P0 Publisher Protocol Upgrade

This document records the first engineering-grade publisher upgrades for ARSVIN. The scope stays intentionally narrow: ARSVIN is a Sampled Values publisher, not a process-bus analyzer.

Implemented in this package

Multi-ASDU Sampled Values publishing

ARSVIN now accepts SCL streams with nofASDU from 1 to 8. The publisher packs sequential samples into one SavPdu frame:

• nofASDU=1: one sample per Ethernet frame.
• nofASDU=2/4/8: sequential samples are packed as multiple ASDUs in one frame.
• smpCnt is incremented per ASDU, not per Ethernet frame.
• publication frame rate is calculated as sampleRateHz / nofASDU.

The current hard limit is 8 ASDUs per frame. Higher values are blocked by preflight to avoid accidentally generating oversized or non-interoperable frames before dedicated test vectors exist.

Second-aligned sample counter start

The runtime initializes smpCnt from the fractional UTC second when publishing starts. This is closer to synchronized SV publisher behavior than always starting at zero. It remains a lab/compatibility behavior and is not a certified PTP implementation.

Publisher validation and frame preview

A publisher-side validation layer now checks the SCL stream before live TX:

• APPID presence and zero APPID warning.
• destination MAC presence, multicast sanity, and common SV multicast range warning.
• svID / smvID presence.
• DataSet resolution.
• confRev, smpRate, smpMod, and nofASDU readiness.
• payload layout support.

Preflight also reports a frame preview summary: nofASDU, sample rate, publication rate, and payload bytes per ASDU.

Quality bit simulation foundation

The engine now has SampledValueQuality helpers for common publisher simulation states:

• good
• invalid
• questionable
• oldData
• test
• operatorBlocked

The UI currently publishes good quality by default. The next UI step is exposing this as a compact quality selector per stream or per channel.

Generated PCAP evidence foundation

The engine now has a small generated-frame PCAP export helper. This is not a network analyzer; it only writes frames produced by the publisher so they can be opened in Wireshark or shared as evidence.

Golden protocol tests

Unit tests were added for:

• multi-ASDU frame build/parse round-trip.
• ASDU sample counter rollover.
• wrong payload batch rejection.
• frame preview publication rate.
• quality bit encoding.
• second-aligned sample counter.
• generated PCAP header/output.

Still intentionally not included

• full network analyzer
• process-bus scanner
• subscriber diagnostics
• GOOSE/SV mismatch live diagnosis
• certified PTP grandmaster/slave operation
• calibrated analog source behavior

Next P0 follow-up recommended

1. Expose SampledValueQuality in the WPF UI as a compact stream-level selector.
2. Add one-click dry-run PCAP export from the WPF app.
3. Add golden PCAP files under tests/golden/ and compare byte-level expected output.
4. Add nofASDU=2, nofASDU=4, and nofASDU=8 sample SCL files to CI test data.
5. Add jitter/p95/p99 TX health statistics to the publish status bar.

P0.1 Publisher evidence update

This revision keeps ARSVIN scoped as a publisher, not an analyzer.

Added evidence-oriented publisher functions:

• TX quality preset selection for the selected publisher slot: good, invalid, questionable, oldData, test, operatorBlocked.
• Runtime payload quality fields now follow the selected preset instead of always emitting good quality.
• Preflight shows a warning when a non-default quality preset is selected.
• Generated PCAP export writes ARSVIN-created SV Ethernet frames without capturing live network traffic.
• Publish status includes per-publisher q= labels so operator intent is visible during live/dry runs.

Generated PCAP export is meant for offline evidence in Wireshark or third-party SV tools. It is not packet capture and it does not turn ARSVIN into a network analyzer.

P0.2 TX Timing Health update

This revision adds publisher-side timing health only. It does not capture external packets and does not make ARSVIN a process-bus analyzer.

The runtime records, per enabled publisher slot:

• target publication rate, derived from sampleRateHz / nofASDU;
• actual publication rate observed from the TX loop;
• target frame interval;
• average and maximum absolute jitter;
• late frame count;
• missed schedule count;
• average and maximum send duration.

The status bar shows a compact global summary using the worst status from all enabled publishers:

TX Timing: GOOD act=1599.8/1600.0fps jitter=8/42us late=0 missed=0 send=6/18us maxLate=20us

Status meaning:

• GOOD: frame rate and local TX timing are stable enough for lab publishing.
• WARN: the publisher is running but jitter/late frames are visible.
• BAD: the runtime missed at least one schedule slot, actual FPS is significantly low, or jitter is larger than one target interval.

This is a Windows/Npcap runtime health indicator, not certified PTP or hardware timestamp evidence. Use it to determine whether the PC/NIC is strong enough for the selected SV profile.