Used Car Inspection Checklist 2026: Paperwork, Walk-Around, Test Drive and the OBD2 Check for Cleared Codes

I built PulsCar after a $380 "full diagnostic" in Phuket in 2025 led to roughly $6,000 over six months across several shops, and the sound never went away (the story is on the About page). That is why this checklist leans on one question a buyer can answer on the seller's driveway: was the check engine light turned off by a repair, or by a scan tool ten minutes before you arrived?

The checklist on one screen

The same five passes, with the home lookups split out and the decision added. Take a friend: one drives, one watches the dashboard.

The time estimates are ours, not a regulator's. The order is the point: paperwork costs little or nothing and can end the visit early, the walk-around needs no tools, the OBD2 check needs a reader (yours, borrowed or the mechanic's), and the professional inspection costs real money.

How we built this checklist, and how the marks work

Every claim below comes from a page we opened on 30 September 2026 and saved as text: chiefly the FTC's Used Car Rule, 16 CFR Part 455, and its consumer guide to buying from a dealer; NHTSA's odometer disclosure rule, 49 CFR Part 580; the Department of Justice's National Motor Vehicle Title Information System; the EPA's OBD test procedure in 40 CFR Part 85, Subpart W, which a 1996 amendment made an official OBD performance warranty short test, and its OBD guidance to state inspection programs (EPA420-R-01-015, June 2001); and California's Smog Check documents where the federal text stops.

Two marks, no third. 🟢 Government means a regulator, regulation or official inspection document printed it. ⚪ Maker's own claim means a company published it about its own product. A ⚪ labels the source, not the truth of the claim.

This is not a price guide (inspection prices and diagnostic prices have their own pages), not a scanner ranking, and not legal advice: where the federal text sends you to your state, so do we.

Step one, at home: the VIN and three lookups

Get the VIN in writing before you drive anywhere; federal rules make it easy to sanity-check. Under 49 CFR Part 565, "Each VIN shall consist of seventeen (17) characters" (NHTSA's decoder does not decode pre-1981 years), with no letters I, O or Q and a check digit in position nine. On passenger cars and light trucks it must be readable "through the vehicle glazing" by an observer standing outside "adjacent to the left windshield pillar." 🟢 A VIN with the wrong length or a forbidden letter is a question for the seller.

Lookup one: decode it. NHTSA's VIN decoder returns what the manufacturer reported for that VIN; in NHTSA's words, "The information displayed through NHTSA’s VIN decoder is reported by the manufacturer." Compare the decoded model year, body and engine with the listing. On passenger cars and light trucks the tenth character's year code can mean 1980-2009 or 2010-2039 depending on position seven, which is why a decoder beats reading it by eye. 🟢

Lookup two: open recalls. The FTC's used car resource page points buyers to NHTSA's lookup: "Enter the vehicle’s VIN to find out if a vehicle is subject to an open recall." An open recall, in the FTC's words, means "the manufacturer has recalled the vehicle but repairs are not yet done." The Buyers Guide gives the address: "To check for open safety recalls, visit safercar.gov." NHTSA's pages, including safercar.gov and nhtsa.gov/recalls, refused our automated requests, so this description is the FTC's, not a reading of NHTSA's page. The FTC gives NHTSA's Vehicle Safety Hotline as 1-888-327-4236. 🟢

Lookup three: title history, which gets its own section next, because it is the lookup most easily over-read.

While you are at the keyboard, spend two minutes on the model itself: if PulsCar has a guide for it, check the known problems for this model before you see the car, so you can ask about a known noise, leak or code by name. The list of guides is further down.

Title history through NMVTIS: what a brand is, and what a clean report does not prove

The National Motor Vehicle Title Information System, run for the Department of Justice, collects data from state titling agencies, insurers, and junk and salvage yards. Before purchase, its consumer page says, a buyer can learn the current title's brand history, "The latest reported odometer readings", any insurer determination that the vehicle is "salvage" (including total losses), and any report of the vehicle going to a recycler or salvage yard. 🟢

The word that matters is brand: "A “brand” is a descriptive label that states assign to a vehicle to identify the vehicle's current or prior condition, such as “junk,” “salvage,” “flood,” or another designation." Once a state brands a vehicle, NMVTIS says the brand becomes a permanent part of its record, which NMVTIS is designed to use against brands being "washed" off by retitling in a state that does not check. The site's own example is flooded cars moved out of Louisiana after Hurricane Katrina and "dried out, cleaned, and readied for sale to unsuspecting consumers in states that do not brand flood vehicles." 🟢

