Electrical Takeoff from 2D Drawings: Workflow, Checklist and Quality Checks
A dependable electrical takeoff from PDF drawings comes from a controlled set, not from faster clicking: current revisions and verified scales, the project's own legend, counts recorded by sheet and zone, run lengths built from stated assumptions, and a reconciliation against the specifications and the bill of quantities. Anything the drawings do not show goes on the sheet as 'to confirm', never as a silent estimate. This guide sets out that method, a blank electrical takeoff template (Excel) you can download, and a two-level quality check.
- Published
- Sources checked
- Reading time
- 12 min
- Written by
- Mirage Metrics team
Scope: building electrical packages (lighting, power, containment, and low voltage or fire alarm devices where they are in your scope) measured from 2D drawings. It covers quantities, not labour units or pricing. Units in the example are feet; the method is the same in metres.
Decision snapshot
- Best suited to
- Electrical estimators at specialist subcontractors and general contractors, and the reviewers who sign off their takeoffs.
- Inputs needed
- The full drawing set with its register, legend, schedules, single-line and panel schedules, specifications, architectural sections, any issued BOQ, addenda and RFI answers.
- What can be automated
- Sheet indexing and revision comparison, symbol counts per sheet and zone against the project legend, measuring drawn routes at a verified scale, assembling the takeoff sheet with a source reference per line, flagging symbols missing from the legend.
- What requires human approval
- Confirming scales, interpreting ambiguous symbols, choosing length assumptions, deciding what to qualify or exclude, raising RFIs, and signing off the reviewed quantities.
- When this approach is insufficient
- When the design is incomplete (performance specification, design-and-build electrical), when routes are not drawn at all, or when a coordinated model exists and should be the source. Then the takeoff is an allowance, and should say so.
Documents to collect before counting
A takeoff started on half a set is redone later. Collect the documents below first and log what is missing, because a missing schedule is often the reason a count cannot be priced.
| Document | What you take from it | Typical failure without it |
|---|---|---|
| Electrical plans by system | Lighting, power, containment, fire alarm or low voltage layouts, floor by floor | Devices counted on the wrong system, or twice |
| Legend and symbols sheet | The meaning of every symbol, suffix and note tag on this project | A symbol read by habit from another job |
| Luminaire and equipment schedules | Fixture types, mounting, emergency variants, equipment ratings | Counts without types, so no price |
| Single-line or riser diagram | Distribution boards, feeders, sizes, vertical routes | Feeders and risers missing from the takeoff |
| Panel schedules | Circuits per board, spare ways, circuit numbering | Home runs impossible to trace |
| Specifications (Divisions 26, 27, 28 in MasterFormat, or your market's equivalent) | Materials, installation methods, testing, what the contractor supplies | A priced item the specification excludes, or the reverse |
| Architectural backgrounds and sections | Ceiling types and heights, floor-to-floor heights, walls | No basis for vertical drops and riser lengths |
| Bill of quantities, if issued | The client's item structure and quantities | A takeoff that cannot be compared line by line |
| Addenda and RFI answers | Changes issued after the set | Pricing a superseded design |
Source: Division numbers from CSI MasterFormat. Projects outside North America use other specification structures; the check is the same.
Revision and scale control
Most takeoff errors that survive review are not counting errors. They are counts made on the wrong revision, or lengths measured at the wrong scale. Both are cheap to prevent and expensive to find later, so run these checks before the first count and record the result per sheet.
| Check | When | If it fails |
|---|---|---|
| Title block revision matches the drawing register | Every sheet | Stop and ask for the current register |
| Revision clouds and deltas read and listed | Every revised sheet | List the changes before counting that sheet |
| Stated scale checked against a known dimension (grid spacing, a dimensioned wall) | Every sheet you measure on | Recalibrate; note the measured scale on the sheet log |
| PDF page size matches the drawing's intended sheet size | Every set | Scale by known dimension only, never by the printed ratio |
| Sheets marked NTS (not to scale) identified | Every set | Count only; take no lengths from them |
| Enlarged plans and details mapped to the general plan area they repeat | Every enlarged plan | Exclude that area from the general plan count |
Why the printed ratio is not enough
A PDF exported at a different page size than the drawing was drawn for keeps its stated scale in the title block but not on screen. Calibrating on a known dimension, such as a structural grid spacing, is the only check that survives resizing. Record the measured scale next to the sheet number so the reviewer can repeat it.
