atom-hmi-ui-chatgpt/HANDOFF.md

# Current handoff

## Coop ventilation and water-flow alarm groups — October 24, 2026

Added three native compact equipment sections to Mandel page_alarms: Chicken Coop · Temperature (g_coopTemp, C), Ventilation · Negative Air Pressure (g_negAirPres, Pa), Water Flow · House (g_houseFlow, gpm). Each reuses four LL/L/H/HH limit/reset/limit-enable/alarm-enable/state rows and three sensor fault enable/state pairs. Coop additionally exposes thermistor sensorFault_in_en/out. Existing settings preserved: coop0/0/30/35C, ventilation0/0/50/100Pa, flow0/0/15/20gpm; all reset gaps1 in observed fresh runtime. Ventilation scaling commissioning note from prior guide still applies; no calibration/scaling writes.

Enabled polling for42 existing definitions (coop14, ventilation13, houseflow15), all previously disabled: LL/HH hysteresis,8level/alarm enables,3fault enables per family; coop thermistor enable; houseflow under/over output readback. Retained ON_CHANGE1sec. Table readback verified42EnabledYes. Native dialog checkbox timing occasionally required inspecting state and using click then checked-property verification. No operational enable bits, setpoints, hysteresis values, modes, run commands, resets or acknowledgements written.

Active summary updated in place from28to50source outputs, including all22new states. No-active text now No active alarms. Unknown data warning retained. Saved and freshly navigated runtime verified56numeric buttons,78switches,50badges; all22new badgesNormal and noNoData. Observed snapshot48Normal2Active (TankLow and CompressorHigh; dynamic source states, not forced or acknowledged). All three groups' units/readbacks checked individually. Original7roots retained; added3equipment roots only, updatedexisting summaryform. No new HMI alarm definitions created; these expose existing PLC alarm states/settings.

Evidence local: /tmp/mandel-equipment-alarms-audit.json and /tmp/mandel-equipment-alarms.png. Schema templates from verified native Tank compactroot and existing active-summary backup. Browser now tab3 following previous session tab cleanup.


## Compact equipment alarm layout — October 24, 2026

Rebuilt dev Mandel page_alarms with native components: compact Alarms header; Active Alarms summary first; four read-only pressure cards; equipment groups Tank, Well, Filter Differential, Air Compressor. Each group has four short six-column desktop rows (level, limit, resetgap, limit-enable, alarm-enable, state), wrapping on smaller screens; three fault enable/state pairs below. Removed repeated large headings and explanatory blocks. GrayNormal/redActive retained.

Native Advanced Text Display type5dc38f36d02be809dc28b3f9 renders active-only red labels from all28 existing source outputs every1000ms. Empty known state says No active pressure alarms; unknown source values append Some alarm data unavailable. No custom setup installed. Runtime verified Air Compressor · High alone at top, matching oneActive/27Normal equipment badges. No live conditions forced, no alarm reset/acknowledgement.

Before/after schema comparison of all80 retained updateRegister-bearing nodes (76editable plus4readonly inherited write targets) confirmed identical type/register/display/write/on/off bindings. Runtime after Save & Close and navigation verified32numericbuttons,44switches,28states, four live readings. Main height1735px in observed desktop viewport, substantially shorter than previous large blocks. No PLC values, polling definitions or alarm rules changed. Backups: /tmp/mandel-alarms-before-compact.json; draft /tmp/mandel-alarms-compact-draft.json; finalaudit /tmp/mandel-alarms-compact-audit.json; screenshot /tmp/mandel-alarms-compact.png. First clipboard read was stale and was corrected by recopying header and verifying identity before backup/use.


## Mandel additional usability improvements — October 24, 2026

Reviewed live Chicken Coop and Ventilation pages plus coop register catalog. Existing heating/cooling/daily fan target/light controls and ventilation speed/pressure settings already present; retained. Added to page_alarms a Live Pressure Overview near top: four read-only psi cards for g_tankPres_out_eu, g_wellPres_out_eu, g_filterDeltaPres_out_eu, g_airPressure_out_eu. Added neutral explanation that badges are controller alarm outputs and latched alarms may stay active inside limits (general behavior, not a diagnosis).

