Short answers, no manual-hunting.

If I ask Dirac to schedule to energize one asset, does it stay focused on just that asset, or can it pull in unrelated work?

By default it stays scoped to that asset's real operational/control dependency chain only. It won't pull in an unrelated asset just because there's spare team capacity, and it also won't pull in an unrelated asset that only shares a control-power hub with your target through the physical-safety gate.

If you specifically want the wider view that also includes safety-linked neighbors, add "including safety chain" (or "with safety dependencies") to the command — e.g. schedule to energize XFMR P including safety chain. It's also in the Dirac Commands panel as "Schedule Energisation (Including Safety Chain)".

What does Dirac need to calculate a fault level, a voltage drop or a loading?

Nameplate data on the assets (for example a transformer's rated power, voltages and impedance; a utility feed's short-circuit power), length and either R/X or size and material on the connections, and Load MW on the assets that consume power. Ask what data is missing for Transformer-A and it lists exactly what is empty. Anything missing is named in the answer and left out of the sum; it is never invented.

I do not have the cable datasheet. Can I still get a voltage drop?

Yes. Enter the length, the size in mm², the material (Cu or Al) and the number of parallel cables on the connection. Dirac then estimates R and X from typical values and marks them Assumed in the answer. Datasheet values, when entered, always win. For a test record or an acceptance decision, use the datasheet.

Why is the fault level on my board higher or lower than my study?

Dirac uses a simplified IEC 60909 method: three-phase, far from generator, voltage factor 1.1, no motor contribution, no transformer correction factor, taps at nominal. It is a quick check from the drawn single line, not a protection study. The answer lists what it assumed and which source data was missing.

Can a checklist step run a calculation and record the result?

Yes. In the Checklist Manager pick a Calculation for the step (or fill the Calculation column in the template XLSX). On the asset's checklist the step shows Run calculation: the working goes into the step's notes and Pass or Fail is set from the result, which you can overrule by tapping another result.

Is there one document that explains the whole application?

Yes, the Application Overview: what each part does, how the parts depend on each other, how it is built and deployed, and what is built versus planned. It is the first tab in the help library after the Dirac manual, and Dirac answers "what does the application do" from it.

Why did "Import .json" on the Project Manager screen reject my file?

The file is checked before a project is created from it. It must contain a blocks array in which every entry has an id, a kind of asset, room or packageManager, a numeric position (x, y) and size (width, height); any connections need an id plus from.blockId/to.blockId, and any platforms need a height above 0 and a footprint with positive width and height. A rejected import shows a red message listing each problem by array index and field (for example blocks[3].position.x: must be a finite number.), up to 20 entries, so a hand-edited file can be fixed and re-imported; nothing is created until the file passes.

If I export Connections to XLSX, edit the file, and reimport it, will it overwrite the existing connections?

Not cleanly — it's additive, not a true round-trip overwrite. Connections are matched purely by (from-asset, to-asset, layer), since the export has no connection-ID column:

• Unedited row: matches the existing connection and is skipped — safe no-op.
• Edited row (changed endpoint or layer): doesn't update the original — creates a brand-new connection alongside it. You'd end up with both.
• Deleted row: the existing connection is left alone — import only processes rows present in the file, there's no delete-reconciliation pass.
• New row: imports correctly as a new connection.

If your edits are pure additions this is safe. If you're changing or removing existing connections, delete the stale ones by hand afterward.

Can I import the issues we track in Autodesk Construction Cloud (ACC) and link them to our assets?

Yes — export them from ACC as a BCF file and use ⬇ BCF in the Issues window. Every topic becomes an issue; the asset link is taken from the asset name in the topic title or description (ACC exports almost never carry IFC GUIDs, so the name is what the matcher looks for), and you can correct each link in the preview before anything is written. Snapshots become photos, comments become the issue's Activity with their original dates, and importing a newer export later updates the same issues rather than duplicating them. See the Commissioning Tracker manual's Issues section for the status mapping and what is kept on re-import. The reverse works too: ⬆ BCF exports the issues (all, open, or the current filter) as a BCF 2.1 or 3.0 archive, with comments, photos and an IFC selection for assets that have a GlobalId.

How do I mark a control connection as a backup/reserve source instead of a normal dependency?

Tick "Backup/Reserve Source" on that connection in the connection-info panel (control-layer connections only). Use it for a real, physically-wired feed that only kicks in once other sources fail — e.g. a battery that's normally charged by a DC source but can become that source's emergency backup. Tagging it this way keeps the wire visible in canvas/BIM, but it creates no scheduling dependency in that direction: it's excluded from final-stage readiness, the L3 startup gate, the control-gates-mains rule, and cycle detection. The normal direction (the thing charging/feeding the backup) stays a real predecessor as usual.

