How to Work Out Whether a Backup Came From Inside or Outside

Why this matters

Every sewage backup answers one question before it answers any other: is the restriction between the fixtures and the building's exit, or is it past that point where the shop has no authority and often no responsibility. Get it wrong in the inward direction and you cable a line that was never blocked while the real head sits outside waiting for the next storm. Get it wrong in the outward direction and you tell a customer to call the utility about a grease plug under their own kitchen. The evidence that answers it is fragile: it is a water level, a set of which-fixtures-participated observations, and one moment at a cleanout, and every clearing tool you own erases all three. So the steps below are ordered by what destroys what, and each one names the action that would have wiped it out.

Before anyone goes down the stairs

Electricity first, and it is not negotiable. A flooded basement contains energized receptacles, a sump or ejector pump, and usually the appliance the customer wants saved. Do not step into standing water until the circuits serving that area are opened at the panel and each one is verified dead with a tester proved live on a known source before and after. Electrical isolation of a panel or branch circuit is 29 CFR 1910.333(b)(2), not the general lockout standard, which excludes work on conductors in electric utilization installations at 1910.147(a)(1)(ii)(C). If the panel, the meter base or the service equipment itself is standing in the water, nobody enters: the utility de-energizes it at the service first.

The water is contaminated, not just dirty. Sewage backup water carries pathogens and whatever the building put down the drain. Waterproof boots, cuffed gloves, and eye protection go on before entry, and any cut or abrasion is covered before the gloves. Wash and change before eating or driving.

Ventilate before you spend time down there. A surcharged line pushes sewer gas into the building ahead of the water. Open the exterior access, get cross-ventilation started, and if the space is small, unventilated or below grade with a single access, treat it as a confined space and work it under 29 CFR 1910 Part 1910.146 in general industry or 29 CFR 1926 Subpart AA on construction work. Hydrogen sulfide deadens the sense of smell well below the concentrations that hurt you, so the fact that the smell is fading as you work is a reason to leave, not a reason to relax.

Step 1: Stop every water-using fixture in the building, and say why out loud

Nobody flushes, nobody runs a tap, and the dishwasher, washing machine and ice maker are all switched off at the appliance. A single laundry discharge is tens of gallons in a couple of minutes, which will raise the level you are about to measure and will contaminate the only test in this procedure that has a clean yes or no answer.

Skipped: your high-water mark now records the customer's shower, not the sewer, and the idle-rise test in step 3 becomes meaningless because you cannot tell the building's own discharge from the street's.

Step 2: Read and mark the water level and the participation set, before anything is pumped

Two readings, both taken with a tape from a fixed datum, and the datum written down. Use the basement finished floor.

• The high-water mark. Find it on cardboard, on a mechanical pad, on the wall, or as a debris line inside the floor drain box. Photograph it with the tape in frame.
• The participation set. Walk every fixture opening in the building and record its rim or outlet height above the same datum, and whether it overflowed, gurgled, or did nothing. Rim height is the number that matters, because a surcharged system relieves at the lowest opening it can find and fills from the bottom up.

That set is the shape of the fault. If everything below one elevation participated and nothing above it did, the branches are all seeing one common water level, which puts the restriction downstream of the point where those branches join. If one group of fixtures participated and a lower fixture on a different branch stayed clean, the restriction is on that branch, upstream, and inside.

Skipped: the first shop-vac pass erases the level, the mop erases the debris line, and no later observation recovers either. This is the step most often lost, because the customer is standing there asking you to start pumping.

Step 3: Run the idle-rise observation before you open anything

With every fixture off, mark the water level, wait a measured interval, and mark it again. Thirty minutes is usually enough and it is time you can spend setting up. A level that climbs with zero building discharge can only be fed from the connection side. A level that holds or falls slowly is consistent with a restriction the building itself filled.

Read the rise as a difference between two tape marks, and treat each mark as an independent random spread rather than a systematic offset, because you are re-finding the waterline by eye each time. Two equal independent spreads combine in quadrature, so the uncertainty on the difference is the square root of two times the single-mark uncertainty, not twice it and not the same as it.

Skipped: the moment any cleanout is opened, the head relieves and the rise can never be observed again for this event. You will be back on the next storm to get the answer you could have had today.

Step 4: Open the exterior cleanout, and watch it while it opens

This is containment and diagnosis in the same motion, which is why it is one step. Relieving at the exterior cleanout sends the surcharge to the yard instead of the finished basement, and the behaviour at the instant the plug breaks free is the single most decisive observation available.

Stand to the side with the plug's path away from your face, wear a face shield, and back the plug out slowly so it vents before it clears the threads. A line under head will drive the plug and its contents out with real force.

