Taste and Odor Complaint Investigation

Purpose

The nose is a real instrument and it is being used badly. A tech leans over a running tap in a room that already smells of the customer's kitchen, decides "yeah, I get it", and the ticket says "odd taste". Under controlled conditions the same nose separates a plastic character from a musty one from a solvent one, and those three route to three different owners: the pipe the shop just installed, the source water, and a contamination event nobody should be selling a filter against. This procedure makes the character a recorded observation rather than an impression.

Scope

Covers investigating a taste or odor complaint on potable water at a house, duplex, small retail unit, small restaurant or office, once a visit has established that the complaint is in the water rather than at the drains.

Does not cover the first-visit triage that establishes the zone, owned by the water quality complaint visit SOP, and this procedure cites that verdict rather than re-deriving it. Does not cover hot-side sulfide and the heater anode, which that SOP owns end to end, nor sewer gas and dry traps, owned by the drain odor complaint investigation SOP. Does not cover laboratory analysis, sample collection protocol, or any statement about whether the water is safe to drink.

Roles and responsibilities

Procedure

1. A petroleum, solvent or chemical character is not a taste complaint, and it stops the visit. Acceptance: water use for drinking and cooking stopped, the building ventilated, the water system and the health authority notified the same day, and the buried service material and any site fuel history written down, because permeation through a plastic service in contaminated soil and a cross-connection are the two mechanisms that produce it. What wrong looks like: proceeding to a treatment quote. Stop rule: no device is sold, no cartridge is fitted, and the shop does not clear the water; the routing is the deliverable. Hazard: never taste water to diagnose, and do not run a hot shower in a closed bathroom to reproduce the smell - a volatile organic that reaches you by ingestion at a tap reaches you far faster by inhalation in a steamed, closed room.
2. Take the complaint apart into one character word, an intensity and a timeline, before you open a tap. Acceptance: a single character chosen from a stated list - chlorinous, sulfurous, earthy or musty, metallic, plastic or waxy, petroleum or solvent, medicinal - plus the onset date, whether it is constant or only after the water sits, which fixtures, and what changed at onset. What wrong looks like: "bad taste" on a ticket, forcing the next visit to rebuild the only diagnostic information there is. Stop rule: where the customer cannot separate a taste from a drain smell, fill a glass and carry it to another room first, because sewer gas follows the room and not the glass. Hazard: none here, an interview at the kitchen table, and the discipline is verbal - no opinion is offered on whether the water is safe to drink.
3. Run the odor test under controlled conditions rather than sniffing at the tap. Acceptance: odor-free glassware rinsed without detergent, a low-odor blank alongside the sample, both warmed to the same temperature near 40 C, the lower of the two temperatures the standard threshold odor method uses, and the result recorded as the dilution at which the character disappears, with the blank presented blind. What wrong looks like: comparing a cold sample against a warm one, which changes volatility and not chemistry. Stop rule: a solvent or fuel character appearing at temperature sends the job back to step 1, whatever was originally reported. Hazard: warming liberates the volatile you are trying to smell, so waft across the vessel rather than inhaling over it, keep the room ventilated while the samples are warm, and handle hot glassware in a rack rather than by hand.
4. Use two testers independently, and manage nose fatigue as part of the method. Acceptance: two character words written down before either person speaks, a rest of at least a minute between presentations, and no coffee, scented product or smoking beforehand. What wrong looks like: one tech saying "sort of musty?" and the second agreeing, producing one opinion wearing two names. Stop rule: two different words means the test repeats on a fresh sample rather than being averaged into a compromise. Hazard: repeated sniffing deadens the sense of smell within a few presentations, so a character that fades during testing is fatigue and not improvement - odor is never an all-clear, and the release criterion is a measurement, never a nose.
5. Cite the zone verdict rather than re-deriving it, and let it close the branches it can. Acceptance: the hot-versus-cold verdict, the fixture map and the first-draw-versus-flushed comparison taken from the water quality complaint visit and written on this ticket, each with its tap named. What wrong looks like: repeating those tests badly under time pressure and getting a different answer than the visit that did them properly. Stop rule: a sulfurous character on the hot side only belongs to the heater and its anode under that SOP, and this investigation closes there. Hazard: running hot at a fixture for a comparison delivers water that can scald quickly, so run it into a vessel standing in the sink, never one held in the hand.
6. Clear the treatment devices before you blame the pipe. Acceptance: a paired character reading upstream and downstream of every device on the affected path - carbon cartridge, reverse osmosis tank, softener, whole-house housing, refrigerator line - plus each cartridge's age against its labeled interval. What wrong looks like: sampling only downstream of an under-sink filter, which cannot separate a device failing to remove a character from one adding it. Stop rule: a device whose downstream sample carries a character its upstream does not is corrected first, and nothing else is credited until the pair matches. Hazard: a filter housing holds line pressure, so close the inlet, open the downstream fixture or the housing's relief button until flow stops, and support the sump before unscrewing it - a sump released under pressure comes off in one movement and takes a hand with it.
7. Match a plastic or waxy character to a construction event, then measure the decay rather than arguing about it. New plastic pipe, solvent cement and flux residue all give off flavor that declines with use and flushing; a garden hose, a temporary connection or a non-potable fitting does not decline, because it is still there. Acceptance: a timeline tying onset to a dated construction event, plus at least two repeat tests at a stated interval showing direction, with at least three volume exchanges of the new pipe flushed between tests. Gallons per foot is 0.0408 times the square of the published inside diameter in inches. What wrong looks like: a single retest called an improvement. Stop rule: a character that does not fall across the stated checks is not new-pipe off-flavor. Hazard: a sustained flush at a laundry box or bib can overwhelm a floor drain, so watch the drain rather than the clock and keep the discharge off finished flooring.
8. Route what is not yours, restore what you disturbed, and set the recheck. An earthy or musty character with no premise-side pattern is usually source water and belongs to the water system; a chlorinous character tracks the residual and a seasonal disinfectant change and belongs there too. Acceptance: the routing named in writing with the reading behind it, every device back in service with a fresh dated cartridge, every stop returned to full open, each disturbed fixture run clear at full flow, and a recheck date set. What wrong looks like: a verbal "it is the city's water" with nothing written, which the customer cannot take anywhere. Stop rule: nothing closes on the day the device was changed, since a character takes days to re-establish if the correction did not hold. Hazard: this step puts pressure back with the customer at the sink, so open each stop slowly with the fixture open to bleed air, keep faces out of line with outlets, and hold every reopened housing under full line pressure for a stated period, watching the sump seal, before anyone leaves the building.

