on the network stage.
2. Missing or invalid network configuration. The TCU thin client relies on DHCP from the NCU. If the NCU's network script fails to bring up eth0 or br0 (the SINUMERIK service bridge), the TCU cannot obtain a lease, producing the warning message observed on the HMI.
3. DHCP server process not running on the NCU. The NCU runs an internal DHCP server (dhcpd) for the TCU subnet. A corrupted init script, missing dhcpd.conf, or a failed bind to the bridge interface will keep the NCU in the 4-6 loop without offering a lease.
4. BIOS / firmware corruption. Less common, but a failed BIOS flash or an interrupted firmware update on the NCU 730.3 PN can leave the unit in a state that displays a steady orange LED without progressing to the kernel phase.
5. Hardware fault on the NCU mainboard. Failing RAM, a damaged network PHY, or a defective CF card slot can produce this exact symptom pattern. Confirm by swapping the CF card into a known-good NCU of the same type.

Diagnostic Procedure

Work through the following checks in order. Do not skip the network-side validation - the HMI's DHCP warning is the single most useful indicator you have that the NCU is alive but unable to complete its service handshake.

1. Confirm 7-segment code is repeatable. Power the machine off for at least 30 seconds, then on. Note whether the 3346 pattern reappears with identical timing. A different code (e.g., 3301, 3350) points to a different failure mode.
2. Check the CF card is fully seated. Open the NCU front cover, press the ejector on the CompactFlash card, reseat it, and power up. The card must be the original Siemens part (e.g., 6FC5310-1AX11-0AA0 or higher revision) with the SINUMERIK image.
3. Validate the TCU link. Connect a service laptop directly to the NCU's X120 (TCU) port. Set the laptop to DHCP. If no lease is offered within 30 seconds, the NCU's internal DHCP server is not running - confirming a Phase 4 failure.
4. Validate the customer-network port. Connect to the NCU's X130 (plant network) port. The NCU should respond on its configured service IP. If the network stack is down, no response will be received.
5. Read the seven-segment digits and the orange LED frequency. A steady orange LED (not the 0.5 Hz slow flash of Phase 2) means the kernel never started cleanly. This points toward CF card corruption rather than a runtime configuration fault.
6. Inspect the Profibus gateway. The Bihler/Wiedemann gateway reporting "Profibus error - no connection" is a downstream effect. Do not attempt to reconfigure the gateway before confirming the NCU's PROFIBUS master (X136 / X136.1) is active.

Network and DHCP Verification

The TCU's DHCP warning is a direct readout of the NCU's network service status. Validate the network stack in this order.

1. From a service laptop connected to the X120 (TCU) port, run:
   ipconfig /release && ipconfig /renew (Windows) or sudo dhclient -v eth0 (Linux). A successful lease proves the NCU DHCP server is running.
2. If no lease is offered, attach a serial console to the NCU's X101 (COM) port at 115200 8N1. The boot log should show the script that brings up br0 (or the legacy eth0) and the call to dhcpd. A missing or aborted dhcpd entry confirms a service-stack failure.
3. Check the dhcpd.conf on the CF card. The default SINUMERIK image writes lease definitions for the TCU subnet (typically 192.168.214.x). A missing or zero-byte file will cause the 4-6 loop without raising a clear alarm.
4. Verify the bridge interface MAC address. The NCU's br0 MAC is derived from the CF card's persistent configuration. If the file /etc/sysconfig/network/ifcfg-br0 is missing or has a malformed MAC entry, the bridge will fail to bind and the DHCP server will not start.

CF Card and BIOS Recovery

If the diagnostic procedure above points to a corrupted CF card image, recover the system using the steps below. The CF card is a service-level spare and is not hot-swappable in production. Power down the NCU before removing it.

