Skip to content

Releases

ARIEC60870 v3.6.17-rc.4

Pre-release

Choose a tag to compare

@github-actions github-actions released this 18 Nov 19:25

ArSee v3.6.17-rc.4 — SCADA Q engineering ranges by bay

Windows x64 · prerelease test candidate · 19 November 2026

This candidate adds the explicit SCADA data-item/bay association received from the operator on 19 November to the already merged and tested PR #263, based on master@b13f47117762f8520e9bf8eefaf54b6daa222101. The older v3.6.17-rc.3 remains immutable. This is not a stable promotion or RTU field FAT certification.

Source-specific Q scaling now visible in the bundled 31-point Signal Map

The older scada.sdb photograph provided two Q engineering ranges but did not identify which bay used each. The later operator clarification resolves that gap for this particular reference:

Operator's bay SCADA data item CA IOA IEC Type Decoded IEC signed source Engineering display range
Feeder Q Diri Sendiri 105 790447 TI11 SVA I16 −32768..32767 −2078..+2078 MVar
Kopel — Lab SOGI 3 Q Lab 105 790439 TI11 SVA I16 −32768..32767 −260..+260 MVar
Trafo — Lab SOGI 4 Q Lab 105 790443 TI11 SVA I16 −32768..32767 −260..+260 MVar

Each bundled reference Q now contains scalingMode=TwoPoint, all four endpoints, unit=MVar, and the corresponding engineering min/max. The same JSON is used for the Windows packaged seed and human-readable docs mirror. This is a source-reported engineering display map, not evidence that every field gateway's raw IEC SVA matches SCADA's internal raw domain. Compare with actual signed RX and independent SCADA measurement before accepting site calibration.

At raw signed X=0, the literal asymmetric −32768..32767 conversion yields approximately +0.0317 MVar Feeder and +0.00397 MVar Lab; the firmware does not silently clamp/alter the original source value or IEC QDS to report an attractive zero. Calibration must be checked at zero and both sign directions before approving.

Safe upgrade of an existing locally saved Signal Map

An installer does not overwrite a saved project JSON or silently remap its Q readings. Open Signal Map / Sync 31 I/O to preview matching CA=105 / exact IOA / TI11 rows; only a truly legacy, uncalibrated Factor identity row with no custom endpoints or min/max is eligible for its corresponding Feeder/Lab SOGI 3/4 Q TwoPoint reference. Any custom Factor or TwoPoint mapping, different station/Type, command config, point identity, feedback wiring, or raw record is preserved. Accept the preview and Save As a NEW JSON profile; that new copy must then be applied to the active monitoring session.

Manual per-signal SCADA reference range preview remains available with explicit confirmation, including a warning when choosing an alternative to the latest operator association; nothing is imposed on user-owned profiles.

Unchanged source semantics / continued RC.3 fixes

  • POAQ inbound TI9 normalized feedback IOA790449 remains 0..1200 MW, and distinct POOP outbound TI48 setpoint command IOA70537 remains 0..1000 command metadata. No setpoint encoder changes, physical command TX or new command privileges.
  • Original CA/IOA/TI, IEC ASDU raw HEX, QDS OV/IV, timestamps, GI/SBO and Event Log forensic evidence are unchanged.
  • Unknown extra received IOAs such as 790349, 1790446, 1790447 continue to show unmapped exact physical identities, not guessed aliases.
  • TI13 P source float identity and TI5/TI32 step-position family remain as before. Historical tap range −4..14 versus 1..19 requires separate station evidence and is not inferred here.
  • Separate old documentation-only ±1732 all-bay nominal CT/PT hypothesis is marked superseded for this source; it must not be mistaken for an approved feeder or lab mapping.

Software qualification versus physical FAT

The source PR #263 passed Windows CI, 1,249 xUnit (0 failures), 19/19 protocol smoke, Signal Map real WPF acceptance, repository hygiene, dependency review, and CodeQL. This source-level qualification is not package certification. Before publishing the public release, the exact RC.4 package build must pass full xUnit, WPF packaged executable smoke, setup EXE, portable ZIP, SHA-256, SBOM and artifact attestation.

Physical IEC101/104 FAT on SICAM/other third-party devices, actual 1200-bps SOE completeness, Lab Q calibration at negative/zero/positive samples, and independent RTU-to-SCADA engineering reference are NOT TESTED by this report. Do not promote to stable or claim field approval based on synthetic tests alone.

Expected artifacts after a successful source-bound public workflow:

  • ARIEC60870-v3.6.17-rc.4-setup-win-x64.exe
  • ARIEC60870-v3.6.17-rc.4-win-x64.zip
  • ARIEC60870-v3.6.17-rc.4-sbom.spdx.json
  • SHA256SUMS.txt

Earlier RC.3 package stays available without modification. Choose the portable ZIP first for isolated read-only evaluation.