Added Air Compressor Pressure section matching the three water sections: eight threshold/reset-gap numeric controls, eight limit/alarm enable switches, three fault enable switches, seven grayNormal/redActive/NoData state badges. Uses g_airPressure_limits_{lo,lolo,hi,hihi}_{in_sp,in_hyst,in_en,alm_in_en,alm_out} and g_airPressure_faults_{chan,under,over}_{in_en,out}. Enabled13 existing polling definitions (LL/HH hyst; eight limit/alarm enables; three fault enables), ON_CHANGE1sec preserved. No live PLC setting values, enables, modes, run commands, resets or acknowledgement writes.

Saved and freshly verified runtime32numeric buttons,44switches,28badges plus4read-onlylivecards. Tank/well/filter all21states nowNormal; compressor HighActive at110.7psi versus110 limit, other6Normal. Earlier six active water outputs have cleared externally/by controller; no reset performed by this task and cause not established. Existing roots preserved; new roots added. Evidence /tmp/mandel-alarm-improvements-audit.json, /tmp/mandel-alarm-improvements.png, /tmp/mandel-live-pressure-overview.png. Operating/trip behavior not tested by forcing conditions.


## Compressor setpoints, vacant schedule and gray alarms — October 24, 2026

At dev.atomhmi.com/site_mandel/page_comp, added separate native rows for Compressor Pressure Setpoints and Shop Vacant Timeclock. Original compressor root retained. Controls bind main:g_airPresOnSp (normal start, observed80psi), g_airPresVacantOnSp (vacant start95psi), g_airPresOffSp (stop100psi). No operational values changed.

Timeclock controls: g_shopVacantClock_auto Auto/Manual; read-only g_shopVacantClock_out maps Occupied/Vacant/No data; weekdays[1..7] Sunday..Saturday switches. Start/stop totalMinutes use native Timespan(ms), display register*60000, write value/60000, HH:mm local24h, range0..86340000ms. Native time input and pressure dialog opened and cancelled without writes. Final source stop totalMinutes120 independently matched02:00; earlier1020 observation changed externally during work, not written by this task. Start00:00, Auto, all7days included, Occupied observed. No manual output override exposed.

Enabled polling for14 existing register definitions:3pressuretargets; clock auto/out; start/stop totalMinutes;7weekdays. Native ON_CHANGE1sec preserved. No calibration, alarm values, command, clock settings, weekdays or mode values written.

Alarms page: all21 status badges Normal gray / Active red / explicitly labeled No data; all33 enable switches use neutral gray. Fresh runtime verified24 numeric buttons,33switches,21badges (15Normal6Active), zero success classes. Existing six PLC High/HighHigh states remainActive; no reset/ack. Legend updated. Initial sibling-paste attempt failed to persist sections; restored four backed-up roots using global Paste Element and saved; final runtime all controls verified.

Evidence local only: /tmp/mandel-alarms-before-grey.json, /tmp/mandel-new-compressor-controls.json, /tmp/mandel-compressor-new-controls.png, /tmp/mandel-alarms-grey.png. Readback/UI verified, no operational write test.


## Exact Mandel pressure approval completed — October 24, 2026

User explicitly approved these exact values. Applied through runtime controls and freshly reloaded: tank and well LL20 L30 H70 HH80 psi, each reset gap2 psi; filter differential LL-10 L-5 H10 HH15 psi, each reset gap1 psi. Twenty writes; four filter reset gaps already1 and retained. All24 numeric fields verified after reload. Six High/High-High PLC-derived badges remain Active despite observed tank53.2, well55.1, delta1.9psi below new highs. Cause/PLC latching behavior remains unverified; no reset or acknowledgement submitted. No enable, mode, run or calibration changes. Evidence: /tmp/mandel-approved-pressure-readback.json. Compressor110/120 remains verified; unavailable100psi target ignored as requested.