1. Power off the machine and wait for the orange LED to extinguish completely.
2. Open the NCU front panel and eject the CompactFlash card using the ejector mechanism. Note the part number printed on the label (for example, 6FC5310-1AX11-0AA0).
3. Insert the card into a Linux workstation with a CF card reader. Mount the partitions read-only first to avoid further corruption.
4. Run fsck -n /dev/sdX1 on the first partition. Repeated read errors indicate the card must be replaced.
5. If the file system is intact, compare the contents against the most recent service backup (typically /card/system/ and /card/user/). The hostname file at /etc/HOSTNAME and the network configuration at /etc/sysconfig/network/ are the most likely corruption points for a 3346 fault.
6. If a service backup is available, restore the system image using the SINUMERIK service menu (see the archive restore procedure below). If no backup is available, the card must be re-imaged with the matching SINUMERIK service pack (e.g., 4.95 SPx or 5.x as appropriate to the machine build sheet) and then re-commissioned.
7. Reinsert the CF card, power up the NCU, and observe the 7-segment code. A successful boot will progress through 3346 to 0 and finally to a steady green LED.

Bihler/Wiedemann Profibus Gateway Checks

The Bihler/Wiedemann gateway (AOI, ABW, or BWU series depending on the machine vintage) is a Profibus DP slave on the NCU's Profibus line. The "no connection" error confirms the Profibus master on the NCU is not yet active. Resolve the NCU boot first; the gateway error will clear automatically once the master is up.

1. Verify the Profibus cable is connected to the correct NCU port (X136 for channel 1, X136.1 for channel 2 on the NCU 730.3 PN).
2. Confirm the Profibus terminator switch on the gateway is set per the machine drawing. Missing termination is a common field fault on retrofitted machines.
3. After restoring the NCU, use the SINUMERIK diagnostics menu to confirm the gateway is visible in the Profibus device list. The gateway's GSD file must be installed in the SINUMERIK project.
4. Do not attempt to change the gateway's Profibus address before the master is up. Address changes made while the master is down will not persist on the next master startup.

Archive Restore Workflow

When the CF card image is intact but the configuration is corrupted, an archive restore is faster than a full re-image. The SINUMERIK service menu (accessed by the appropriate service login) supports restoring from a USB stick or from a network share.

1. Locate the most recent commissioning archive. DMG MORI machines typically store this on the service laptop or in the customer's backup share under MACHINE_NAME\SINUMERIK\archive\.
2. Copy the archive to a FAT32-formatted USB stick. The file name must follow the SINUMERIK convention: arc.arc or *_arc.arc.
3. Insert the USB stick into the operator panel front USB port.
4. From the service menu, navigate to Commissioning > Archive > Restore. Select the USB source and the matching archive.
5. Confirm the restore. The NCU will reboot automatically. Observe the 7-segment code during the next boot. A clean restore typically passes through 3346 to 0 within 90 seconds.
6. If the 7-segment code returns to 3346 after the restore, the archive itself contains the corrupted file. Restore from a known-good earlier archive, or re-image the CF card from a Siemens service pack.

Verification Checklist

Use this checklist before returning the machine to production. Each item must be confirmed independently.

Troubleshooting Matrix

Use this matrix when the initial diagnosis is ambiguous. Match the observed 7-segment code to the most likely root cause.

Safety and Operational Notes

• Power down and lock out the machine before opening the NCU front cover. The NCU 730.3 PN is not hot-pluggable and the CF card is ESD-sensitive.
• Always create a fresh commissioning archive before attempting any recovery. A failed recovery without a backup forces a full re-image, which is significantly longer.
• Do not interchange CF cards between NCU 720 and NCU 730 units. The hardware abstraction differs even though the form factor is identical.
• If the NCU 7-segment remains in any non-zero code for more than five minutes, the boot has failed. Do not continue to wait; initiate the recovery procedure.
• Escalate to Siemens service support through the official Siemens Industry Online Support portal if the NCU 7-segment code is not documented in the machine's commissioning manual. Quote the NCU part number, the CF card part number, and the SINUMERIK service pack version when raising the case.

FAQ