Symbols and legends
Read symbols against the project's legend, not against memory. Standard symbol sets exist: the IEC maintains the IEC 60617 database of symbols for electrotechnical diagrams, and IEEE 315 served the same purpose in North America before being made inactive-reserved in 2019. Building layouts rarely follow either strictly. Design firms keep their own symbol libraries and add suffixes, subscripts and note tags that change what a symbol means on this project.
Four legend checks that catch most misreadings
- Suffixes and subscripts. The same downlight symbol with "EM", a circuit letter or a type tag is a different item. List every suffix found on the plans and match it to the legend.
- Symbols used but not in the legend. Each one is an RFI, logged with the sheet and location.
- Symbols in the legend but not used. Harmless, unless a schedule lists quantities for them: then something is missing from the plans.
- Notes that change quantities. "Typical of 6 units", "mirror of level 2", "provide in all corridors": notes carry quantities that no symbol count will find.
Counts by zone and by sheet
A total without a location cannot be reviewed. Define zones before counting: a floor, a wing, a grid range, or the zones the client's BOQ uses if it has them. Then count each symbol type per sheet and per zone, and keep those sub-totals on the takeoff sheet. When a reviewer disagrees with a total, the zone sub-totals show where to look.
- Count one system at a time (lighting, then power, then containment) so a device is never counted on two layers.
- Handle typical floors explicitly: count the typical floor once, record the multiplier and the sheets it applies to, and count the exceptions separately.
- Mark counted symbols on the drawing (a layer, a colour, a stamp) so the reviewer sees what was counted and what was not.
- Keep a sheet log: sheet number, revision, scale verified, systems counted, date, estimator.
Run lengths and their assumptions
Lengths are where 2D takeoffs are weakest, because electrical drawings often show a device and a destination but not the path between them. Split every length into what was measured and what was assumed, and write the assumptions down in words the client can check.
- Measured. Containment and conduit routes that are actually drawn, measured at a verified scale along the drawn path.
- Derived. Vertical drops from ceiling height (from a section or reflected ceiling plan) to device height (from the specification or a mounting note). The inputs are documented, so the result is checkable.
- Assumed. Home runs drawn as an arrow to a panel, feeders with no route, allowances for terminations and waste. These are your estimate, stated as one: for example, "home run length = measured distance to the panel location along the corridor containment, plus drops at both ends". The percentage you add for waste is a company rule, not a standard.
The line software cannot cross
A tool can measure a drawn route faster than a person and flag every arrow that has no route. It cannot know a length the design has not defined. Any product that returns a confident length for an undrawn home run is estimating on your behalf, and you should ask what assumption it used.
Home runs are the largest of these assumed lengths on most jobs. The guide to electrical home run calculation works one through, with the voltage drop, conduit fill and derating checks its length drives.
Reconciling drawings, specifications and the BOQ
When the client issues a bill of quantities, reconcile it against your own takeoff line by line before pricing. The three sources disagree more often than they should, and each disagreement is a pricing risk until someone decides which one governs.
| Item | Takeoff | Specification says | BOQ | Variance | Action |
|---|---|---|---|---|---|
| Downlight type A | 38 | Type A specified in 26 51 00 | 40 | -2 | Recount zone L1-North; check the enlarged lobby plan |
| Emergency downlight A-EM | 6 | Battery pack integral, 90 min | 6 | 0 | None |
| Receptacle 'W' | 4 | Not described | 0 | +4 | RFI-07 open; price as qualified item |
| Cable tray 12 in | 212 ft | Galvanized, with 25% spare capacity | Lump sum | n/a | Price per ft internally; state the measured quantity in the qualifications |
The order of precedence between drawings, specifications and the bill is set by the contract documents, not by the estimator. Where it is silent or unclear, the discrepancy goes to the designer as a question during the bid period, and the assumption you price on goes into your qualifications.
Missing and conflicting information
Every takeoff has gaps. What separates a defensible bid from a risky one is whether those gaps are visible. Give each one a status and a decision:
- To confirm. The item is counted but its meaning or type is uncertain. It carries an RFI number and a pricing assumption.
- Missing. The item is required (by a schedule, a single-line or the specification) but not shown on any plan. It carries an allowance, clearly labelled.
- Conflict. Two documents disagree. It carries the question and the assumption you priced on.
- Excluded. Deliberately not priced, with the reason, and listed in the bid's exclusions.
The bid qualifications then list every assumption and allowance that remains open at the submission date, in the same words as the takeoff sheet.
