The As-Found Drawing and Why You Make One
Why this matters
The moment you open a panel or a cover, you become the last person who knows what the inside looked like before anyone touched it. If the job goes sideways, if a second tech comes behind you, or if the customer says the unit worked differently before your visit, the only thing that settles it is a record made before you moved a wire. An as-found drawing is that record. It also happens to be the print you wish had been in the door pocket, because on older equipment and on anything that has been serviced three or four times, no published drawing describes what is actually in front of you.
What an as-found drawing is
An as-found drawing is a sketch of the installation exactly as it exists at the moment you arrive, made before you change anything. Not a design. Not a corrected version. A photograph in line form, with the terminal designations, the conductor routes, and the device positions written down as found.
If you are drawing inside an electrical enclosure, open the disconnect, apply your lock and tag before you put a hand or a pen in the box (29 CFR 1910.333(b)(2) requires the circuit be de-energized and locked or tagged before employees work on or near exposed energized parts), then prove it dead: test your meter on a known live source, test the conductors, test the known source again (NFPA 70E-2021, 120.5). If what you need to record is an operating state rather than a wiring path, do not compromise. Record it from outside the enclosure with the cover on, as observed behavior, and note in the drawing that it was taken energized and from outside.
The negative space: what deliberately stays off it
Most as-found drawings fail not because something is missing but because something extra got in. Every one of these belongs somewhere, and none of them belongs on this sheet.
Design intent. What the circuit is supposed to do is not what the circuit is. The moment you draw the interlock where the sequence says it belongs rather than where the wire actually lands, the drawing becomes a copy of the manual and stops being evidence.
Corrections. If you find a conductor on the wrong terminal, draw it on the wrong terminal. The correction goes on a second sheet, or in your notes, dated and separate. A drawing that silently shows the fixed state cannot answer the question "was it like that when he got there," which is the exact question it exists to answer.
Your judgment about what matters. You do not know yet which wire matters. The abandoned-looking conductor taped off behind the board is the one that turns out to be the disconnected safety three visits later.
A cleaned-up symbol set. Redrawing an odd device as the nearest standard symbol hides the oddity. If you cannot name the device, draw its outline, transcribe every marking on its body, and label it "unidentified, markings transcribed."
Wire colour as a function label. Write the colour as an observation ("blue, 18 AWG, field-added, non-factory jacket"), never as an identity ("neutral"). Colour is a claim someone else made. See the wire-colour card in this category for why that claim fails.
Anything you did not personally see. A conductor that disappears into a conduit and reappears is drawn as entering and leaving the conduit, not as one continuous run, unless you proved continuity end to end.
The minimum field set
Everything below earns its place because leaving it out is what makes a drawing unusable to the next reader.
| Field | Why it is on the sheet |
|---|---|
| Date, time, tech name | Fixes the drawing to a moment; without it the drawing cannot prove sequence |
| Unit identity transcribed off the unit | Model and serial as printed, not as written on the work order |
| Scope boundary | One sentence naming what you drew and what you did not, so absence is not read as "nothing there" |
| Supply source and its overcurrent device | Where the power came in and what protects it |
| Every terminal, landed or empty | Empty positions are data (see below) |
| Terminal designations as printed | Transcribe the marking even when it looks wrong |
| Factory versus field wiring, marked | Different jacket, colour, gauge, or routing; the single highest-value distinction on the sheet |
| Device positions as found | Switch positions, dip switch pattern, valve handle orientation, setpoint dial |
| Photo keys | Numbered arrows on the sketch matched to photo file numbers |
Photograph first, draw second
Take photographs before the pen comes out: one wide shot with the cover off, one per terminal strip square to the strip, one of every label and every handwritten tag, and one of every dip switch bank close enough to read the individual switch positions. Then draw from the photographs, on the tailgate or in the truck, with the panel still open so you can go back and check.
The order matters for two reasons. Photographs are complete in a way a sketch never is, so anything you failed to notice is still recoverable later. And drawing from a photograph rather than from the live panel keeps your hands out of the enclosure while you are thinking, which is when people lean on things.
The drawing is still necessary. A photograph shows you fourteen wires in a bundle and cannot tell you which one goes where. The sketch is where the tracing lives.