ARIEC60870 v3.6.17-rc.3

Pre-release

Choose a tag to compare

@github-actions github-actions released this 18 Nov 11:48

ArSee v3.6.17-rc.3 — Verified Values scaling and unmapped IOA diagnosis

Windows x64 · prerelease test candidate · 18 November 2026

Built from merged PR #262 on master@5503f1ac8fefe177775217a2f800dd32a2bcab92, plus this branch's exact RC.3 version stamp and release notes. This release supersedes v3.6.17-rc.2 for evaluation, without modifying that immutable release. The last stable v3.6.16 remains unchanged.

Reproduced screenshot issues and software corrections

  • POAQ feedback at CA105, TI9, IOA790449: the operator-corrected bundled reference is TwoPoint X0=0, X100=1, Y0=0, Y100=1200 MW (separate from POOP TI48 outbound command 0..1000). Incoming normalized X=32767/32768 projects to 1199.96337891 MW, retaining any QDS OV quality indication. Previously the Values fallback could display preformatted 0.999969482422 MW even when this mapping was not successfully applied.
  • Q at CA105 TI11 IOA790447 (FEEDER), IOA790439 (KOPEL), IOA790443 (TRAFO): bundled profile remains uncalibrated counts. An incoming signed SVA=13641 projects to 13641 counts, not 13641 MVar unless an explicit profile-approved Q TwoPoint is selected. The SCADA screenshot contains two different Q ranges, ±2078 and ±260, but does not bind either range to the three specific IOAs. Blind default calibration would assert invented MVar measurements.
  • Unmapped received IOAs such as 790349, 1790446, 1790447 are not present in the supplied 31-point CA105 Signal Map. Values now uses exact physical RX identity and a one-time diagnostic instead of letting an upstream guessed signal name or mapped flag hide the discrepancy. No automatic IOA spelling/digit correction; check original HEX against the simulator/gateway database before editing a copied JSON.
  • A single source-qualified measurement projection now serves both live Values and cached Save & Apply reprojection. Only a verified canonical IEC source object and a strict CA/IOA/TI-family point can yield engineering output. Missing source proof visibly says [scaling not verified] and cannot silently append a guessed MW/MVar unit.
  • Values compact header now shows the count of active Signal Map references alongside received points. It is not a claim that all expected points were received or that every IOA has a certified calibration.

Regression tests reproduce the TI9 source value and quality flag, signed TI11 Q, cached ValueRow profile reproject, the missing-source fallback, and all three extra original IOAs. Windows Release/xUnit/protocol-smoke and CodeQL on source PR #262 passed.

Read-only acceptance checklist (virtual inline VSPE or authorized isolated RTU)

  1. Save a copy of the station's own Signal Map JSON; do not overwrite an existing site-approved profile. Start with a read-only monitor of actual traffic, not an unsolicited SCADA command.
  2. Confirm active mapping reference count, then compare Values against Signal Map. For incoming CA105 IOA790449 TI9 X≈0.999969, expect ≈1199.963 MW only if that reference profile is the active strict match, and separately see OV if wire QDS has OV.
  3. For Q with the uncalibrated default, expect signed raw integer counts, not unproven MVar. The existing More → SCADA reference range for selected signal... allows explicit single-point ±2078 or ±260 candidate previews and Ctrl+Z; Save As requires a distinct file. Obtain actual eTerra/RTU data-item-to-IOA association before using MVar.
  4. Confirm incoming 790349, 1790446, 1790447 appear unmapped with exact CA/IOA/TI/source diagnostics rather than being silently translated to 790439, 790446, 790447. Only change a saved project map if verified against original HEX and RTU/simulator source configuration.
  5. For tap IOA790448, the screenshot observed untimed TI5 and value -4 while the original mapping lists TI32/1..19; this is a pending source specification conflict. Never invent an offset (+5), convert TI5 into TI32, or mistake historical profile limits for physical travel proof.
  6. Re-check Event Log/Trace/Values original frame parity, quality, Ctrl+Z, Save & Apply and WPF rendering at 100/125/150% DPI. Record actual evidence or mark NOT TESTED.

Proof boundaries and distribution

Package publication is allowed only after its exact-source GitHub Actions workflow passes Windows xUnit, smoke, packaged EXE/WPF, installer build, ZIP structure, SHA-256, SPDX SBOM and artifact attestations. Publishing this notes file does not alone establish package success.

Neither the source CI nor Windows Package can establish real SICAM A8000 timing/reconnect, RTU FIFO 60/500/2,000 event parity, the source-specific Q MVar calibration, multi-vendor interoperability or command execution/physical switchgear travel. Draft IEC101 core #145 remains separate.

On successful publication, release assets will be named:

  • ARIEC60870-v3.6.17-rc.3-setup-win-x64.exe
  • ARIEC60870-v3.6.17-rc.3-win-x64.zip
  • ARIEC60870-v3.6.17-rc.3-sbom.spdx.json and SHA256SUMS.txt

Do not reuse or mutate the public v3.6.17-rc.1 / v3.6.17-rc.2 tag or binaries.