Worked example: a takeoff sheet with its sources
Illustrative example, invented figures: first floor, north zone, of an office fit-out
Assumptions, all invented for this example. Sheets E-101 (lighting), E-201 (power), E-301 (containment) at revision C, scale verified on the structural grid. Ceiling height from section A-301. Switch mounting height from the specification. Assumption A1 for home runs: measured distance to panel LP-1A along the corridor containment, plus a drop at each end.
| Line | Item | Plan source (sheet, zone) | Qty | Unit | Basis | To confirm | Status |
|---|---|---|---|---|---|---|---|
| E-01 | Recessed LED downlight, type A | E-101 lighting, zone L1-North | 38 | ea | Counted on plan, type from luminaire schedule | Counted | |
| E-02 | Recessed LED downlight, type A-EM (emergency) | E-101 lighting, zone L1-North | 6 | ea | Suffix EM per legend note 4 | Counted | |
| E-03 | Duplex receptacle, 20 A | E-201 power, zone L1-North | 54 | ea | Counted on plan | Counted | |
| E-04 | Receptacle, symbol with suffix 'W' | E-201 power, zone L1-North | 4 | ea | Suffix not in legend | Is 'W' weatherproof? RFI-07 | To confirm |
| E-05 | Cable tray 12 in, horizontal | E-301 containment, zone L1-North | 212 | ft | Measured along drawn route at verified scale 1/8" = 1'-0" | Measured | |
| E-06 | Branch circuit home runs, lighting | E-101 arrows to panel LP-1A | 14 | circuits | Count only; route to panel not drawn | Length assumption A1, see notes | Assumed length |
| E-07 | Vertical drops to switches | E-101 plus section A-301 | 22 | ea | Ceiling height 10 ft from section, switch height 48 in per spec | Derived | |
| E-08 | Feeder LP-1A from MDB | Single-line E-601 | 1 | ea | Size from single line | Route not shown on any plan | Missing route |
Three lines need a person before pricing. E-04 has a suffix the legend does not define, so it is counted, priced on a stated assumption, and raised as RFI-07. E-06 is a count with an assumed length, and the assumption is written where the reviewer will see it. E-08 is a feeder that the single-line requires and no plan routes: it cannot be measured, so it becomes a labelled allowance or a question, never a number that looks measured.
Download the blank electrical takeoff template (Excel, 30 KB): the columns of this sheet with system, revision, assumption and reviewer notes added, a sheet log, and drop-down lists for basis and status.
Two levels of quality control
Level 1: the estimator's own checks, before handing over
- Every sheet in the register appears in the sheet log, with revision and verified scale
- Every symbol and suffix found on the plans maps to a legend entry or an RFI
- Zone sub-totals add up to the item totals
- Typical floor multipliers and their sheets are recorded
- Every length line says measured, derived or assumed
- Every open question has a status, an RFI number and a pricing assumption
Level 2: an independent review
- Recount on the drawings, not on the takeoff sheet: the top items by value, every 'to confirm' line, and a random sample of the rest
- Remeasure two or three drawn routes at the recorded scale
- Re-derive one vertical drop and one home run from the stated assumptions
- Check the reconciliation against the specification and the BOQ
- Record differences by zone, agree the correction, and keep the review with the takeoff
Set your tolerance in advance: for example, which differences on a count trigger a full recount of that zone. The threshold is a company decision; having one written down is what makes the review consistent between estimators.
Limits when there is no BIM model
A 2D set supports a sound device count and measured drawn routes. It does not give you coordinated three-dimensional routes, clash-free containment, exact vertical routing in risers, or conditions hidden in existing buildings. When those drive the price, ask for what would remove the uncertainty (a model, a survey, a site visit) or carry it explicitly as an allowance.
The reverse also holds: a model is only as complete as what was modelled. Check whether wiring, small containment and supports were modelled before trusting its quantities over a careful 2D takeoff.
