Ductless Indoor Unit Cleaning and Odor Complaint
Purpose
This standing instruction guarantees that a ductless odor complaint is either resolved by a wet clean of the surfaces that actually hold the growth, or proved not to be a head problem at all, with the customer standing there when it is verified.
A ductless wall head has no return filter upstream of anything, so the cross-flow blower wheel and the coil face collect what the room gives them. The growth lives down between the wheel vanes and on the leaving face of the coil, and neither surface is reachable by washing the front screen, which is what most previous visits did. The second half of the failure is that a clean head with a lazy drain smells again within weeks, so a cleaning procedure that does not prove the drain has bought the customer a repeat.
Safety actions that gate the work
- Open the head's breaker and prove dead at its terminal block before any cover comes off, proved on a known live source before and after per NFPA 70E-2021, 120.5, with work practices at 29 CFR 1910.333(b)(2), written for qualified persons under 1910.332. This procedure sprays water at a cabinet that contains a live control board.
- The dry deposit on a blower wheel goes airborne the moment the cover moves, not when the spray starts. Respiratory protection matched to a bioaerosol route, under a written program per 29 CFR 1910.134, goes on before the cover comes off and stays on through the rinse. Gloves and glasses do not address an inhalation route.
- A respirator named on a product label is a condition of using that product. If nobody on site is in a 1910.134 program, the crew does not do the work that required one; it is rescheduled with someone who is.
- Read the SDS for the cleaner before it leaves the truck, and match eye, skin and ventilation controls to what it actually is. Alkaline coil cleaners burn skin slowly enough that people do not notice until afterward.
- Never treat smell as an all-clear. The nose fatigues, so a space that stops smelling bad is not evidence of anything; the acceptance conditions below are visual and volumetric.
Scope
Covers odor complaints and wet cleaning of residential and light-commercial ductless indoor heads, wall-mount and floor-mount, from complaint triage through verified odor resolution.
Does not cover the annual service inspection that finds the growth in the first place, which the mini-split annual service SOP owns and which hands off to this document. Does not cover condensate diagnosis beyond the flush proof in step 5, which the ductless condensate SOP owns. Does not cover ceiling cassettes, which are an overhead-access job with their own platform requirements, or whole-house indoor air quality investigations.
Roles and handoffs
| Role | Owns | Hands off |
|---|---|---|
| Office | Capturing the complaint in the customer's own words at booking, including mode and timing | A ticket that says musty, sour, sewage or burning, because those four go different places |
| Service tech | Every step below, including the verification with the customer present | The cleaning record with before and after photographs |
| Service manager | The call when the odor survives a verified clean | A revisit scoped to the drain and the room rather than another clean |
| Customer | Filter washing between visits, and telling the shop when the smell returns | A written interval and a described normal |
Procedure
Take the complaint apart at the door before you open anything. Acceptance: the mode it happens in, whether it appears at startup and fades, roughly how long it lasts, and whether any other head or room is affected, all in the customer's words. Wrong looks like hearing "musty" and going straight to a wet clean. Stop rule: a sewage smell goes to the condensate SOP because that is a drain connection, and a burning smell means the breaker goes off now and the fault is diagnosed electrically before any cleaning is discussed. Hazard: none at this step, it is a conversation at the door, which is exactly where the two dangerous cases get separated out.
De-energize and build the containment before the first cover comes off. Acceptance: breaker off and proved dead at the head's terminal block, cleaning bib hung and draining to a bucket, floor and furniture covered, and the electronics compartment bagged against overspray. Wrong looks like hanging the bib after the cover is off, which is when the loose deposit lands on the customer's floor. Stop rule: no bib, no wet clean; a dry brush-out of a growth-laden wheel puts the whole deposit in the room. Hazard: water and a live board share this cabinet, so the proving step is not a formality here and the bag goes on before any liquid is opened.
Inspect the wheel vanes and the leaving face of the coil, photograph both, and set the level of clean. Acceptance: images clear enough to see between the vanes, and a decision recorded as wet clean or filter-and-face only. Visible growth down in the vanes is a wet clean; a light grey dust film is not. Wrong looks like judging from the front screen, which is clean on almost every head that smells. Hazard: pulling the cover disturbs the deposit, so respiratory protection is already on per the safety block; this is the step the exposure actually happens at, before any spray.
Wet clean the wheel and coil to the label and to the coil manufacturer's material compatibility. Acceptance: dilution as the label states, dwell for the stated time, rinse until the water leaving the bib runs clear, and the vanes clean to the base of the vane on inspection rather than at the tip. Wrong looks like a strong cleaner left to dry on aluminium, which etches fins and takes the coil with it. Stop rule: a wheel that will not come clean in place gets pulled, or the head gets quoted for replacement; it does not get signed off dirty. Hazard: brushing and rinsing aerosolizes both the biofilm and the cleaner, so the respirator stays on through the rinse and the room is ventilated, and any eye or skin control the SDS names is worn for the whole wet stage.