ARIEC60870 v3.6.17-rc.2

Pre-release

Choose a tag to compare

@github-actions github-actions released this 18 Nov 08:20

ArSee v3.6.17-rc.2 — Event FC evidence and SCADA scaling review

Windows x64 · Public release candidate · 18 November 2026

This candidate extends v3.6.17-rc.1, keeping its compact Material toolbar, Excel-like Signal Map, descriptive IEC dropdowns, and shared WPF styles. It does not replace the latest stable v3.6.16.

Changes since rc.1

  • #256: Event Log FC column shows FC10 (IEC-101 Request user data, Class 1) or FC11 (Request user data, Class 2) only with a verified source/link-qualified request–response pair; unmatched/other-protocol events show —. Hover explains that FC11 can carry Class 1 fallback and that FC10/FC11 describes the observed poll, not a permanent IOA class. The RequestFC evidence is also included in Event CSV.
  • #257: reconcile operator-supplied eTerra SCADA range reference, including the later correction POAQ TI9 measured feedback 0–1200, distinct from POOP TI48 outbound setpoint 0–1000. Default profile retains the 31 CA/IOA/TI points and safe uncalibrated TI11 Q counts. Inbound TI9 uses TwoPoint: X0=0, X100=1, Y0=0, Y100=1200 on decoded NVA per-unit, not 32767 raw counts.
  • #258: optional Signal Map → select a signal → More → SCADA reference range for selected signal.... Deliberately opt-in choices: Q Diri Sendiri signed raw −32768..32767 to engineering ±2078, Q Lab same raw to ±260, or corrected POAQ feedback 0–1200. Only matching CA105/IOA/TI measurement points; shows a confirmation and supports Ctrl+Z. The photographed Q data items were not mapped to specific FEEDER/KOPEL/TRAFO IOAs; no bulk Q assignment, no command encoder changes.
  • #259: audit fix. Sync 31 I/O reference now correctly labels POAQ 0–1200. When an unverified SCADA reference range is first staged, the editor requires Save As to a new JSON file, prevents saving to the original station profile, and permits subsequent edits in the new copy. Adds field-isolation/rollback regression tests, including preservation of POOP setpoint.

Short test checklist

  1. For a non-destructive trial use the portable ZIP and back up the user-owned mapping JSON. Load the 31-point reference profile; confirm POAQ 790449 (TI9) feedback span 0–1200 and POOP 70537 (TI48) metadata 0–1000.
  2. Select one Q TI11, choose a Q reference under More, verify X0/X100/Y0/Y100, then test Ctrl+Z. Do not accept Q Diri Sendiri versus Q Lab for a particular IOA until eTerra and RTU evidence match.
  3. Apply one reference, press Save or Save & Apply, and verify the app requests a new JSON filename instead of overwriting the original. Attempting the original file path must be refused.
  4. Open Event Log during a source-qualified IEC-101 unbalanced poll capture: matched FC10 and FC11 responses show their real request FC; missing TX, mismatched source, and IEC-104 show —. Verify tooltips and Event CSV export.
  5. Confirm search/filter, direct cell edit, Undo/Redo, Mapping/Scaling tabs and shared style have not regressed.

Engineering limits

The photographed SCADA note suggests S = √3 × 500 kV × 2000 A ≈ 1732.05 MVA; 120% ≈ 2078.46 MVA. The Q Diri Sendiri ±2078 and Q Lab ±260 endpoints remain reference ranges, not confirmed per-IOA calibration. A full-scale TI11 signed mapping has a small nonzero output at raw=0 due to asymmetric −32768..32767 limits; do not silently force zero centering. TI32 tap positions and TI13 float units are not inferred from this photo. SCADA endpoint equality requires live RX bytes, source data-item association, known configuration revision and independent FAT proof.

No physical RTU/SCADA, 1200-bps SOE, 125/150% DPI, or command execution FAT is claimed from CI. IEC-101 P0 work #145 remains unmerged and excluded. This RC must not be advertised as production IEC conformance or station calibration approval.

Distribution and provenance

Publish only after the exact-source Package workflow succeeds and verifies portable executable GUI, installer, checksums, SPDX SBOM, and attestations. Official assets:

  • ARIEC60870-v3.6.17-rc.2-setup-win-x64.exe
  • ARIEC60870-v3.6.17-rc.2-win-x64.zip
  • SHA256SUMS.txt and the published SBOM/provenance

The v3.6.17-rc.2 tag is immutable. Never reuse RC.1 or overwrite its released bytes.

ARIEC60870 v3.6.17-rc.1

Pre-release

Choose a tag to compare

@github-actions github-actions released this 18 Nov 04:49

ArSee v3.6.17-rc.1 — Signal Map dropdown and toolbar acceptance

Release candidate · Windows x64 · 18 November 2026

This pre-release is for UI evaluation of the latest Signal Map engineering workbench. It is not a stable release, a protocol conformance certificate, or physical RTU/SCADA FAT evidence. The latest published stable fallback is v3.6.16.