Now the limits, in the system's own words. "A NMVTIS Vehicle History Report is intentionally concise." It covers five indicators aimed at fraud and theft; none of them, in our reading, is a mechanical one. The page states that 87 percent of the U.S. vehicle population is in the system, that more than half of the states report data to it, and warns of "the possibility that a search may yield a false negative indication, such as "no junk or salvage history."" And states brand differently: damage one state brands "salvage" may fall under another state's threshold. 🟢

Getting a report. The approved-provider page lists providers "with no preference indicated" and warns: "Consumers CANNOT receive NMVTIS Vehicle History Reports from Carfax, DMVDesk, or Experian; these entities provide information only to car dealerships." That undated sentence concerns NMVTIS reports only; the FTC lists Carfax.com among providers of other, separate history reports. The FTC's used car resource page says you choose a provider, enter the VIN and "pay the provider’s fee", and that other providers' reports "sometimes have additional information, like accident and repair history". 🟢

The flood check. The FTC's flood alerts (2022, 2021) point to the National Insurance Crime Bureau's free database, which shows flood damage, a theft never recovered, or a salvage declaration, "but only if the car was insured when it was damaged." NICB's page refused our request, so we quote the FTC's description. 🟢

A brand is a reason to walk or re-price; a clean report is the absence of one kind of bad news. In the FTC's words, "A vehicle history report is not a substitute for an independent vehicle inspection."

The Buyers Guide on the window: what "As Is" actually means

If the seller is a dealer, the key document is taped to the car. Under 16 CFR Part 455, a used-vehicle dealer must display a Buyers Guide on each used vehicle offered for sale, "in such a fashion that both sides are readily readable." It may come off for a test drive, "but you must return it as soon as the test drive is over." The rule's "vehicle" excludes motorcycles and, among other limits, anything with a GVWR of 8,500 lbs. or more or a curb weight of 6,000 lbs. or more, so some heavy trucks and vans carry no Guide. The FTC's Dealer's Guide adds that a Guide in a glove compartment, trunk or under a seat "is not conspicuous because it is not in plain sight." 🟢

The official English form opens with "Spoken promises are difficult to enforce. Ask the dealer to put all promises in writing. Keep this form." Then the warranty box that applies is checked.

The FTC's consumer page translates the first box: "As Is - No Dealer Warranty means the dealer won’t pay for any problems or needed repairs. You’re assuming the risk of anything that goes wrong after the sale." On an as-is car, your own inspection before signing is the protection.

Three details matter. State law can override it: if your state "limits or prohibits “as is” sales of vehicles, that State law overrides this part", and the dealer must use the Implied Warranties Only version; the FTC sends buyers to their state attorney general for the local rule. The Guide beats the contract: if the Guide shows a warranty and the contract says as is, "the dealer must give you the warranty described in the Guide." A service contract can bring warranties back: the form says a service contract bought "within 90 days of your purchase of this vehicle" may give you additional rights under state implied warranties. 🟢

The back of the form carries "a list of some major defects that may occur in used vehicles." It becomes our walk-around below.

Private sellers, online sellers, and the cooling-off myth

The Used Car Rule applies to dealers, defined in the regulation as a person or business that sells or offers a used vehicle "after selling or offering for sale five (5) or more used vehicles in the previous twelve months", excluding banks, businesses selling to their own employees and some lessors. The FTC's Dealer's Guide describes it as "more than five used vehicles in a 12-month period"; the regulation is the text that binds. Either way, a private owner selling their own car is outside the rule: no Buyers Guide, no as-is box, no warranty disclosure. 🟢

The rule has a geography too: "The Used Car Rule applies in all states except Maine and Wisconsin", which have similar rules of their own, and it applies in the District of Columbia, Puerto Rico, Guam, the U.S. Virgin Islands and American Samoa. Buying online does not remove the Guide; the FTC's Buying a Used Car booklet says: "You’re also entitled to a Buyers Guide if you buy a used car online." A 2024 FTC alert says an online seller "broke the law when it failed to give people required Buyers Guides, misrepresented vehicle inspections", and that for online sales dealers must offer a cancellation and prompt refund when a car does not ship on time. 🟢

There is no federal three-day return. The FTC's exact sentence: "Federal law doesn’t require dealers to give you three days to cancel the deal and return the car." Some states require a right to cancel; elsewhere a return window exists only if the dealer offers one. Get any return policy in writing before you sign. On a private sale, assume there is no way back once money and title change hands. 🟢