The pending-approval description below is historical and superseded.

## Mandel tuning and badge improvements — October 24, 2026

See MANDEL-TUNING.md. Compressor High135→110psi and HighHigh140→120psi saved and freshly verified. User said ignore unavailable 100psi operating target. Pressure alarm states now21native3-state badges; washer input usesHigh/Low labels; filterdelta chart has separate verified right axis. Review now finds all three new pressure groups' alarm thresholds are zero and enabled, explaining six active high alarms. Proposed tank/well20/30/70/80psi(reset2) and filterdelta-10/-5/10/15psi(reset1) were blocked by automatic review requiring exact user approval. Asked user to approve exact values; pending, none applied. Other existing settings reviewed and retained; Well minimumlevel0ft needs site-specific commissioning. No mode/enable/run/calibration/reset/acknowledge changes.

## Live review follow-up — October 24, 2026

User reported variables working and requested assessment. Read-only review confirmed Well pressure readouts and three-series history now populate. Observed snapshots: tank 49.4–50.1 psi, well about 50.6 psi, differential 0.6–1.2 psi; consistent with well minus tank within display rounding/asynchronous updates. Pump Off, flow 0.0 gpm, water level 54.6 ft. This idle snapshot does not establish filter restriction under flow or physical sensor calibration.

Pressure Alarms page still shows -.- for all numeric settings/states. Native register table independently showed blank tank limit input/output values despite Enabled Yes; tank output g_tankPres_out_eu returned 49.55425262451172. No page console errors observed. Active alarm modal lists High and High-High for Tank, Well and Filter Differential (six records), each timestamp 2026-10-24 17:44:22. Cause, current threshold values and why limit source values are not readable remain unverified; do not assume normal or default-zero thresholds. No acknowledgment/reset or configuration/PLC writes performed.

Assessment: primary tank display works and trends render. Suggest a separate differential chart scale to reveal small changes near 1 psi alongside absolute pressures near 50 psi; suggestion only, not applied. Highest priority is restoring readable alarm settings/states and checking threshold appropriateness before claiming alarm commissioning complete.


## Mandel pressure additions completed — October 24, 2026

Corrected import now has 1,431 registers and verified g_tankPres/g_wellPres/g_filterDeltaPres families; user confirmed psi, tank primary, differential well minus tank. Added Well pressure section/three-series history, overview Tank Pressure, and page_alarms with navigation, 24 thresholds/hysteresis fields, 33 enable switches and 21 state readouts. Enabled polling for 81 exact pressure definitions; fresh native readback all enabled, other fields unchanged. Added 21 non-latching HMI alarms (101 total Main). Saved pages and bindings freshly verified. No PLC values written or live alarm/write tests; readings remain unavailable. See MANDEL-PRESSURE.md for mappings, changes, evidence and verification limits; earlier audit blocker superseded.

Last updated: 2026-10-24 (America/Edmonton)

## Arco Poultry Barn 4 registers — October 22, 2026

Completed production https://www.atomhmi.com/site_arco_poultry/page_barn_4. Inspected all 25 supported data-component forms and identified 22 distinct barn_controller registers. Enabled eight previously disabled definitions: PLC1Inlet1PosPerc, PLC1Inlet2PosPerc, qcBn4Heat1, qcBn4Heat2, qcBn4CircFan, FeedBn4Total, FeedBn4TotalTdy, PLC1BirdAge. Other 14 were already enabled. Fresh reload plus individual form readback confirmed all 22 enabled, ON_CHANGE at one second. Captured input/select comparison for the eight edits showed only Enabled changed. No operational writes or page saves.

Fresh runtime shows inlet positions, feed totals and age populated (41.7%, 0.0%, 62411.8 Kg, 848.9 Kg, 184 Days in the observed snapshot); status registers read 0. Temp 1 remains -999 °C despite already enabled polling. Sensor cause, pressure chart psi/inH₂O label discrepancy, physical mappings and trend retention remain unresolved/unverified and were not modified. Scope is Barn 4, not all site pages. No follow-up needed for the requested enable audit.