Scope since stable v3.6.16

  • PR #255 integrates the More circular icon-only action at 36×36 DIP, optically aligned with adjacent Add, Save, and Save & Apply 36-DIP pill controls. It inherits the same Main Window circular appearance and input-state template.
  • Technical Signal Map fields now expose a compact native WPF dropdown of common IEC values with readable short labels and a secondary explanation for each option: IEC Type ID (TI), IEC Type/role, Class, COT, Command policy, TYPE/category, Engineering Unit, and Scaling Mode.
  • Each selection saves the exact existing field code, not a UI description: e.g. TI 31 saves 31; IEC role M_DP_TB_1 and policy MonitorOnly remain independent values. Non-catalogued project-specific codes remain editable. The UI never auto-pairs TI, role, unit, scaling, command policy or calibration from another choice.
  • Retains the stable grid-first workbench: native spreadsheet editing, filtering, Save/Apply validation, resizable inspector, paired two-point scaling, and current user-owned JSON profile schema. No IEC-101/103/104 parser, TX state machine, GI policy, physical serial wiring, command permission or source evidence changes.

Recommended desktop test (Signal Map)

  1. Back up the current engineering profile JSON first. For an isolated side-by-side evaluation, use the portable ZIP rather than upgrading a production installation.
  2. Open Signal Map and check that More remains circular and is exactly as tall as adjacent action pills.
  3. Choose digital TI 30 / TI 31 and analog TI 9 / TI 11 / TI 13 from the Mapping inspector; verify option descriptions and that the selected cell/JSON retain numeric codes.
  4. Check role, Class, COT, policy and category dropdowns; confirm no adjacent fields are silently rewritten. Choose an unfamiliar project-specific token, then verify edit/reopen fidelity.
  5. Confirm Scaling tab still exposes Factor and TwoPoint workflows with X0/X100/Y0/Y100; avoid interpreting IEC TI11 signed counts as MVar without authoritative field calibration.
  6. Confirm Save, Save & Apply, Copy/Paste, Undo/Redo, profile loading and the 31-point default profile have not regressed. If testing with an RTU, perform only site-authorized, isolated operations and retain source HEX.

Qualification and release invariants

The merged PR #255 passed Release build, formatting, protocol smoke, xUnit, native WPF acceptance, CodeQL, dependency review and repository hygiene; this RC's own Package workflow independently builds and checks its release source branch commit before publishing. Windows physical DPI 125/150%, genuine relay command transactions, SICAM serial handshakes, SOE/FIFO integrity, and verified field scaling remain NOT TESTED by these GUI gates. Unmerged IEC-101 P0 work in PR #145 is explicitly excluded.

Use the GitHub Release assets when and only when the Package run passes: ARIEC60870-v3.6.17-rc.1-setup-win-x64.exe, ARIEC60870-v3.6.17-rc.1-win-x64.zip, SHA256SUMS.txt, and SPDX SBOM. The immutable v3.6.17-rc.1 tag must refer to the accepted candidate source; do not replace the tag or republish different bytes under it.

ARIEC60870 v3.6.16

Choose a tag to compare

@github-actions github-actions released this 18 Nov 03:31
9892280

ArSee v3.6.16 — Signal Map workbench and Windows engineering test release

Windows x64 · Source-qualified stable release preparation · 18 November 2026.

Publication is not established by this notes file. Official Windows Setup/portable ZIP, SHA-256, SBOM, provenance and an immutable v3.6.16 tag must first pass the exact-source Package workflow. Until then, v3.6.15 remains the previously verified public stable version.

Signal Map: grid-first engineering workflow

Changes since v3.6.15 on the current master source:

  • #248: find by exact CA/IOA, signal name, bay, TYPE or mnemonic; combine bay/kind filters; an eight-column daily grid with optional advanced columns; explicit BAY TYPE + TYPE + MNEMO (+ existing number) name composition. Exact numeric IOA search prefers the primary point over command feedback references.
  • #250: Excel-style direct cell edit, copy/paste TSV, multi-cell selection, Fill Down and Undo/Redo. Pasted batches are validated against a staged profile and rejected atomically for malformed, out-of-range or invalid IEC data. Esc restores native WPF cell edits.
  • #251: grid-first layout. A compact header/search/action row gives the DataGrid maximum working area; the resizable Details inspector opens on demand with Ctrl+I instead of permanently narrowing the grid.
  • #252: toolbar Material vector icons and exactly the same action-pill, primary-pill, icon-only and search-shell WPF style resources as Main Window.
  • #253: Scaling/Mapping use the same shared toggle-pill styling and Material icons, with consistent field spacing, centered content and compact inspector controls.

For analog TI9/10/11/12/13/14/21/34/35/36, the Scaling page continues to expose Unit and either Factor or TwoPoint; X0/X100/Y0/Y100 remain above the fold in tested CI viewport. Digital/command points open Mapping. User-owned JSON profiles, command policy, IEC source identity, raw frame/quality evidence and Save & Apply semantics remain unchanged.

Field evidence worksheet (included in source)

Signal Map → More → Export blank Field FAT worksheet... creates an Excel-compatible UTF-8 CSV from the validated current Signal List. It records mapped identity and active scaling parameters while leaving independent field observations, calibration authority and evidence references blank with result NOT TESTED. Export is not a calibration approval or a command action. Reference field FAT matrix.

