Diagnose PCS 7 V9.0 SP2 Trend Gaps When Archive Data Is Missing

When PCS 7 V9.0 SP2 trend curves stop recording for tens of seconds while the OS server remains connected to the AS, first determine whether values are absent from the archive or merely missing from the displayed trend. In this report, the operator later said the values were absent from the database, and isolating other devices from the plant bus was followed by a problem-free observation period. That correlation makes plant-bus interaction worth investigating, but it does not establish a universal third-party-device cause or a confirmed permanent fix.

Distinguish an archive gap from a trend display gap

A trend can fail to show values for more than one reason. If the stored archive contains samples for the interval but the trend does not display them, investigate the trend time range, selected tags, archive source, and rendering path. If samples are actually absent from the archive store, the issue is an acquisition or archival gap; a display-side limit alone does not explain missing stored records.

The discussion includes a reply claiming a Tag Logging SCS display restriction of 120 MB. No Siemens reference for that value or its applicability was established during this review. More importantly, the reporting operator said the corresponding data was absent from the database, so the display-buffer explanation does not account for that reported condition unless the database check itself was incorrect.

Interpret the plant-bus observation carefully

The installation had seven PLCs on the plant bus plus other devices not included in the PCS 7 multiproject. The archive issue was intermittent, and the operator reported that after isolating the other devices the system had run without the problem so far. This is useful before/after evidence, but the observation was still ongoing and no packet capture, switch diagnostics, or root-cause analysis was supplied. It supports testing the network hypothesis; it does not prove that all non-PCS 7 equipment is incompatible with a PCS 7 plant bus.

Troubleshooting sequence

Record the exact time of the next trend gap and whether current process values and alarms continued updating. Save the relevant WinCC/PCS 7 system and Tag Logging diagnostics around that time.
Determine whether values were written for the affected time interval using the site's approved WinCC archive inspection method. Do not rely only on what the trend graph draws. Keep any database inspection read-only and version-appropriate; this article does not prescribe undocumented SQL queries or table names.
Compare the affected interval with a working interval and, if possible, a different trend/archive on the same OS server. Record the tags, archive configuration, sampling/archiving settings, and whether the loss affects one archive or several.
Map every device sharing the plant bus, including devices outside the PCS 7 multiproject. Work with the network and control-system owners to inspect switch-port errors, link events, traffic changes, and topology during the time of the gap.
If operationally safe and approved, compare behavior with the unrelated devices isolated or on a properly engineered separate network segment. The source reports this as a provisional field observation, not as a guaranteed fix. Do not disconnect plant equipment or reconfigure switches on a live system without authorization and rollback planning.
If gaps persist with the network condition controlled, investigate the OS server's Tag Logging/Archive System diagnostics, storage availability, archive configuration, and project compile history. A full OS compile with memory reset was suggested in the discussion but was not tried in the reported case; do not perform it as a generic remedy without version-specific justification and change control.

How to treat the proposed 120 MB explanation

The reply says Tag Logging SCS shows data up to 120 MB and suggests that the trend may omit periods because of a display restriction. That numerical limit and its exact meaning were not verified in Siemens documentation during this research. Distinguish compressed-segment/display behavior from whether samples exist in archive storage: confirm stored values first. Do not change archive compression, segment sizes, SQL structures, or other internal settings based on the unverified number.

Expected results

A controlled diagnosis should classify each event as either stored data that is not displayed or an actual gap in stored samples, and identify whether the loss correlates with a particular archive, OS station, network event, or device topology. The source says isolation was followed by a trouble-free observation so far, but it does not provide the length of that observation or prove a lasting correction.

Verification

For each reported gap, compare the exact time interval against the approved archive inspection method and the Trend Control view. Confirm whether the record is present in archive storage, whether online values remained current, and whether other archives on the same OS were affected. After an approved network or project change, observe through a period comparable to the prior intermittent recurrence window (reported as about two to five days) and record the results. This is an observation plan, not a claim that the issue was lab-tested or solved.

Limitations and Operational Precautions

The ticket lacks trend screenshots, detailed archive diagnostics, actual Tag Logging configuration, database evidence, traffic captures, and a completed post-isolation observation. The cause and exact fix therefore remain unresolved. Do not treat 120 MB as a confirmed PCS 7 V9.0 SP2 limit. Network segmentation, device isolation, compile/download actions, and database inspection can affect a running process-control system; coordinate them with operations, automation engineering, and IT/network personnel, and use approved backups and rollback procedures.

Engineer FAQ

Does a connected OS server prove that archived trends are recording? No. A live server/controller connection and archive storage are distinct functions; verify stored records for the affected interval.

Does the case prove that third-party devices on the plant bus caused the gaps? No. The operator reported improvement after isolation, but the observation was ongoing and no network-level cause was demonstrated.

Should I run a full compile with memory reset immediately? No. It was suggested but not attempted in the case. First establish whether this is an archive-storage gap and review diagnostics; use compile/download changes only when justified by the project and version-specific guidance.

Technical References

Siemens, Compendium Part A – Configuration Guidelines (V9.0), Configuration Manual, 03/2018, A5E43228900-AA, §4.3.7 “Configuring the plant bus,” p. 43. It documents PCS 7 plant-bus interface/protocol configuration; it does not establish that non-project devices caused this archive symptom: https://support.industry.siemens.com/cs/attachments/109756485/pcs7_compendium_part_a_en-US_en-US.pdf
Siemens, PCS 7 V9.0 – List of messages, including OS Archive System / Tag Logging messages. The official support PDF returned HTTP 403 during this review; the indexed Siemens text lists archive-configuration read errors, but does not identify the cause of this ticket's intermittent gaps: https://support.industry.siemens.com/cs/attachments/37696303/37696303_pcs7_messages_V90.pdf

Diagnose PCS 7 V9.0 SP2 Trend Gaps When Archive Data Is Missing