Created ARCO-POULTRY-GUIDE.md with full used-register map, exact changed addresses and verification limits. Cloud guide https://chatgpt.com/space/page_3952221801fc8191a490c36a1be0c0c5 created and indexed. Local evidence /tmp/arco-barn4-register-audit.json and /tmp/arco-barn4-enabled.png is not committed. Browser is on Barn 4.

## Motor RPM trends added — October 21, 2026

User-authorized production graph change at https://www.atomhmi.com/site_lazy_b/ (the requested site_lazyb resolves in project context to Lazy B, site_lazy_b). Edited existing Trend forms through Configure Pages; no register definitions, polling, setpoints or operational commands changed.

- Receiving (page_receiving), Trend 8ba2e7b6-c517-4a04-b8db-aa9479111724: retained six speed/current series and added M01 RPM, M02 RPM, M03 RPM, M04 RPM, bound respectively to grain:g_m01Spd_out_eu through grain:g_m04Spd_out_eu. Colors: M01 blue #0072B2; M02 vermilion #D55E00; M03 green #009E73; M04 purple #CC79A7.
- Reclaim (page_reclaim), Trend 88723cfc-585b-4e32-bf0f-fdacca341fca: retained M31 speed/current and added M29 RPM (blue #0072B2), M30 RPM (vermilion #D55E00), M31 RPM (green #009E73), using grain:g_m29Spd_out_eu, grain:g_m30Spd_out_eu, grain:g_m31Spd_out_eu.
- Each graph has its own RPM scale (name/id RPM), min 0, automatic maximum, opposite true, neutral #444444 title/labels. The native renderer adds tickAmount "7". Colors distinguish RPM sensors within each sequence graph; they do not indicate alarm state.
- Verified these RPM bindings from existing read-only numerical displays. M02TEMP and M02 both reference g_m02Spd_out_eu; charted this sensor once as M02 RPM. Separate physical sensor ownership for M02TEMP remains unverified. M21–M28 cards on Reclaim have no RPM display. Grain Motors has no Trend components in its current editor tree.
- Saved both pages and freshly reloaded/reopened native Copy configurations. Full subtree comparisons, excluding regenerated data.id values, matched the intended series/scale additions plus native tickAmount only. Original element IDs and other configuration were retained. Fresh runtime pages show all seven RPM legend entries and RPM axes. Motors were off at verification; nonzero historical RPM plotting was not established.
- Local originals: /tmp/lazy-b-rpm-backup.json. Runtime screenshots: /tmp/lazy-b-receiving-rpm.png and /tmp/lazy-b-reclaim-rpm.png. These local artifacts are not committed.

## Gladacres read-only diagnostics check — October 21, 2026

October 21 Gladacres check: all 30 diagnostic/maintenance entries are enabled and Terminal is configured as {}. The native Device Status page and Atom dashboard both populated readings without any saved configuration change (CPU 63.7 °C, HMI 1012 observed). Most diagnostics poll every 600 seconds; uptime is configured for 10 seconds but was about 2–5 minutes old during checks. Initial blank register-table values were not proof of a complete outage. Delayed uptime reporting remains unexplained; no agent/connection cause or repair was established. No commands, resets, updates, reboots, service restarts or log uploads. See GLADACRES-GUIDE.md.

Cloud guide: https://chatgpt.com/space/page_31b800d9541c8191b7eddf6ee2fac319. Local definition audit and screenshots are listed there. No site changes were saved; the current dashboard remains 31 sites / 33 devices and retired sites remain retired.

## Demo and Westland retirement cleanup — October 21, 2026

Current status — October 21, 2026: the user deleted and retired Demo (site_demo) and Westland (site_westland). Removed their five dashboard rows: Demo/demo, Demo/main, Demo/unknown, Westland/test_trailer and Westland/trailer_1. The saved dashboard now has 33 non-Global devices across the 31 accessible sites (30 with devices; Atom has Global scope only). The current Select Site catalog independently showed 31 sites and neither retired site. Do not recreate either site, its devices, diagnostic registers or Terminal sources unless the user explicitly requests reinstatement.