Evidence and safety boundaries

The merged GUI PRs passed Windows Release compilation, formatting, protocol smoke, xUnit, native WPF screen/interaction tests, CodeQL and repository hygiene on their reviewed heads. This release additionally requires exact merged master Windows CI, self-contained executable GUI probe, independent installed-updater FAT, Package asset verification, SHA-256 checksums, SBOM and provenance. Source-PR green checks alone do not satisfy the release workflow.

No physical IEC-101/104 RTU or SCADA FAT is claimed. Signed TI11 Q counts must not be displayed as MVar without independently approved RTU scaling (see #243); a 1200-bps SOE stress test requires physical FIFO/wire evidence. Open IEC-101 reliability work in #145 is not merged and not included in this release. Never perform active command/GI, serial-control or clock changes on a production link without site authorization.

Distribution

When the Package workflow succeeds, use its official v3.6.16 GitHub Release assets:

  • ARIEC60870-v3.6.16-setup-win-x64.exe — installer for managed Windows use.
  • ARIEC60870-v3.6.16-win-x64.zip — self-contained portable Windows application.
  • SHA256SUMS.txt, SPDX SBOM and published build provenance.

Do not reuse an older v3.6.16 tag for different source bytes or advertise those download paths before publication succeeds.

ARIEC60870 v3.6.16-rc.1

Pre-release

Choose a tag to compare

@github-actions github-actions released this 17 Nov 12:01
81b922c

ArSee v3.6.16-rc.1 — IEC-101/104 physical FAT candidate

Classification: public pre-release for authorized, isolated field-equivalent testing, not a stable release, IEC conformance certificate, or proof of correct RTU/gateway engineering calibration.

Source branch: release/rc-3.6.16-rc.1 (immutable candidate tag will point to its exact source commit). Published stable fallback: v3.6.15.

Candidate features

  • Includes merged Signal Map workbench #248: visible Ctrl+F search, bay/kind filters, focused eight-column grid, advanced columns in More, easy BAY TYPE + TYPE + MNEMO (+ #) generation and Scaling-first analog inspector.
  • Adds More → Export blank Field FAT worksheet... to the same editor. This is a read-only CSV derived from the current validated 31-point or custom user-owned Signal List, even when the user has unsaved edits. Excel-compatible UTF-8 BOM and quoted cells; textual formula injections are prefixed; source-specific connection passwords/hosts are not exported.
  • The worksheet contains effective CA/IOA/TI, signal name, bay/business TYPE/mnemonic, the active Factor or TwoPoint settings, and blank fields for independent observation time, decoded raw value, engineering display, independent RTU/SCADA reference, COT, QDS, source, original HEX/frame reference, calibration authority and field engineer notes. Every result is NOT TESTED until a human records physical proof. Dormant settings are intentionally not shown as active calibration.
  • Unverified TI11 reactive power counts remain counts, even when the bundled I/O source documents CT 2000/1 A and PT 500/0.1 kV. No guessed MVar, no implicit IEC unit conversion, no command permission changes, no native engine/FCB/GI timing modifications.
  • Includes a complete RC field execution and evidence matrix for the SICAM A8000 and other independently documented IEC-101/104 RTU profiles.

Software gates required for release

  • Exact-head Windows build and all regression suites, 19 protocol smoke vectors, formatter and repository hygiene.
  • Real WPF packaged-executable GUI smoke (including Scaling above-fold and Signal Map workbench checks), 1400-frame OBS-C acceptance.
  • Native Windows installed-update FAT from previous published Setup plus CodeQL/dependency checks on PR/source; Package creates checksum, SPDX SBOM and provenance.
  • A published prerelease, not a draft, with installer ZIP, portable ZIP, checksum and SBOM, with tag SHA identical to the accepted release-branch commit. Never overwrite an existing tag or change released binary bytes.

Physical testing — still outstanding

This release does not establish any of the following by software testing alone:

  • Physical COM/RS232/RS485 wiring, modem control RTS/CTS, SICAM firmware behavior, actual LA/COT/CA/IOA interoperation.
  • The true signed TI11 Q counts-to-MVar calibration for FEEDER IOA 790447, KOPEL 790439 and TRAFO 790443 (issue #243).
  • Hardware SOE/FIFO integrity during 60/500/2000 changes at 1200 bps; a test batch may physically exceed available throughput.
  • GI expected point completeness, command/SBO actuation, dual-circuit failover and sustained production-grade interoperability.

Use the field matrix to record observed stages and attach sanitized original HEX/verified capture. Do not operate breakers, modify clock, send setpoints, reset a production link or disconnect a live SCADA master to execute a software checklist. The on-site engineer must separately authorize and isolate any active Master tests.

ARIEC60870 v3.6.15

Choose a tag to compare

@github-actions github-actions released this 17 Nov 09:29
78356d0

ArSee v3.6.15 — Visible IEC-101/104 Scaling in Signal List

Release candidate notes. Not published until the immutable exact-source Package workflow succeeds. The current published stable release is v3.6.14. The new v3.6.15 assets/URLs must not be advertised until verified.