The odometer: what the seller must sign, and which cars are exempt

Federal law puts the mileage in writing. Under 49 CFR 580.5, the seller must disclose the mileage on the title (a seller in whose name the car is titled must use the title itself), sign it, with the odometer reading "(not to include tenths of miles)", the date, both parties' names and addresses, and the vehicle's identity including VIN. The seller then either certifies that the reading reflects the actual mileage to the best of their knowledge or, knowing otherwise, states that it exceeds the odometer's mechanical limit or does not reflect the actual mileage "and should not be relied upon", with a warning of the discrepancy. 🟢

Which cars are exempt in 2026. The exemptions in 49 CFR 580.17 split at the 2011 model year. A vehicle "manufactured in or before the 2010 model year" is exempt once it is transferred at least 10 years after January 1 of its model year, which in 2026 covers every 2010 and older vehicle. A vehicle "manufactured in or after the 2011 model year" stays covered for 20 years; the regulation's own example says that during 2031, model year 2011 and older vehicles are exempt. So in 2026 a 2011 or newer car needs the federal odometer statement and a 2010 or older one does not (vehicles over 16,000 pounds GVWR are exempt too). Your state's title form may still ask; the exemption is from the federal disclosure only. 🟢

Odometer tampering is prohibited by federal statute. The statute, 49 U.S.C. 32703, prohibits a person from acting to "disconnect, reset, alter, or have disconnected, reset, or altered, an odometer of a motor vehicle intending to change the mileage registered by the odometer." 🟢

Cross-check the number against the latest NMVTIS reading (a dashboard number lower than an earlier reported one needs an explanation), service records and inspection stickers, and visible wear.

The walk-around, using the FTC's own list of major defects

The back of the federal Buyers Guide carries a list headed "Here is a list of some major defects that may occur in used vehicles." Written for dealers' disclosure, it doubles as the best public walk-around list we found: short, organized by system, printed by the regulator. We abridged it, reordered it into walking order and kept the FTC's wording where it matters.

Flood signs, from the FTC. The FTC's flood alerts give four questions: "Is there mud or sand under the seats or dashboard? Is there rust around the doors? Is the carpet loose, stained, or mismatched? Do you smell mold or decay — or an odor of strong cleaning products — in the car or trunk?" Look under the trunk floor too. The same alert says a mechanic can check for water damage "that can slowly destroy mechanical and electrical systems and cause rust and corrosion." 🟢

Tire age, from the sidewall. Tread depth is on the FTC list; age is not, but every tire carries it. Under 49 CFR 574.5, the tire identification number ends in a four-digit date code, and "The date code must identify the week and year of manufacture": the first two digits are the week, the last two the year. 🟢 Four very different dates tell you something about how the car was kept (our reading); we quote no age limit because the federal rule sets none.

The VIN, three times. Before you leave the walk-around, compare the VIN through the windshield, the VIN on the certification label (under 49 CFR 567.4 it sits on the hinge pillar, door-latch post or door edge next to the driver's seat, or failing that the left of the instrument panel or the door's inner face, and it must include the VIN), and the VIN on the title. They should match character for character; NMVTIS's consumer page describes a South Florida ring that replaced stolen cars' VINs with VINs from other vehicles of the same make, model and year.

The walk-around is triage, not diagnosis: a frame weld or a flood smell ends the visit, a worn tire goes into the price, and anything unexplained goes on the list for a mechanic.

The cold start: what to watch in the first minute

Ask the seller not to start the car before you arrive. That request is ours, not a regulation's: a cold start lets you hear the engine before anything has warmed up, and if the engine is already warm, ask why.

Key on, engine off: the check engine light should come on. The EPA's 2001 guidance says "The MIL should illuminate (on some vehicles, only for a brief period of time)", so watch closely. The check is written into the EPA's emissions-warranty OBD test procedure: under 40 CFR 85.2222(d)(4), a car fails the OBD inspection "If the MIL does not illuminate at all when the vehicle is in the key-on/engine-off condition" (the MIL is the check engine light), even with no codes stored. California's OBD test reference lists the four cases: lamp lights with key on and engine off (pass), goes out once the engine runs (pass), fails to light with key on (fail), stays lit with the engine running (fail). 🟢 In our reading, a lamp that never lights may be a dead bulb, a disconnected lamp, or something somebody preferred you not to see.