Prove the drain runs before you close anything up. Pour a measured 8 oz into the pan and watch the outdoor termination. Acceptance: discharge at the termination inside 30 seconds with no rise standing in the pan afterward. Wrong looks like water sitting in the pan, which is a growth surface that will bring the smell back within weeks whatever you did to the wheel. Stop rule: no discharge inside 30 seconds stops the reassembly and the job goes to the condensate SOP before this one finishes. Hazard: you are pouring water into an open cabinet, so the breaker stays off and the board stays bagged until this step is signed.
Reassemble, restore power, and prove what you disturbed still works. Acceptance: every cover and the electronics compartment dry and seated, breaker closed, head restarts, blower runs at each speed with no imbalance rattle, and after a 20 minute cooling run the pan is empty and the termination is discharging. Wrong looks like a wheel reinstalled out of balance or a float switch left unplugged, both of which are quiet until the customer is asleep. Stop rule: any imbalance noise means the wheel comes back out. Hazard: this is the step that puts power back next to a cabinet you have just had water in, so confirm it is dry, stand clear of the head when you close the breaker, and let the unit start on its own rather than helping it.
Verify the odor with the customer at the head, and set the follow-up rule. Run 20 minutes in cooling, then switch to fan only, and have the customer say what they smell at the discharge. Acceptance: their statement recorded, and a written interval for filter washing and for the next inspection. Wrong looks like verifying it yourself after two hours in that room, when your own nose has adapted to it. Stop rule: an odor that survives a verified clean is not the head, so stop cleaning and open a drain and room-side investigation rather than quoting a second clean. Hazard: none at this step, it is a conversation, but it is the step that decides whether the shop owns a callback.
The record this produces
A cleaning record filed to the equipment:
- The complaint in the customer's words, with mode, timing and duration
- Before photographs of the wheel vanes and the coil leaving face, and after photographs of both
- The product used, its dilution and dwell time, and the material compatibility source
- The 8 oz flush result, in seconds, and whether water stood
- Blower speeds checked, float switch confirmed connected, pan condition after the 20 minute run
- The customer's own statement at step 7, quoted
- The interval given for filter washing and for the next inspection
The next tech reads the before and after photographs before quoting anything on this head, because the second clean on a head that was genuinely clean the first time is a refund waiting to happen. The manager reads the flush line across the book: a run of odor complaints on heads whose drains all ran slowly is an install habit.
Worked pass: wall-mount head in a bedroom, musty at startup in cooling
Step 1: customer says musty, only in cooling, strongest for the first two or three minutes after it starts, then fades. No other room affected. Not sewage, not burning, so it stays in this procedure.
Step 2: breaker off, proved dead at the terminal block, bib hung and draining to a bucket, bed and floor covered, electronics compartment bagged.
Step 3: respirator on before the cover moves. Photographs show growth down between the wheel vanes across most of the wheel and a film across the leaving face of the coil. Front screen is nearly clean, which is consistent with a customer who has been washing it monthly. Wet clean recorded.
Step 4: cleaner selected against its SDS and the coil manufacturer's compatibility note, mixed at the label dilution, dwelled the label time, rinsed until the bib runout ran clear. Vanes inspected and clean to the base.
Step 5 FAILS. 8 oz poured into the pan. Nothing at the outdoor termination at 30 seconds, and the water is still standing in the pan with the level just under the front lip. Stop rule taken: reassembly halted, job handed to the condensate SOP, which finds a partial blockage in the drain hose inside the wall sleeve and clears it. Back to step 5 with the same measured 8 oz: discharge at the termination at 11 seconds, pan empty afterward.
Step 6: covers and the electronics compartment confirmed dry and seated, breaker closed, head restarts on its own, blower runs at each of its speeds with no rattle. After the 20 minute cooling run the pan is empty and the termination is discharging.
Step 7: after the 20 minute cooling run the mode is switched to fan only with the customer at the head. Her words, recorded: "nothing, it just smells like the room." Filter washing set monthly and the next inspection at the annual service.
Checking the pass against the sections above: step 5's acceptance names both a time and the absence of standing water, and the run prints the failing result against both halves and the passing result against both, using the same measured 8 oz. Step 4's acceptance names the vanes clean at the base rather than the tip, and the run reports the inspection at the base. Step 7's acceptance is the customer's own statement rather than the tech's nose, and the run quotes her, which is the safeguard the safety block's point about nose fatigue exists to create.
References
- The coil and head manufacturer's cleaning instructions and material compatibility notes; an aggressive cleaner on the wrong coil is a coil replacement
- The safety data sheet for the product actually used, which governs dilution, dwell, ventilation and personal protection
- 29 CFR 1910.134 for the written respiratory program that gates steps 3 and 4, and 29 CFR 1910.333(b)(2) with NFPA 70E-2021, 120.5, for de-energizing and proving dead
- See related: mini-split annual service, ductless condensate and line set troubleshooting, indoor air quality assessment visit