What changed

The Signal List Editor previously placed Engineering scaling mode after CA, IOA, Type, Name, Group, Type/Role and Unit within one long scrolling Selected signal panel. On a normal 1360×820 Windows window, a selected TI9 TRAFO SOGI TM POAQ could show only Bay / Description near the end of the panel while all scaling controls were off-screen, despite the scaling engine being correct. This was unusable for routine FAT/SAT calibration.

PR #245 changes the actual WPF editor:

  • Engineering Scaling and Mapping & Commands are separate, compact tabs in the fixed right inspector, not one endless form. A measured analog TI9/10/11/12/13/14/21/34/35/36 opens Scaling automatically; a digital or command row opens Mapping/Commands.
  • Scaling offers Unit, mutually exclusive Factor / TwoPoint mode, and X0, X100, Y0, Y100 in a compact two-column layout. The full four-input TwoPoint range is directly visible in the default 1360×820 WPF viewport without dragging the scrollbar.
  • Factor mode retains Scale/Offset/Ratio primary/secondary and normalized full-scale, independent from the TwoPoint formula. Changing mode never compounds the two calibration mechanisms.
  • CA, IOA, TI and Name remain frozen at the left side of the large signal DataGrid while the user scrolls horizontally. Mapping/Commands retains BAY TYPE, business TYPE, MNEMO, state map, class/COT, feedback and command policy.
  • Type-aware guidance reminds engineers that TI9 normalized X is decoded per-unit, TI11 scaled X is signed I16 counts (not automatically MVar), and TI13 IEEE short float is already in the upstream unit unless project mapping specifies otherwise.

Why this does not change field scaling accuracy

The prior v3.6.14 Factor and TwoPoint implementations remain unchanged. The document-backed 31-point I/O mapping still uses CA=105 and exact IOA/TI identity. The PDF does not provide an approved gateway signed integer-to-MVar factor for FEEDER, KOPEL or TRAFO reactive power TI11 points; these remain raw counts until field engineering data is supplied. Track physical bipolar Q calibration in #243.

The editor continues to preserve existing profile JSON, user-owned calibration, original HEX and IEC quality, timestamps, GI/SBO/control identity and all IEC-101/103/104 engine behavior. The new UI does not silently overwrite old profiles; Sync 31 I/O still requires explicit review and Save As.

Real-WPF acceptance

The exact self-contained Windows executable runs a UI acceptance probe as part of the existing --passive-gui-smoke flow. It opens the actual Signal List Editor at 1360×820, selects TI9 POAQ, and asserts each X0/X100/Y0/Y100 editable field falls within the visible scaling viewport, not merely within the full scrollable content. It also checks TI11 warnings, a digital row's Mapping default, frozen identity columns, and unchanged persisted row data. A first PR candidate failed the above-fold test and was redesigned; the corrected version passed PR software CI.

Publication acceptance remains a separate gate: native installed-update FAT on the exact merged commit, Windows/WPF tests, CodeQL, release executable probe, verified Setup/ZIP, checksums, SPDX SBOM, provenance, immutable GitHub tag and Pages documentation update. This release is not an IEC conformance certificate or field SAT.

See engineering scaling guide, 31-point source analysis and Release Packaging.

ARIEC60870 v3.6.14

Choose a tag to compare

@github-actions github-actions released this 17 Nov 08:50
631d6ee

ArSee v3.6.14 — IEC-101/104 engineering scaling and source-grounded Signal Mapping

Release candidate notes — not a statement of publication. This document is bound to the source version 3.6.14 in Directory.Build.props. Public links, actual tag/commit, installer SHA-256 and provenance must be filled from a successfully completed immutable Package workflow. Until then, the latest published stable remains v3.6.13.

1. Engineer-configurable analog scaling

  • IEC-101/104 measured values may be normalized (TI9/10/21/34), signed scaled (TI11/12/35), or IEEE-754 short float (TI13/14/36). The IEC type identifies encoding, not an RTU-specific engineering-unit factor.
  • Signal List provides mutually exclusive Factor and TwoPoint projections. Factor uses engineering = source × normalized-full-scale(if NVA) × scale × primary/secondary + offset. TwoPoint uses Y = Y0 + (X−X0) × (Y100−Y0)/(X100−X0), where X is the decoded source value (NVA in per-unit, SVA signed integer, IEEE float as sent).
  • Both negative and positive ranges are supported. Invalid/nonfinite endpoints or zero X span are refused. TwoPoint does not additionally multiply by Factor, CT/PT or NVA full-scale. Off-range values are extrapolated rather than silently clamped; never interpret extrapolation as RTU certification.
  • Original capture bytes, IEC quality, CA/IOA, timestamps, controls and GI evidence are preserved. The engineering projection appears in Values, Event Log and its history/CSV and Frame Trace when the exact point identity and trusted measurement are available.
  • TI13 P is identity when the upstream device already sends MW. TI11 Q requires the authorized gateway signed counts → MVar conversion. A raw observation such as 13641 must not appear as 13641 MVar merely because a unit field says MVar.