Start it and watch the rest of the dashboard. The FTC's defect list names "Gauges or warning devices" among inoperable accessories, and its brake section lists a "Failure warning light broken". Ask about every lamp that stays lit; the OBD2 check later reads only the engine side. 🟢

Listen and look. The FTC list includes "Knocks or misses related to camshaft lifters and push rods" and "Abnormal exhaust discharge". 🟢 Outdoors, stand beside, not behind, the tailpipe for the first minute and note what you see or smell; we do not interpret smoke colors, because no government source we read does. Listen at the engine bay for ticks, knocks and whines that change with engine speed, keeping hands and clothing clear of belts and fans.

What "warmed up" means. The OBD counters below count warm-up cycles, and California's Bureau of Automotive Repair defines one: "A warm-up cycle means driving a vehicle so that the engine coolant temperature rises by at least 40 degrees Fahrenheit after the engine is started and reaches at least 160 degrees Fahrenheit." The 2006 amendment text of the OBD II regulation sets the diesel figure at 140 degrees. 🟢 In our reading, a car that has only been idled briefly may not have completed a single one.

The test drive, system by system

Drive only if the walk-around found no bald or damaged tires and the brake pedal feels firm, and only if you are insured to drive it. Start in an empty parking lot with gentle stops, then one firm stop at low speed, and full-lock turns; then, if the seller agrees, a road and highway run. Radio off.

Write down when, not just what. Braking-only versus highway-only, engine speed versus road speed: a mechanic will ask exactly this.

After the drive, park with the transmission in Park or neutral and the parking brake set, leave the engine running and look at the ground under the car from outside (do not reach or crawl under a running car): a drip that was not there before is the most honest thing the car will tell you all day. Then plug in the OBD2 reader, engine still running. The drive has warmed the car, which helps some self-tests finish, though monitors that need specific conditions may still show not ready.

The centerpiece: an OBD2 check for codes that were just cleared

The situation this section exists for: a car has a fault that lights the check engine lamp. Somebody clears the codes with a scan tool, or disconnects the battery. The lamp goes out, and for a fault whose self-test needs particular driving conditions it can stay out through a short test drive, because the car has not yet re-run the self-test that found the fault (our reading of how the monitors below work). No special equipment is needed; clearing codes is a standard function of consumer readers, and the OBDLink and FIXD pages we read both advertise it. ⚪

The EPA named this in its June 2001 guidance to state inspection programs, EPA420-R-01-015. Among five reasons a car can arrive with its self-tests incomplete, it lists "A fraudulent attempt to avoid I/M program requirements by clearing OBD codes just prior to OBD-I/M testing (by, for example, temporarily disconnecting the battery)." The other four are not fraud: not enough of the right kind of driving, a recent battery disconnection or replacement or routine maintenance, a vehicle-specific OBD fault, or a design anomaly. That is the honest framing for a buyer too. A cleared-code signature proves a recent reset, not a motive. It is a reason to ask and to inspect further, not an accusation. 🟢

The same regulators built the fingerprints into OBD II cars: readiness monitors from the start, the two counters on newer cars, and permanent codes phased in over the 2010 to 2012 model years. The California Air Resources Board's OBD II fact sheet says "All 1996 and newer model year gasoline and alternate fuel passenger cars and trucks are required to have OBD II systems", with diesels from 1997, and that the U.S. EPA sets OBD requirements for 1996 and newer cars sold in any state. California's OBD II regulation (13 CCR 1968.2, 2006 amendment text) puts the connector in the driver's side foot-well region, "no higher than the bottom of the steering wheel when in the lowest adjustable position." 🟢

Three fingerprints follow (the first two reset together; the third survives a clear): readiness monitors, the warm-up and distance counters, and permanent codes.

How can you tell on the driveway whether a seller cleared the check engine codes, and can PulsCar do it? Plug an OBD2 reader into the port under the dashboard after your test drive and read three things. First, the readiness monitors: after a code clear or a battery disconnect they show "not ready" until the car's self-tests run again, and the EPA's OBD test sends any unset monitor back to drive more, except one on a 2001 or newer car. Second, Mode 01 PID $30 and PID $31, the warm-ups and distance since codes were cleared; numbers far below California's 15 warm-ups and 200 miles on a daily driver point to a recent reset. Third, permanent codes, read with Mode $0A, which a scan tool cannot erase on 2012 and newer cars. PulsCar's free sound check does none of this; it listens and reads no codes. PulsCar's app works with the OBDLink MX+, but any reader showing these three screens will do.