• It gushes and the riser is standing full: there is head in the piping downstream of that cleanout. The fault is outside.
• It is empty and the basement level does not fall: the restriction is between the basement fixtures and that cleanout. The fault is inside.

Never open an interior cleanout that sits below a full exterior riser. You will put the head that is currently outside into the room you are standing in. Establish the exterior condition first, every time.

Step 5: Camera before cable, once the line will accept a camera

If the fault reads as outside, the line gets looked at before it gets cabled. A cable rearranges roots, punches a channel through grease, and moves debris off the defect that caused it, so the camera run after cabling shows you a line that has been cleaned rather than a line that failed. Cross-reference the sibling card on camera inspection for how to read what you see; this card's job ends at deciding whose line it is.

Skipped: you clear it, it flows, you leave, and the report says "roots" when the actual finding was a bellied section that roots merely colonized.

Worked example: one house, one storm, one answer

A 1950s two-storey with a full basement, called in during the tail of a heavy rain. Datum is the basement finished floor.

Participation set, measured with a tape:

High-water mark on the boiler pad and the boxes beside it: 4.0 in above datum.

Read that set before doing anything else. One common level at 4.0 in, the only opening below it relieved, the two above it did not, and the highest one is far enough above to be irrelevant. The first-floor gurgle is displaced air, not water, which tells you the stack is passing air past a restriction rather than passing water. That pattern is a fault downstream of where the branches join, not a branch fault.

Idle-rise observation. All fixtures off for 30 measured minutes. First mark 4.0 in, second mark 4.6 in, so the rise is 0.6 in. Each mark carries about plus or minus 0.1 in of read uncertainty, independent on each mark, so the difference carries the square root of two times 0.1, about 0.15 in. The rise is four times its own uncertainty and the sign is not in doubt. Had it come back at 0.1 in, that is inside the noise and the test would have said nothing at all.

Converting the rise to a rate is worth doing once, with the approximation named. The wetted floor area is roughly 600 sq ft, which is an eyeball figure and not a measurement, so this result is an order of magnitude rather than a value: 600 sq ft x 144 sq in/sq ft x 0.6 in = 51,840 cu in, divided by 231 cu in per gallon = 224 gallons, over 30 minutes = about 7.5 gpm. Every fixture in the building is off. Nothing inside this house is producing 7.5 gpm.

Exterior cleanout. Plug backed out slowly with the face shield on and the body clear. It vented, then discharged under head, and the riser stood within about 8 in of grade before it started running out at the top of the box. The hydraulic grade line outside the building was roughly 100 in above the basement floor.

Conclusion, and what each piece contributed. The participation set said one common level, so not a branch. The idle rise said water was arriving with the building idle, so not the building's own discharge. The cleanout said the head was outside, and gave its elevation. Three independent observations, all pointing the same way: the restriction is downstream of the exterior cleanout, and this is a call to the utility with a written record attached, not a cabling job.

Failure mode. The version of this call that goes wrong starts at step 5. The tech arrives, pulls the basement cleanout plug because it is closest to the van, takes the head into the basement, and then cables the kitchen branch because that is what came back on the cable. The line drains once the street relaxes, the ticket says grease, and the customer floods again in six weeks with a new finished floor in between. Nothing in that sequence was incompetent. The evidence was simply spent before anyone read it.

Verifying the call before you write it down

• Re-read your participation set against your conclusion. If you wrote "everything in the basement backed up", count the basement openings and confirm every one of them did, including the one you are least likely to have rechecked. In the case above only one of the two basement openings participated, and that is the fact that makes the level meaningful.
• State the idle-rise result with its uncertainty, not as a bare number. A 0.6 in rise against a 0.15 in combined uncertainty is a finding. The same 0.6 in reported without the uncertainty is an assertion.
• Check timing against the rainfall record rather than against the customer's memory of when it rained. A backup that tracks rainfall closely belongs with the wet-weather cards; a backup with no rainfall relationship at all points back inside regardless of what the cleanout showed on one visit.
• Write the datum on the ticket. Heights above "the floor" are worthless to the next tech if nobody said which floor.

References

• 29 CFR 1910.333(b)(2), electrical work practices for de-energizing and verifying circuits; 29 CFR 1910.147(a)(1)(ii)(C), the carve-out that sends electrical exposure there rather than to the general lockout standard
• 29 CFR 1910 Part 1910.146, permit-required confined spaces, general industry; 29 CFR 1926 Subpart AA, the construction counterpart
• The model plumbing code as adopted and amended by your local jurisdiction, which is what sets where the building's responsibility for the lateral ends
• See related: What Inflow and Infiltration Look Like From the Building Side; What a Backwater Valve Protects and What It Breaks; Sewer Camera Inspection Reference