This is different from "AC Auxiliary," which still gates final-stage readiness — backup/reserve gates nothing.

How do I model an LV AC auxiliary circuit (heaters, lighting) without it counting as real DC control power?

Tick "AC Auxiliary" on that control-layer connection. It's excluded from the L3 startup gate and the control-gates-mains rule (since it isn't real DC control power), but it still has to be satisfied for the asset's actual final stage — whatever that resolves to (L4 normally, or L3 if L4 is waived/not required).

If a project has waived L4 (or any final stage), does the validation report show it as pending?

No. The per-asset Stage Status table, the summary counts (Done / Not Released / Pending), and the Test Execution and Documentation Requirements sub-tables all check the same effective-final-stage logic used everywhere else in the app (project-wide waiver, or an individual asset's own "Final Stage Not Required" tick). A waived stage renders as "NOT REQUIRED" (or is skipped entirely in the checklist/requirements sections) and is excluded from the tallies — it's never counted as pending, and never force-marked complete either.

Does the validation report's "Prepared By" field fill in automatically?

Yes — it uses the name you type into the E-Signature field before generating the document (plus today's date). If you generate without typing a name, it falls back to a blank "Name: / Date:" line as before. "Reviewed By" and "Approved By" are left blank on purpose — those are different people who haven't signed yet at export time.

Does the validation document's Protocol No. reflect the project's mode (datacenter/substation/pmOnly), or is it always pharma-style?

It's mode-aware. A pharma project keeps its usual "IQ"/"DQ-OQ"-style code. Every other mode uses the same short stage label(s) shown elsewhere in the app instead (e.g. "L2", "L2-L3", "FULL") — so the filename and title-page Protocol No. match the mode-appropriate title, instead of always reading "...-IQ-123456" regardless of mode.

Does the "Print / Save PDF" export always come out portrait?

Yes — the report's print styling forces portrait orientation, so it doesn't depend on your browser/printer's last-used page-orientation setting.

Does the generated validation report include Documentation Requirements, or only Checklists?

Both. Each asset's "Test Execution" section has its own "Documentation Requirements" sub-table per stage, alongside the existing Checklist sub-tables — covering the equipment-type template requirement, any per-asset additional requirement, and any fully-shared requirement, with the same real completion status shown in the Commissioning Tracker.

Files attached to a requirement are not printed inline in that table (they might be lengthy source documents) — instead every attached file across the whole document is listed once in an "Addendum: Attached Documentation Files" section at the very end, as an asset/stage/requirement name plus a clickable link to the actual file.

The closing "Final Approval Signatures" table's role names are also mode-aware: a pharma project shows "Project Manager (PMC)" / "QA / Reviewer (SPC)"; datacenter/substation/pmOnly projects show the same roles without the pharma-specific abbreviation.

Does "Generate Word Document" produce a real .docx file now?

Yes. It generates a genuine Word (.docx) file with real page breaks and exact-width table columns, built from the same data as the Print/PDF export — a reliable improvement over exporting an HTML file with a .doc extension, which Word's own import renderer could handle unreliably (no guaranteed page breaks, no reliably fixed-width table columns). The "Print / Save PDF" button is unchanged and still opens a print-ready window.

One deliberate difference: the Word export always uses the local (non-AI) Summary & Conclusion template, so it stays instant and works fully offline. The AI-drafted conclusion (when a project's T0.5 Dirac tier is enabled) is only available in the Print/PDF export.

Why do all the tables in a validation document now have the same column widths, even when one asset has much longer text than another?

Every table "shape" (e.g. the Checklist sub-table, the Stage Status table) uses a fixed percentage-width profile shared across every asset and every document, rather than auto-sizing each table's columns to that one instance's own content. If a checklist item's text is too long to fit, it wraps onto a new line rather than squeezing the column narrower or getting cut off — this applies to both the Word (.docx) and Print/PDF exports.

Can I generate a validation document for just the assets in one room, or one custom asset group?

Yes. The Asset Scope dropdown has two more options alongside "All assets", "Selected asset", and "By equipment type": "By room" (every asset spatially inside a chosen Room block on the canvas) and "By asset group" (every asset that's a member of one of your named custom asset groups from the Scheduler's Parallel Capacity Windows — the same groups used for group capacity limits and joint final-test scheduling). Both options only appear if the project actually has rooms or asset groups defined.

Why does my BIM Status Report show its column headers again partway through the report when I print it or save it as PDF?