Fingerprint one: readiness monitors that say "not ready"

An OBD II car runs self-tests on its emission systems and records which have finished since its memory was last cleared. The EPA's 2001 guidance describes the system as monitoring "up to 11 emission control related subsystems". Three run all the time (misfire, fuel trim and comprehensive components); the other eight run only when their conditions are met, and the EPA names the common five as catalyst, evaporative system, oxygen sensor, heated oxygen sensor and EGR. Each monitor reports as "ready", "not ready" (meaning the monitor has not yet been evaluated) or "not applicable". 🟢

The reason monitors catch a clear is in California's inspection manual, the OBD test reference: "Readiness monitors must be rerun after a repair activity like disconnecting a vehicle’s battery or replacing a defective emissions component." Clearing codes has the same effect. A 2013 EPA memo transmitting a workgroup's best-practices paper on diesel readiness (EPA says it is not official guidance) recommends that diesel drivers operate vehicles "for two or three weeks after codes are cleared to provide for readiness monitor setting", against the one week often used for gasoline cars, and notes that many newer diesels may struggle to meet the one-unset-monitor limit "if codes were recently cleared". 🟢 So a car whose codes were cleared this morning is likely to show monitors not ready this afternoon, however good it feels on the road.

How many "not ready" is normal? The regulators give two useful yardsticks, and they differ.

The EPA numbers are the exception in 40 CFR 85.2222(c)(2) to its default in (c)(1), under which any incomplete monitor sends the customer away to drive and return, and a second incomplete reading is a fail; California's come from Table 1 of its OBD test reference, implemented May 4, 2015; California counts propane and natural gas as "gas". BAR's 2018 workshop slide words the 2007-and-newer diesel rule differently (any incomplete monitor except the particulate filter fails); we follow the test reference, BAR's inspection manual. California ignores the continuous monitors on gas vehicles, and the EPA's 2001 guidance recommends that programs disregard them, explaining that a small number of cars flag them "not ready" through scanner-software incompatibility, which "is not indicative of a fault with the vehicle’s OBD system." 🟢 So count only the periodic monitors (catalyst, EVAP, oxygen sensor, oxygen sensor heater, EGR and similar) and ignore a "not ready" misfire, fuel system or components line.

Reading it as a buyer. One incomplete monitor on a car that is otherwise clean is normal; the evaporative monitor that California exempts needs conditions such as fuel level and ambient temperature (BAR's 2018 workshop), and BAR's test reference says "other vehicles have more trouble either because the vehicle design requires unusual operating conditions or because the owner rarely drives the car in the necessary operating mode." 🟢 In our reading, two or more periodic monitors not ready on a 2001 or newer car (three or more on a 1996-2000 car) that the seller says is driven every day is consistent with a recent reset, though a faulty component can also keep a monitor from completing. It does not tell you why. The next two fingerprints tell you when.

Fingerprint two: warm-ups and distance since codes were cleared

This is the most precise of the three. Many OBD II cars keep two running counters that start again from zero when the codes are cleared. In the SAE J1979 numbering that inspection systems and consumer apps use, they sit in Mode (Service) $01:

The first two rows come from a 2018 Bureau of Automotive Repair workshop on permanent codes, whose slide says the counters are "Supported by most 2006+ diesel and most 2008+ gasoline vehicles"; a later slide puts majority coverage at 2006 for diesel and 2007 for gasoline. The third row is quoted from BAR's OBD test reference. The regulatory basis is the 2006 amendment text of California's OBD II regulation, 13 CCR 1968.2, which requires 2005 and subsequent model year vehicles using the ISO 15765-4 (CAN) protocol to make available the "distance traveled since fault memory last cleared, and number of warm-up cycles since fault memory last cleared." We did not read SAE J1979 itself; we cite the public documents that use its numbering. 🟢

California uses them as a clock, and so does the 2013 workgroup paper. Since July 1, 2019, California's OBD test reference says a permanent code "will be ignored if the vehicle has completed at least 15 warm-up cycles and been driven at least 200 miles since its OBD information was last cleared (as determined by Mode $01 PID $30 and $31)." The 2013 workgroup paper uses more than 500 miles and more than thirty warm-ups since the last clear in one of its decision paths for diesels. 🟢

How to read them on a driveway. The workgroup paper states its own assumptions, and they give a buyer a rough ruler: 30 warm-up cycles "is based on an assumed average of 2 warm-up cycles per day or 14 per week", and its mileage target is based on an estimated annual 12,000 miles, "about 500 miles in two weeks". 🟢 Those are the workgroup's planning averages for light and medium-duty diesels, not a rule for your car, but the arithmetic helps; the yardstick that follows is ours. A car the seller says is driven to work every day that shows 3 warm-ups and 11 miles since codes were cleared points to a reset within the last few days. Warm-ups in the high double digits and hundreds or thousands of miles since the last clear point away from a recent clear.