2. SCADA I/O list — 31 exact point names

Based on the operator-provided I/O LIST PENGUJIAN SCADA (31 rows, page 1) previously reconciled in #240, the bundled example uses BAY TYPE + TYPE + MNEMO + LAB (when present), preserving CA=105, original IOAs, IEC TI, command policy and feedback:

  • FEEDER TM P — IOA 790446, TI13; FEEDER TM Q — IOA 790447, TI11.
  • KOPEL TM P 1 / KOPEL TM Q 1 — IOA 790438/790439, TI13/TI11.
  • TRAFO TM P 2 / TRAFO TM Q 2 — IOA 790442/790443, TI13/TI11.
  • TRAFO SOGI TM POAQ — IOA 790449, TI9, source-documented real power setting (0/1000); X0=0, X100=1 NVA, Y0=0, Y100=1000 engineering.
  • Existing digital indication, remote command, and command-feedback identities remain source-aligned, with no guessed station alias for an unrelated CA or Type ID.

CT 2000/1 A and PT 500/0.1 kV from the I/O list provide a theoretical apparent-power nominal magnitude (~1732.05 MVA) but do not establish the RTU TI11 signed-count scaling to MVar. The bundled default deliberately labels those three uncalibrated reactive-power readings counts. The separate provisional scaling example is not a default or FAT-approved calibration.

3. Upgrade existing Signal Lists safely

Already-saved local user Signal Lists are never replaced by installation. For an older project mapping, open Signal List → Sync 31 I/O, check the preview of exact CA+IOA+TI matches, and confirm only if those points belong to the intended project. Save & Apply to a new JSON filename. This updates names and business metadata, replaces an uncalibrated legacy TI11 Q label MVar with counts, and applies the documented TI9 POAQ mapping only to legacy identity/unconfigured rows.

The action keeps custom Factor/TwoPoint calibration, command policies, feedback IOAs, station identity, communications, and other stations unchanged. The editor also now retains business type and optional engineering min/max metadata on save/duplicate. No user profile is silently migrated at startup.

To display actual MVar, have the approved RTU/gateway factor or two independent endpoint pairs and then enter them in the selected Q point. The software must not infer a usable calibration from one observed 1730 MVar reference.

4. Software release acceptance and boundaries

The source-milestone PRs #239, #240 and #241 have been merged. Their software regression work and the prior merge commit's exact-source CI/CodeQL/hygiene were green before preparing this version. These results are not qualification for a future v3.6.14 release binary.

A public stable release requires all gates from Release Packaging: exact-source Windows build, protocol and all xUnit suites, real-WPF acceptance of the packaged executable, independent genuine previous-release installed-updater FAT, immutable package verification, SHA-256 sums, SPDX SBOM and provenance. Physical RTU/SCADA signal scaling, serial 1200-bps 2,000-event SOE parity, and multi-vendor controls/GI tests remain separate field acceptance. The binaries are currently unsigned; nothing here claims IEC conformance certification.

Sources: engineering scaling guide, document-backed mapping analysis, and the immutable v3.6.13 stable release.

ARIEC60870 v3.6.13

Choose a tag to compare

@github-actions github-actions released this 17 Nov 02:16
289caef

ArSee v3.6.13 — Reliable IEC-101 Master and verifiable IEC-104 GI (Windows x64)

This document describes the v3.6.13 release candidate source contents, prepared 16 November 2026. It does not assert that v3.6.13 is already published or that any later binary passes acceptance. The latest published stable remains v3.6.12 until a successful tagged release completes.

IEC-101 single-link and redundant A/B Master

The reviewed standalone hardening PRs #219–#225 now use one bounded transaction and FCB/FCV authority per physical IEC-101 COM link, correlate observed secondary responses with the active request, handle delayed ambiguous FC11 responses, preserve ACD Class1/Class2 fairness, and recover real serial-port faults with independent A/B ownership and controlled backoff. No application-level mutating control is automatically replayed after a transport ambiguity.

The next stages #226, #227 and #228 harden SBO selection/execute source and command epochs, feedback-target identity, and the IEC-104 TCP connection association. GI is counted at application level only for a request-correlated positive ACTCON and matching ACTTERM on the same physical owner/association. A GI ACTTERM does not prove all configured IOAs were received. A command ACTCON or ACTTERM does not prove physical breaker operation.

The broad historic draft #145 remains unmerged; its entire branch was not rolled into this release.

Mapping, SOE and bounded memory

#229 makes read-only names consistent across online Values and re-opened portable .ariec captures, with a source-scoped CA-ambiguity check before allowing unique process IOA display aliases. The bundled mapping for FEEDER CB STATUS = CA105 / IOA16712689 / TI31 is available when the operator loads an appropriate Signal List. If two stations or other IOA identities conflict, the original CA/IOA/TI remains visible and the name is not guessed. This never changes command addresses, GI expected-point proof or original wire bytes.