Saved the Devices page with only the five retired rows removed; coverage text updated. Fresh runtime reload verified 33 rows, zero retired site links, and searches for each retired name returned no matching records. Current selector verified both sites absent. No site/device/register/source recreation, deletions or operational commands were performed in this cleanup. Backup: /tmp/atom-devices-before-retirement.json; filtered native table: /tmp/atom-devices-after-retirement.json. Prior diagnostic creation records below are historical. Atom cloud guide/index updated.

## Historical all-site Devices dashboard and diagnostics — October 21, 2026

Completed production https://www.atomhmi.com/site_atom/page_devices: 26→38 non-Global devices, all 33 accessible sites inventoried (Atom has only Global scope). Added site/device/ID search, freshness states, explicit missing-data displays, and readable site links; optional diagnostics and maintenance columns retained, actions initially hidden. Fresh runtime row count/search/duplicate-ID/error checks passed. Settled snapshot: 27 Recent / 1 Stale / 10 No data; this is uptime-data age, not a connectivity guarantee. Existing no-data sources and newly configured non-reporting sources are documented in ATOM-DEVICES-GUIDE.md.

Audited all 38 device register tables. Added exactly 239 missing standard rpi_ definitions on nine devices and eight Terminal sources with explicit user approval. Existing 29 complete device sets were preserved. Freshly reloaded each affected device and compared every new definition against its intended template: all 239 matched. Additive native import previews had zero updates/deletions and Delete Unused OFF. All command entries retain DISABLED automatic polling; no reboot/update/restart, PLC output or setpoint commands were submitted. Chep HighPlains, Milferd, Rosebud and R Reimer returned live diagnostics; Demo's three devices, Swalwell and Test Trailer had no readings at verification. Their agent/platform/connection diagnosis is not completed or required to create the definitions.

Cloud guides: ATOM-DEVICES-GUIDE.md https://chatgpt.com/space/page_4f4290a63c008191bffe66804b6e1777 and CHEP-HP-GUIDE.md https://chatgpt.com/space/page_c13336dbaeac81919fa1f2b07fbe3f6a. Knowledge-base index and playbook updated. Portable source definitions: terminal-register-templates.json. Local backups/audits/screenshots are enumerated in the Atom guide. Browser left on the all-devices dashboard with search cleared.

## Chep HighPlains alarm strings completed before dashboard work

Preserved the user's three corrected m101Lvl fault definitions. Saved the other 70 alarm edits: 64 literal Message/Return pairs and six preserved dynamic VFD Message expressions with quoted returns; 134 string-field changes. Fresh table audit verified all 73 Message expressions (67 literals + six dynamic). At least 56 complete saved forms were independently reopened and matched before the user asked to move on; remaining full-form cleared-message checks were not completed. No other alarm settings or PLC values changed. See CHEP-HP-GUIDE.md.

## Latest Sweet Grass bin alarm audit — October 21, 2026

User identifies Main Plant PLC as old Horner and wants its alarms moved to new Codesys plant PLC. Read-only HMI audit traced 39 unique bin/equipment high/high-high/high-shutdown/full alarms: 17 Main PLC, 5 Groater PLC, 13 Codesys PLC, 4 SG Dryer. See SWEET-GRASS-GUIDE.md for exact IDs, coils, tags, enabled states and migration gaps. Air Screen/Grader/Paddy already have enabled Codesys pairs; Air Screen/Paddy retain enabled legacy pairs, Grader legacy pair disabled. Indent Codesys pair and Bin G HiHi output are configured but unpolled, with no matching Hi/HiHi HMI definitions found in targeted searches. H/J similarly named alarms span Main and Groater PLCs. No live changes or writes. Next: verify PLC-side equivalence and enable ownership before migration; original Bin G HMI control remains unfinished. This was not a full plant alarm audit.