It isn't a stray "Description column" — it's the same header row (Asset Name / Equipment Type / Stage / FR Ready / L1-L4 / Open Issues) reappearing. Browsers' print engines automatically repeat a table's header row at the top of every printed page once the table runs longer than one page — that's standard browser print behavior, not something this report opted into. At this table's row height a page break lands roughly every ~15 assets, so the header can print again there, which reads like a stray row when you scroll continuously through an exported PDF.

The print stylesheet overrides this so the header prints once, at the top, and never repeats. The live in-app panel already has its own sticky header and isn't affected by this — it's scoped to printing/PDF export only.

Why can't I add or delete items on a fully-shared (🔗🔗) checklist right away?

By design — click "🔒 Enable Edit" in that checklist's header first. Fully-shared checklists are one shared record every member asset sees, so add/delete controls stay hidden until you deliberately unlock editing, and it resets to locked again the next time you reopen the modal. Deleting an item or section after unlocking also asks you to confirm first, since it removes that item for every member asset at once and can't be undone. Per-asset (🔗) additional checklists aren't affected — only one asset is at stake there, so those keep the old single-click add/delete.

I couldn't scroll down in Checklist Management (or Requirements Management) to reach or delete items past the first few — is this fixed?

Yes. Scrolling now works correctly in both the equipment-type Checklist/Requirements template editors and the per-asset Checklist popup: the scrollable stage-list area shrinks correctly within its fixed-height window, and each individual stage panel (header + expanded items) keeps its full natural height instead of being squeezed. That means real overflow is detected and the scrollbar appears whenever a stage has enough items to need it.

How do I let an asset reach its final stage without waiting for its upstream to be ready?

Tick "Bypass upstream readiness for final stage" on that asset (next to "Final Stage Not Required" in the commissioning panel). It only affects the final-stage gate — whatever stage that asset's final actually is (L4 normally, L3 if waived). It has no effect on the L3 startup gate or the control-gates-mains rule; those stay fully enforced regardless.

This is a different flag from "Final Stage Not Required" — that one changes which stage counts as final; this one changes whether that final stage waits on upstream readiness.

How do I test a loss of utility on a plant that is not commissioned yet?

Press the amber Scenario button on the canvas toolbar. It works on a private copy of the project where, as designed, every asset counts as installed and commissioned and the plant starts from its normal state, so nothing needs to be marked Available or signed off to L4. Run simulate loss of Utility-A in Dirac and read the timeline. Nothing is saved; Exit scenario returns to the real project untouched. Switch to As built in the strip to test only what is Available and signed off. Not offered in Project Management Only projects, or where there is no utility feed or generator.

How do I get a logic diagram for just one asset / one room / one group?

Ask Dirac: logic diagram for <name>. Variants: upstream logic of X (what gates X), downstream logic of X (what X releases), control logic of X / field logic of X (one layer; in PM mode office/field/support), logic diagram for all transformers (equipment type), … depth 4 to widen. The result is a derived diagram: boxes are asset × stage tasks, &/≥1 gates and coloured edges come straight from the Scheduler's predecessor graph, columns are the order of execution, joint/lockstep groups are drawn as brackets, and the whole thing is read-only ("mirror only"). Dirac also answers with the same logic in words.

Why do the layer buttons say Field / Office / Support instead of Operational / Control / Mechanical?

Because the project is in Project Management Only mode. The three layers are the same layers with management words: Field (operational — "must be finished before"), Office (control — "needs this decision/approval before"), Support (mechanical — access/materials/permits). Switch the mode dropdown in the canvas toolbar and the words switch back; the wires, ports and rules never change.

I don't know what to wire to what — is there a guided way?

Type plan this with me in Dirac. It asks for the work items, then per item "what must be finished before it?" and "which decision or approval must be in place before it?", shows a review, and draws the plan on the canvas when you say apply. Then ⚡ Reschedule in the Schedule. Naming something not yet on the list adds it; naming an existing asset reuses it; cancel exits at any point. See the Dirac AI manual's "Plan this with me" section for more.

Assets share a tagging checklist — will the schedule treat them as one?

No. Each asset keeps its own stage, colour, requirements and tasks; the group is a section on top. The canvas shows it as a dashed ring with a badge (gathering n/m → ready → tagged → signed), the Tracker shows the same line, and the Schedule's 🏷 Sections list gives the section date. If you want the schedule to pull a section's assets forward together, tick Guide by sections in the Schedule and press ⚡ Reschedule; it only reorders tasks that compete for the same free slot. With no tagging checklist nothing changes. See the Commissioning Tracker manual, "Sections", and the Scheduling manual.