UV Disinfection Not Killing Bacteria - Decision Tree

Why this matters

A homeowner gets a positive coliform or E. coli result on a well-water test, calls their UV maintainer, and demands to know why the UV system they just paid for "is not working." A large share of those calls turn out not to be the UV at all. Sample collection, source-side contamination, plumbing between the UV and the tap, and a half-dozen UV-specific failure modes all have to be ruled out before swapping a lamp. This decision tree walks the diagnosis in the order that catches the most causes fastest, calibrated to the dominant residential UV brands (VIQUA / Trojan, Luminor, Atlantic Ultraviolet, R-Can).

The decision flow at a glance:

  UV light not curing bacteria - why?
  |
  +-- 1. Sample collection wrong? --> RESAMPLE CORRECTLY
  |
  +-- 2. UV not energized? ---------> POWER / LAMP /
  |                                   BALLAST
  |
  +-- 3. Sleeve fouled? ------------> CLEAN SLEEVE
  |                                   (MATCH TYPE)
  |
  +-- 4. Intensity under 50%? ------> REPLACE LAMP
  |
  +-- 5. Bypass / parallel flow? ---> SEAT BYPASS VALVE
  |
  +-- 6. No 5-micron prefilter? ----> ADD PREFILTER (NSF
  |                                   55)

Quick triage table

Use this to pick the starting step. Work down from whatever line matches what you found on arrival.

Two rules that save callbacks. Never change a lamp as your first move; it fixes one failure mode out of eight and hides the rest for a month. And always shock-chlorinate and resample after any repair that broke the water seal, because the repair itself introduces organisms into a line the UV cannot reach.

Step 1 - confirm the sample collection

Bacterial samples are easy to contaminate. Before troubleshooting the UV:

• Sample bottle: 120 mL sterile, sodium-thiosulfate dechlorinated. Old reused bottles or non-sterile mason jars produce false positives.
• Tap: a cold-water tap upstream of any faucet aerator, run for at least 5 minutes before sampling. Garden hoses and bib spigots are notorious for biofilm contamination - never sample from them.
• Hands: gloves. The homeowner's hand on the bottle's lip plates skin flora.
• Transit: refrigerated, delivered to the lab within 24 hours (the EPA holding-time limit for coliform).

Pull a fresh sample yourself before assuming the original was valid. If you and the lab get a clean result on a properly collected sample, the original may have been a collection error.

Step 2 - confirm the UV is actually energized

Walk to the UV reactor. Confirm:

• The system controller LED is illuminated and shows normal status (green steady, not red / blinking).
• Lamp life remaining (display reads, typically 365 days of cumulative on-time before lamp swap). A VIQUA Sterilight uses a year-count; a Luminor LB6 displays in months.
• The audible alarm is not silenced (the silence button often disables the alarm but not the fault state - look at the LED).
• If a UV intensity monitor is present, the readout is above the manufacturer's alarm setpoint for Class A service, which NSF/ANSI 55 sets at a minimum dose of 40 mJ/cm2. Class B is 16 mJ/cm2 and is a supplemental-treatment rating for water that is already known to be safe; it is not a disinfection rating for a coliform-positive well. If somebody has this system commissioned to a Class B setpoint on a well with confirmed bacteria, that alone is the fault.

A failed power supply, blown fuse, or unplugged controller delivers zero UV. The lamp can be physically intact and still putting out no UV - intensity monitor or no UV at all is the only way to know.

Step 3 - check the quartz sleeve

Pull the quartz sleeve and inspect:

• Heavy iron / manganese staining on the sleeve = effectively zero UV transmittance. Acid-soak the sleeve (CLR, or a dilute muriatic solution around 5 percent, 10 minutes, then triple-rinse with deionized water). Acid work means gloves and eye protection, ventilation, and acid poured INTO water rather than the reverse. Never bring acid near a line or vessel that still holds shock chlorine; acid plus hypochlorite makes chlorine gas. Flush the chlorine out completely first.
• Cracked or chipped sleeve = leaking water onto the lamp, full system failure, lamp shorts.
• Cloudy or hazy sleeve = sleeve at end of life (10 years typical, less in iron / hardness conditions). Replace.

Sleeve P/Ns for the major brands:

• VIQUA Sterilight S5Q-PA / S8Q-PA / S12Q-PA: QS-810, QS-330, QS-1240 respectively.
• Luminor LB5: LR5-058A sleeve.
• Atlantic Ultraviolet Sanitron S5: 05-1480.