Empty terminals, spare conductors, and other things you are tempted to skip
Account for every terminal position, including the empty ones, and say in a note that you did. An unlabeled gap in a drawing is ambiguous forever: did the drawer skip it, or was it empty? Writing "positions 10 and 11 empty, no evidence of prior landing" costs one line and converts a hole into a finding.
The same goes for a conductor that is landed on nothing. Draw it, note where it originates if you can see it, note where it terminates ("cut back and taped, 3 in. free length, behind the board"), and stop there. Do not label it abandoned. Abandoned is a conclusion.
A worked as-found: an eleven-position control terminal strip
A packaged unit with a control section nobody has a print for. Disconnect opened, locked, tagged, meter proved live-dead-live on a known source before the pen came out. The control terminal strip is marked TB1 and has 11 numbered positions.
What the count came to: positions 10 and 11 are empty. Of the 9 landed positions, 7 carry a single conductor and 2 carry two conductors each, so 11 conductor ends land on the strip. Two of those ends belong to the same conductor, a field jumper running from position 5 to position 8, which means 10 distinct conductors are involved. Writing the count out that way, ends first and conductors second, is what forces you to notice the jumper at all. If you count wires you get 10 and the strip looks tidy. If you count landings you get 11 and one of them has to be explained.
The jumper is the finding. It has a jacket that does not match anything else in the panel, it is a lighter gauge than the factory control wiring around it, and both ends are crimped with a ring terminal where every factory landing on the strip uses a fork. Three independent observations, all pointing the same way: somebody added it after the unit shipped. That goes on the sheet as an observation, in those words, and not as "unauthorized jumper" or "bypass."
Positions 6 and 7 carry conductors whose printed designations on the strip are worn to the point that only the second character is legible. Those get drawn with the designation in brackets and a note: "position 6 and 7 markings illegible, numbering inferred from position order, not read." Inferred data is flagged as inferred, every time, or the next reader will treat your guess as a transcription.
Time on the as-found: about 0.4 hours including photographs, all of it unbilled effort absorbed on the first visit. What it displaced: on each of the three later visits to this unit, the tech who arrived would otherwise have spent about 1.0 hour re-establishing the same panel layout from scratch, so roughly 3.0 hours of the same kind of unbilled effort, spread across three visits. Both figures are absorbed effort hours, which is the only reason they can be compared at all. The return is on effort, not on revenue.
What it protects you from on the second call
Three specific situations, and they are the reason this is worth the 0.4 hours even when the drawing is never opened again.
"It was working before you touched it." The as-found is dated evidence of the state you inherited, including the field jumper you did not install. Without it, the argument is your memory against theirs.
A different tech, a different day. The second visit starts from the strip layout instead of from zero, and the jumper is already flagged as a question rather than being rediscovered as a surprise.
The revision that supersedes the print. When a published drawing does turn up and it disagrees with the panel, you have a dated record of the panel to compare it against, which is what lets you tell a superseded drawing from a wrong installation. That adjudication is its own subject and has its own card in this category.
Verifying your as-found is usable by someone else
Hand it, or the photograph of it, to a tech who has never seen the unit and ask three questions. Can they find the supply and its overcurrent device? Can they name what is landed on any position you point at? Can they tell, for any conductor on the sheet, whether it is factory or field?
If any answer is no, the missing piece is almost always one of three: no scope boundary sentence, so they cannot tell what is off-sheet; no factory-versus-field marking, so every wire reads as original; or a designation transcribed from memory after the panel was closed, which reads exactly like a transcription and is not one. The last is the one that bites, because it is invisible. Transcribe while you are looking at the marking, or bracket it as inferred.
References
- 29 CFR 1910.333(b)(2), selection and use of work practices for de-energizing electrical circuits before work on or near exposed energized parts
- NFPA 70E-2021, 120.5, the process for establishing and verifying an electrically safe work condition, including the test-before-touch sequence on a known source
- See related: How to Read a Wiring Diagram You Have Never Seen; Labeling an Install for the Next Tech
- See related: When the Drawing and the Installation Disagree; The Wire Colour Conventions That Are Not Universal