2020 test6 productsAuto Parts & Repair

Which Car Air Filter Brand Wins?

A head-to-head test of 6 car air filter options with the measured results for each. See how they ranked and watch the full test video.

The verdict
Ranked first

EPAuto

Price shown in test: $8.87

Check price on Amazon

Affiliate link. We may earn a commission at no extra cost to you.

The measured results

Every number below is read straight from the test. Scroll sideways to see all measurements. Products are listed in the order they finished.

ProductConstructionNew Filter Airflow0-60mphTest(2003 Suburban, best of 2 runs)ContaminationTest(smallParticles 1-5 micron)ContaminationTest(largeParticles 5+ micron)Static Pressure After Contamination
1EPAuto$8.8780 pleats (most of any filter tested), 2 layers of adhesive, 507 gramsstatic pressure 2.76 in H2O, airspeed 54.7 mph (baseline with no filter: 59.4 mph airspeed, about 0.96 in H2O static pressure)9.4s and 9.4s, best 9.4s, initially the time to beatphase 1 (flour spread) +107; phase 2 (half-second air pulse) 165; phase 3 (1 min continuous airflow) 587phase 1 +76; phase 2 61; phase 3 26increased from 2.76 to about 3.49, the lowest (best) final static pressure of all six filters
2Fram Extra Guard$12.4776 pleats, 3 layers of adhesive, 441 grams, larger pleat hole size than EPAutostatic pressure 2.75 in H2O, airspeed 54.9 mph (best airspeed among filters until K&N was tested)9.5s and 9.5s, best 9.5s (chapter metadata confirms '9.5 seconds average & best time'), slightly slower than EPAutophase 1 173; phase 2 361 (more than twice EPAuto's phase-2 count); phase 3 697phase 1 126 ('next to last' place); phase 2 252 (about four times EPAuto's phase-2 count); phase 3 166 (worst/last place of all six filters)increased from 2.75 to a figure captioned '23.82', almost certainly a caption-mangled decimal for approximately 3.82 based on the closing ranking (EPAuto best, then K&N, then Fram); kept the literal caption value and flagged rather than silently corrected
3AC Delco$16.2376 pleats, 3 different adhesive applications, mesh offset in three areas, 577 grams (heaviest filter tested)static pressure 3.03 in H2O, airspeed lower than Fram's 54.9 mph (exact mph not stated)9.4s and 9.6s, best 9.4s, matched the fastest time at that point in testingphase 1 76 (better than EPAuto and Fram); phase 2 326 (better than Fram, not as good as EPAuto); phase 3 697 (tied with Fram)phase 1 58; phase 2 182; phase 3 157 ('next to last' place, just ahead of Fram)increased from 3.03 to 4.18
4Purolator One$21.5975 pleats, a few lines of adhesive, part of the filter media appeared pressed, 451 gramsstatic pressure 3.09 in H2O, airspeed 54.6 mph (slower than Fram)9.8s and 9.4s, best 9.4s, matched the fastest time despite a notably slow first runphase 1 count actually decreased by 13 from the baseline in the test box; phase 2 increased to 116 (best of all filters at this phase); phase 3 increased by 468phase 1 unchanged from baseline (essentially 0 particles got through); phase 2 increased by only 35 (best of all filters at this phase); phase 3 increased by 14increased from 3.09 to 4.48, one of the two highest post-contamination pressures, framed by the narrator as evidence of superior contamination trapping
5WIX$22.1272 pleats (fewest of any filter), no adhesive applied to the filter media (unlike all other brands), four areas of wire mesh extending outward, 413 grams (lightest filter tested)static pressure 4.07 in H2O (highest/most restrictive of any filter before K&N... actually the single highest new-filter static pressure of all six), airspeed 52.1 mph (lowest of any filter)9.4s and 9.3s, best 9.3s (chapter metadata confirms 'Best 9.3, average 9.35')phase 1 count decreased by 2 from baseline; phase 2 increased by 43 (best of all filters at this phase); phase 3 peaked at 401phase 1 count decreased by 1 from baseline (essentially 0 particles got through); phase 2 increased by 18 (best of all filters at this phase); phase 3 did not increase at all (0, best of all filters)increased from 4.07 to 6.63, the highest post-contamination pressure of any filter, framed by the narrator as evidence of the best contamination trapping
6K&N$44.1932 layers of oiled cotton, wire mesh integrated directly into the filter media (different design than the paper filters), 516 grams, noted as 'very oily right out of the packaging'static pressure 1.72 in H2O (lowest/best of all filters, and lower than the no-filter control's own 0.96 baseline once considered relative to the paper filters), airspeed 57.39 mph (highest/best of all filters)9.3s and 9.2s, best 9.2s, the fastest time of any filter testedphase 1 3604 (by far the worst of any filter, orders of magnitude above the others); phase 2 572; phase 3 407phase 1 2972 (worst of any filter, 'really struggled'); phase 2 421 (worst of any filter at this phase); phase 3 5increased from 1.72 to a figure captioned '2 3.79' (rendered here as '23.79'), almost certainly a caption-mangled decimal for approximately 3.79 based on the closing ranking placing K&N second-best behind EPAuto for post-contamination static pressure; kept the literal caption value and flagged rather than silently corrected

How it was tested

  • baseline (no filter) static pressure and airspeed measurement
  • new-filter static pressure and airspeed measurement
  • 0-60 mph acceleration test in a 2003 Chevrolet Suburban, two timed runs per filter, best time counted
  • particle contamination test with a particle counter tracking 1-5 micron ('small') and 5+ micron ('large') particles across three phases: after spreading 1/2 cup of flour on the filter, after a half-second air pulse, and after one minute of continuous airflow
  • static pressure change from clean to contaminated filter
  • filter construction comparison (pleat count, weight, adhesive layers, wire mesh design)

More Auto Parts & Repair