Step 4 - confirm lamp age and lamp output

UV lamps lose output over time. A 9000-hour-rated lamp at 8000 hours is producing roughly 65-70% of its initial UV output. NSF/ANSI 55 Class A requires the system to maintain 40 mJ/cm2 at end of lamp life - if the lamp is past end-of-life, dose may have fallen below the bacterial-inactivation threshold even with the sleeve clean and the controller showing healthy.

Replace any lamp that is past its rated hours regardless of whether the controller says it is still in service. Common lamp P/Ns:

Look the lamp up by the model number on the reactor label, not off a chart. Reactor bodies of different lengths take different lamp lengths, so a bigger unit in the same family does NOT share the smaller unit's lamp, and it does not share its sleeve either. If a parts list ever hands you the same lamp for two units with different sleeves, the list is wrong.

Always use OEM lamps. Aftermarket 254 nm lamps without low-pressure amalgam construction underperform; the few dollars saved per lamp cost the homeowner a re-contamination event.

Step 5 - measure UV transmittance (UVT) of the source water

UV dose is a function of intensity and water UV transmittance. NSF/ANSI 55 Class A systems are rated at 95% UVT at 254 nm; well water can be much lower:

• Clear, clean well water: 90-95% UVT.
• Slightly tannin-stained water: 75-85% UVT.
• Iron-bearing water (over 0.3 ppm): often below 70% UVT.
• High dissolved organic carbon (DOC): below 60% UVT.

If UVT is below the system's rated UVT, the effective dose at the reactor exit is below the inactivation threshold even with a perfect lamp and clean sleeve. The fix is pre-treatment - sediment filtration, iron filter, carbon filter, or water softener upstream of the UV. Measure UVT with a Hach DR1900 or a Hanna HI801 spectrophotometer at 254 nm.

Step 6 - check flow rate against the rated capacity

UV dose is also a function of contact time. Push more water through the reactor than its rated flow rate and dose drops below inactivation threshold. Confirm:

• Peak GPM flow through the reactor (run all major fixtures - shower, kitchen sink, washing machine simultaneously).
• Manufacturer's NSF/ANSI 55 rated flow at 95% UVT, read at the Class A 40 mJ/cm2 dose. Every residential reactor is published with two flow figures, one at 40 mJ/cm2 (Class A) and a much larger one at 16 mJ/cm2 (Class B), and the big number is the one on the sales sheet. Sizing a disinfection job off the Class B flow is the single most common way a correctly functioning UV system fails to disinfect: the reactor is doing exactly what it was rated to do, at a third of the dose the water needs.

A homeowner who just added a whirlpool tub or a second bath is often exceeding the UV rating without realizing it. Solution: upsize the UV reactor.

Step 7 - look for post-UV contamination

The most overlooked cause. The UV inactivates organisms passing through it, but if any plumbing downstream of the UV is biofilm-fouled, bacteria can re-colonize:

• Long copper or PEX runs between the UV and the kitchen tap, especially if the system was added after the plumbing was already biofilm-fouled. Sanitize the entire downstream system with a chlorine shock (1 gallon household bleach per 1000 gallon storage, 24-hour soak, flush).
• Cold-side dead legs (capped lines that hold stagnant water for days).
• Failed in-line carbon filter downstream of the UV (carbon supports bacterial growth; a carbon block should always be UPSTREAM of the UV, not downstream).
• A pressure tank that has not been sanitized in years.

If post-UV contamination is the cause, the lab will return a positive on a sample at the tap but a clean sample at the UV outlet. Pull a sample at the UV outlet bypass port to confirm.

Step 8 - rule out the well itself

If everything downstream is healthy and the UV is delivering dose, the source is the issue:

• Well cap loose or damaged (surface water intrusion).
• Casing crack or corroded steel casing within 50 ft of grade.
• Septic system or livestock pen too close (most state codes require 100 ft from septic to well).
• New construction nearby disturbing the aquifer.

This is a well contractor's call, not the UV maintainer's. Document and refer out.

References

• NSF/ANSI 55-2023 - Ultraviolet Microbiological Water Treatment Systems
• US EPA Safe Drinking Water Act, 40 CFR Part 141 - Total Coliform Rule
• VIQUA Sterilight Series Owner's Manual, document 520091-R Rev N (2023)
• Luminor LB5 Residential UV System Installation and Operating Instructions, 2023
• EPA Ground Water and Drinking Water Coliform Sampling Guidance, EPA 815-R-13-002
• WQA UV Disinfection Technical Application Bulletin (2024)