The record this produces

One investigation sheet per complaint: each tester's independent character word; the sample and blank temperatures and the dilution at which the character disappeared; the cited zone verdict with tap names; the upstream and downstream pair at every device with each cartridge's age and interval; the construction timeline where one exists; the flush volume and exchanges between retests; the decay series with dates; the routing and to whom; and the restore checks.

The decay series is the part that earns its keep. A single reading proves nothing about new-pipe off-flavor; three dated readings either support it or kill it. Where the shop's own repipe is the source, that series also tells the customer when it will stop, which is a better answer than a filter.

One worked pass, including a failure

Townhouse repiped in PEX five weeks ago, complaint of a plastic taste in the cold water, worst first thing in the morning. Office confirms the water system reports no main work, no flushing and no disinfectant change.

Step 2: character offered as "plastic or waxy", onset about a week after the repipe, cold and hot, every fixture, worse after the water sits. Step 3: sample and blank both warmed to 40 C; the character disappears at a four-fold dilution and the blank is called clean blind. Step 4: both testers independently write "plastic", no discussion. Step 5: zone verdict cited from the earlier visit - both sides, all fixtures, first draw stronger than flushed, which puts it in the premise piping rather than upstream.

Step 6 fails. The kitchen under-sink carbon cartridge is 14 months old against a labeled six month interval, and the paired samples do not match: upstream reads plastic at a four-fold dilution, downstream reads plastic plus a distinctly musty note the upstream does not carry. Stop rule taken: the device is corrected first. Cartridge replaced and dated on the housing, housing sanitized per its instruction, pair re-run - downstream now matches upstream at a four-fold dilution with no musty note.

Step 7: 120 ft of 1/2 in PEX at a published inside diameter of 0.475 in gives 0.0408 times 0.2256, that is 0.0092 gal per ft, so 120 ft holds 1.10 gal and three exchanges is 3.31 gal, delivered in 3.31 divided by 2.6 gpm measured at the laundry box, about 1.3 minutes. Flushed on that basis before each retest. The series falls twice: at week 5 the character disappears at a four-fold dilution, at week 7 at a two-fold dilution, and at week 9 it is detectable only in the undiluted sample. Step 8: no external routing needed, the finding written up as new-pipe off-flavor with the three dated readings, stops reopened full, the new cartridge held under line pressure for 15 minutes with a dry sump seal, and a final recheck booked at week 13.

References

• Standard Methods for the Examination of Water and Wastewater, threshold odor test, in the edition your laboratory works to; it sets the glassware, the blank and the two test temperatures.
• Your water system's consumer confidence report and its notice of any seasonal disinfectant change, which explains a large share of chlorinous complaints at no cost.
• The adopted plumbing code's restriction on plastic piping in contaminated ground, in the edition your jurisdiction enforces, which is what makes a petroleum character on a plastic service a routing question rather than a treatment one.
• Filter and cartridge manufacturer instructions for interval, sanitizing and relief before opening a housing.
• See related: the Plumbing water quality complaint visit SOP, the drain odor complaint investigation SOP, and the point of use filter selection and installation SOP.