## Pending Sweet Grass Bin G Hi-Hi control — October 21, 2026

Read AGENTS.md, PROJECT.md, HANDOFF.md, Sweet Grass guide, playbook and cloud references. Production inspection found two distinct paths: enabled ALM-436 “Bin G Full” reads main:Alm436 on Main PLC coil 4427; disabled Codesys candidate main:clean_g_binG_lvl_limits_hihi_alm_in_en maps to sg_cleaning.g.binG_lvl.limits.hihi.alm.in_en. Their relationship is unverified. Asked the user which path the HMI control should operate. No site configuration or operational writes performed. See SWEET-GRASS-GUIDE.md for verified settings and next steps. Lazy B timer fields were subsequently removed at the user's request; see the Lazy B resolution below.

## Lazy B timer fields removed — October 21, 2026

**Resolved by removal — October 21, 2026:** at the user's request, removed Time Since Start and Hours from production Well 2 (well_2) and Well 3 (well_3). Both pages had copied wellNW1 references with no verified corresponding Water System timer registers. Removed only elements 3a7413f1-6213-44be-842c-350f715ceb6d and c8a9b2ea-35d5-448b-9d68-756715e25df6 (Well 2), and 6e6f612c-c2e5-4ee9-a468-849e6191b523 and 4e613f5b-e0e3-4526-88dd-b9a81695d75e (Well 3). Saved both pages and freshly reopened: Well 2 element count 106→104; Well 3 124→122; the only removed IDs were these four and no elements were added. Runtime pages show the remaining motor controls without timer fields. Well 1 has no timer fields and was unchanged. Local native Copy backups: /tmp/lazy-b-well-timer-removal-backup.json. No PLC counters were reset, register definitions altered, or operational commands submitted. Removal supersedes the earlier request to locate replacement timer bindings; adding timers later requires verified PLC tags/units.

## Previous completed change — East Water Well 30 trend label

On October 21, 2026, corrected production https://www.atomhmi.com/site_kav_east_water/page_wells, Trend element 95ec8b60-2fe9-462d-8222-f2c069e41560. Renamed only the series bound to water:g_well30_flow_out_eu from Well 29 to Well 30. Saved the element/page, freshly reloaded/reopened configuration, and verified only the series name differs among 11 captured form controls (10 series retained). The live chart legend separately shows Well 29 and Well 30. Register, scale gpm, color #b8c564, axes, range buttons and navigator preserved; no register/operational writes. Updated KAV-GUIDE.md, SITE-STUDY.md, PROJECT.md and the relevant cloud guides. Local screenshot: /tmp/kav-east-water-well30-trend-fixed.jpg (not committed).

## Previous completed change — Well 10 chlorine binding

On October 21, 2026, the user authorized correcting production https://www.atomhmi.com/site_kav_water/ Well 10 New (well_10_new). Element Chlorine Current PPM / visible Target ppm, ID 4f503dce-01d9-4339-9d14-8668186b2cea, now has updateRegister water:g_well10ChemPPM instead of water:g_well1ChemPPM. Its display already referenced Well 10. Saved the element and page, observed Successfully Saved Page, then freshly reloaded/reopened and compared all 37 configuration controls: only updateRegister changed. No chlorine value, live register write or operational dosing test was submitted. KAV-GUIDE.md, SITE-STUDY.md, PROJECT.md and their relevant cloud notes updated. Evidence screenshot retained locally at /tmp/kav-well10-chlorine-fixed.jpg (not committed).

## Previous completed survey

Atom GIT connection recovered. Read repository AGENTS.md, PROJECT.md, HANDOFF.md, playbook/study/inventory and current cloud knowledge before inspection. Used the signed-in production browser at www.atomhmi.com.