Accepted IEC-101 spontaneous digital objects have their own cumulative counter. 60/500/2,000-event deterministic synthetic tests validate software accounting and bounded report retention, not physical RTU-generated events. Both single/A/B IEC-101 and IEC-104 now use the shared O(1) bounded evidence FIFO, with exact in-memory eviction counters and unchanged live event delivery.

#230 additionally prevents IEC-104 falsely terminating the GI receive burst or incrementing its GI counter from wrong-source/CA/OA/QOI, negative/TEST or pre-ACTCON responses. The lifecycle is evaluated only for accepted IEC-104 I-frames after an actual successful GI TX on the current connection, using the same tracker implementation as IEC-101.

Software acceptance and release delivery

Before publication, confirm the actual v3.6.13 tag/commit passes Windows Release (warnings-as-errors), all six xUnit suites and 19 protocol smoke tests, native real-WPF OBS-C 1,400-frame acceptance, CodeQL, repository hygiene, release-package executable check, official Windows installer/portable ZIP structure, SHA-256 checksums, SPDX SBOM and build provenance. A green PR is not a green release artifact.

The stable v3.6.13 Package workflow will refuse publication without a successful exact-commit native Windows independent updater FAT including genuine installed-old-release update-worker acceptance. This safeguard must not be bypassed. The real installed updater requires operator approval; the portable ZIP does not update itself. Windows builds remain unsigned.

FIELD FAT PENDING: No claim of complete 60/500/2,000 IEC-101 SOE acquisition at 1200 bps from an actual RTU/gateway until a separately implemented reference master confirms event-count parity on the same cable/adapter, using original RX/TX frames and independently generated RTU event totals; device FIFO overflow, physical USB COM recovery, A/B failover, real command feedback, multiple IEC-103 devices and multi-vendor IEC-104 TCP peers remain independently unqualified. The application is an engineering tester/analyzer, not a certified production SCADA master or IEC conformance test suite.

The website/README links deliberately continue to point at published v3.6.12 until v3.6.13 is actually published, then the public landing, download guide and checksums must be synchronized to the immutable new release assets.

ARIEC60870 v3.6.12

Choose a tag to compare

@github-actions github-actions released this 16 Nov 04:15
01ac138

ArSee v3.6.12 — Consistent IEC-101/104 mapped signal names (Windows x64 stable)

Engineering usability: one signal identity across the application

An IEC-101/104 process point from the active IOA profile now keeps the same engineering signal name in the live Events, Values, Frame Trace and Frame Inspector, as well as paged Event History, history search, offline capture review and CSV exports. This includes the bundled sample profile's CA 105 / IOA 16712689 / TI 31: FEEDER CB STATUS. Previously this point could appear as the raw fallback IOA 16712689 despite an available configured name.

Value rows re-project when a profile is changed, and history pages re-apply the profile's current names without editing the original source evidence. A saved imported profile now restores from its actual local file path; a sanitized path remains only a presentation value.

Read-only name recovery tolerates a single-station process signal with a uniquely matching IOA and compatible signal kind where a received CA or timed Type ID disagrees with the profile. This never rewrites physical CA/IOA/Type, never increases GI proof coverage, and is deliberately rejected on conflicting source identities, multi-station ambiguity and command identity. Unmapped points such as bundled-profile-absent IOA 388725 remain explicitly unmapped; they are not guessed to be other signals.

Command events can show the safely mapped descriptive name while preserving original command target and evidence. Command feedback lookup remains strict by CA/IOA and monitoring type. The release incorporates the reviewed fixes and regression cases from PR #210.

Preserved IEC-101 work from the concurrently developed previous release

This build also retains all public v3.6.11 IEC-101 service-correlated reply validation from PR #208, the earlier v3.6.10 forensic Event Journal, original TX/RX frame inspection, bounded 1200-bps FT1.2 receive, 31-point bundled profile, installer updater safety and the existing IEC-103/104 workspaces. The bilingual public website/download documentation contributed by PR #211 remains part of the mainline history.

Validation and physical field limitations

The Windows x64 installer and portable ZIP require a successful exact-commit GitHub Actions package build, six xUnit suites, protocol smoke tests, real-WPF executable/Frame Trace acceptance, SHA-256 checksums, SPDX SBOM and build provenance, plus independently successful Windows installed-updater FAT before stable publication. Verify the specific successful workflow runs; merging the code or passing PR-only CI does not qualify public installer bytes.

FIELD FAT PENDING: Actual SICAM A8000 / SCADA site verification of the corrected CA 105 / IOA 16712689 display, cross-device IEC-101 interoperability at 1200 bps, 60/500/2,000-event SOE FIFO completeness, physical COM reconnect, dual-link failover and breaker GI/SBO command lifecycle. No physical IEC conformance certification is claimed. The draft #145 shared FCB/retry/core rearchitecture remains unmerged.

Windows executables are currently unsigned. Download only the official GitHub Release assets and verify SHA256SUMS.txt. The installed updater needs operator approval and does not force-stop active IEC capture; portable ZIP does not self-update.