Evaluating a takeoff tool on a real bid
Demonstrations are built on clean drawings. The test that tells you something uses a bid you have already closed, with a takeoff your team trusts, so every difference can be checked against a known answer.
| Criterion | What to ask | How to test it |
|---|---|---|
| Traceability | Can every quantity be opened at the exact sheet and location it came from? | Pick ten random lines and trace each one in under a minute |
| Legend handling | Does it read this project's legend, including suffixes and notes, or a built-in symbol library? | Include a sheet with a non-standard suffix and see what happens |
| Revision handling | What happens to counts when a sheet is reissued? Does it show what changed? | Feed the tool revision A, then revision B of the same sheet |
| Scale | How is scale set, verified and recorded per sheet? | Use a PDF printed at the wrong page size |
| Uncertainty | Does it mark what it could not read, or return a number for everything? | Count how many lines come back flagged; zero is a warning sign |
| Lengths | Where do lengths come from: drawn routes, stated assumptions, or inference? | Ask which lengths were measured and which assumed |
| Output | Does the export keep your item structure, zones and the source column? | Import the result into your estimating sheet unchanged |
| Effort | Time to a reviewed takeoff, including the estimator's checks | Time the whole run, not just the extraction |
A test protocol in five steps
- Pick a closed bid with its final reviewed takeoff and the drawing set as issued, including at least one revised sheet and one non-standard legend entry.
- Give the vendor the drawings only, not your takeoff.
- Compare counts by symbol type and zone. Note agreements, differences and items the tool flagged as uncertain.
- Trace a random sample of lines to the drawing and check each source reference.
- Measure the total time including your estimator's review, and compare with the time the original takeoff took.
Where Mirage fits
Mirage deploys agents that read a plan set, extract quantities and link each one to the sheet and location it came from, for the estimator to check before anything reaches a bid; the electrical takeoff page describes that workflow. The published deployments are on other trades: at OCBM, a Moroccan general contractor, on formwork and structural drawings with every quantity traceable to its source; at John Cockerill, on piping, boilermaking and structural steel drawings. Mirage has not published an electrical deployment. The fair test for an electrical lot is the protocol above, run on one of your own closed bids.
Questions readers ask
Can software measure cable lengths from a 2D electrical plan?
It can measure what is drawn: a containment route or a conduit run at a verified scale. It cannot know a length the drawings do not define, such as a home run shown only as an arrow to a panel, or a vertical drop with no ceiling height given. Those lengths come from assumptions the estimator states and the client can check.
Which symbol standard should I read electrical plans against?
The project's own legend first. Symbol databases such as IEC 60617 and older standards such as IEEE 315 (inactive-reserved since 2019) describe symbols for diagrams, but design firms adapt symbols, add suffixes and define notes per project. A symbol that is not in the legend is a question for the designer, not a guess.
How do I avoid double counting on enlarged plans?
Map every enlarged plan to the area of the general plan it repeats, and count that area once, on the enlarged plan. Record the mapping in your sheet log so a reviewer can see why the general plan count excludes it.
What should a reviewer recount first?
The items with the highest value or the highest quantity, any line marked 'to confirm', and a random sample of other lines. Recount on the drawing, not from the takeoff sheet, and compare by zone so a difference points to a place.
Is a BIM model always better than 2D drawings for an electrical takeoff?
A coordinated model gives routes and lengths that 2D drawings leave to assumption, if the electrical systems were actually modelled to that level. Many models stop at devices and major containment. Check what was modelled before treating model quantities as complete.
Sources and scope
Checked on . Scope: quantity takeoff for building electrical packages from 2D drawings. The method, checklists and example are Mirage's own and use invented figures; the external sources below cover naming standards only.
- Construction Specifications Institute, MasterFormat: Division 26 Electrical, 27 Communications and 28 Electronic Safety and Security as the specification divisions for electrical work
- IEC 60617, Graphical Symbols for Diagrams (database): the international database of graphical symbols for electrotechnical diagrams
- IEEE 315, Standard for Graphic Symbols for Electrical and Electronics Diagrams: status: inactive-reserved since November 7, 2019
Related on this site
- AI electrical takeoff software: the Mirage workflow for electrical sets
- AI quantity takeoff for contractors: the same workflow across all trades
- Electrical home run calculation: the manual method and the NEC checks a home run's length drives
- Plan revision comparison: what changed between two issues of a sheet
- CSI MasterFormat document extraction: organising Divisions 26 to 28 before pricing
- AI takeoff and estimating software: how the tools differ: Mirage's buyer's guide to the categories of tools
- Accubid, McCormick and ConEst alternatives: the electrical estimating platforms compared
- Voltage drop calculator: the check a home run length drives
- OCBM case study: traceable takeoff on structural and formwork drawings
Run the test on one of your closed bids
Tell us about a closed electrical bid and we will arrange a secure way to review the drawing set. We will return a takeoff with a source reference on every line and the items we could not read flagged, so you can compare it with your own.
Prefer to talk it through? Book a call