- Created KAV-GUIDE.md, LONE-PINE-GUIDE.md, LAZY-B-GUIDE.md and SWEET-GRASS-GUIDE.md with verified configure-page maps, subtree patterns, representative read/write bindings, units and limits.
- Opened every Lone Pine (16), Lazy B (21) and KAV East Mill (4) page tree. Refreshed all seven site catalogs and selected active pages in the other KAV sites and Sweet Grass; preserved the previous full survey for remaining pages.
- KAV Main Mill has 43 listed pages, including newly observed Mill 2 Test (mill_2_test). Other KAV counts remain 4/29/38; Sweet Grass 51.
- Resolved Sweet Grass Motors and old Motors inspection gaps: 1,435 and 1,246 elements respectively; native root Copy verified.
- Corrected stale clipboard inference, expanded PAGE-INVENTORY.md and added large-tree Copy inspection guidance to ATOMHMI-PLAYBOOK.md.
- Saved cloud copies of the four guides and updated the knowledge-base index/common notes. Repository and cloud documents are manually maintained, not automatically synchronized.

## Previous survey validation and scope

Verified page Title/URL and visible element structure; native Copy root identity checked before interpreting JSON. Findings distinguish this session's inspections from October 20 register/alarm observations. No production page saves/pastes, register/setpoint writes, operational buttons, device actions, alarm modifications or live control tests. No raw JSON/credentials/camera/map endpoints uploaded. Browser timeouts were recovered with clean Configure Pages reloads.

## Findings and resolution status

1. Resolved: Lazy B Wells 2/3 copied timer fields removed; no replacement timer mapping is currently needed.
2. Lazy B: duplicate main/Main navigation targets; M03 PID Tn suffix KP; Loadout internal 100% Button name on an 80-value command. Internal names are not behavioral proof.
3. Lone Pine dryer uses legacy (Dont Use) components. Inspect supported semantics before reusing/migrating.
4. Alarm latching, notification policies, PLC interlocks and operational write behavior were not audited here.

## Next task

Read AGENTS.md, PROJECT.md and the specific site guide; reconnect the signed-in browser and inspect only the current affected subtree. This baseline context survey need not be repeated. Verify intended target before any change to a flagged binding. User's work environments are dev.atomhmi.com and atomhmi.com.

## Prior work retained

Earlier repository import preserved project knowledge. Mandel dashboards/settings, navigation cleanup and 81 non-latching HMI alarms; Shop overview/heating/actuator/trends/diagnostics and 22 non-latching alarms remain in their guides, not new work in this survey. Shop temperature limits remain 0 °C pending commissioning; a separate boiler-water sensor/target was not found. Do not edit Shop atGVL_alarmManager.

Original screenshots/JSON artifacts remain outside this repository. Browser sessions and chat history do not transfer through Git.

## Resume prompt

Use Atom GIT to open Auto/atom-hmi-ui-chatgpt on master. Read AGENTS.md, PROJECT.md, HANDOFF.md and the relevant site guide. Target site: [dev.atomhmi.com or atomhmi.com URL]. Task: [requested change]. Inspect current affected elements, update guides with verified changes, then commit and push documentation.

### Alarm history lookup — October 21, 2026

User requested checking whether any Bin G alarm previously occurred. Opened the alarm icon → Alarm History at https://www.atomhmi.com/site_sweetgrass/alarm_history. Requested Last 30 Days (September 22–October 21, 2026), then freshly reloaded and requested October 21, and finally Yesterday (October 20). Every attempt remained on Loading... without a history table, records, or an explicit empty-result message. No browser warning/error logs were captured. Occurrence, timestamps and retention therefore remain unverified; this is not evidence of no prior Bin G alarms. No acknowledgement, reset, alarm edits, register writes or page saves performed. User-confirmed update on October 21, 2026: Atom HMI was just updated and alarm logs moved to a new database; previous history is no longer available. This supersedes the earlier plan to retry for pre-migration records. Historical Bin G occurrences cannot be determined from the current database. The migration explains unavailable prior history, but the cause of the observed Loading... UI was not independently established. The intended Bin G enable/disable alarm path remains unresolved; determine it from current configuration or user/PLC confirmation.