Three cautions from the same documents. First, the workgroup paper: "Temporary disconnection of the battery may cause these counters to reset to zero." A new battery is an innocent reason for low numbers; ask for the receipt. Second, "Vehicles with multiple controllers may report different values", and the workgroup recommends using the minimum reported; if your reader shows two values, use the lower. 🟢 Third, not every car supports the counters; blank PID $30 and $31 on an older car tells you about the car, not the seller. Our tip: check whether your app shows PID $31 in kilometers or miles (200 miles is about 320 km).

Fingerprint three: permanent codes that a clear cannot erase

The last fingerprint was designed specifically to defeat code clearing. A BAR workshop slide, citing SAE J1979, gives the purpose: "Stated purpose is to prevent DTC clearing to obtain a passing test". A permanent diagnostic trouble code (PDTC) is written when the check engine lamp is commanded on, stored in non-volatile memory, and, in BAR's words, "Cannot be erased by DTC code clearing or battery disconnect". 🟢

The workgroup paper attached to the same EPA memo gives the mechanics and the dates, citing California's 1968.2 and the EPA's 40 CFR Part 86. Permanent codes "cannot be erased with scan tools (generic, aftermarket, or manufacturer-specific) or by disconnecting the battery, but instead can only be erased by the OBD system itself." If nobody cleared the codes, it disappears when the lamp turns off on its own (the paper's example: the fault absent on three consecutive trips). If somebody did clear them, "the Permanent-DTC will not be erased until the monitor responsible for setting that specific Permanent-DTC has run at least once and confirmed that the fault is no longer present." They are read with a separate request, "J1979 Mode/Service $0A", so a reader can tell them apart from ordinary codes. And they were phased in from the 2010 model year: "Fifty percent of the fleet was required to be Permanent-DTC compliant in MY 2010; 75% in MY 2011, and 100% in MY 2012." 🟢

What that means for the buyer: on a 2012 or newer car, a permanent code with no ordinary code and no lamp is the clearest single sign on this page that a fault was active and the codes were cleared before the car's own self-test re-checked it. One exception, from the same California test reference: its Table 6 lists specific models and codes that are not self-clearing, among them U0140 on 2012-2015 Dodge Challenger, Charger, Dart and Durango, P0315 on 2011-2014 Ford Mustang and C2200 or B22A9 on 2017-2019 Chrysler Pacifica Hybrid, and more models besides; on those, the listed code alone does not prove a recent clear, while other codes still count. It names the fault, too, because a permanent code is the same code number as the ordinary one it shadows.

California's first month of permanent-code testing, the closest public analogue to a car presented for sale, shows this matters. BAR's fall 2019 newsletter reports that during the first month, "3,124 vehicles failed for PDTCs that would have previously passed a Smog Check inspection." It adds: "These vehicles all had recent codes cleared and a PDTC stored in the computer’s memory." The failure rate for 2010 and newer vehicles rose from 4.2% to 5.6%. When BAR checked later, 2,385 of the 3,124 had passed a re-inspection, 1,974 of them because the car's own system had cleared the permanent code, and 411 because they had reached the 15 warm-up and 200-mile limit. 🟢 That is not a claim that those owners hid anything; many may have cleared a code after a repair.

Reading the result: four patterns and what each one means

Put the three fingerprints together with the code read and the lamp check, and most cars fall into one of four patterns. The interpretations are ours, built from the rules above.

Pattern A is the best a driveway can give you, and it still means only that the engine computer has nothing to report; everything it does not monitor is unexamined.

Pattern B needs a question and a receipt. Ask the seller directly whether the battery was replaced or disconnected, or the codes cleared, and when. A dated battery, repair or dealer reconditioning record that matches the counters is a normal answer; the EPA lists routine maintenance among innocent causes. Mixed results that fit no pattern get the same treatment. Otherwise, walk, come back after a week of driving (the EPA's outreach advice after repairs is to "allow for approximately one week of normal vehicle operation" before testing), or pay for a professional inspection. 🟢

Pattern C is the strongest evidence of a hidden fault this check can produce. Look the code up; if it is on the list of the most expensive check engine light codes, that tells you what is at stake before you negotiate. If the fault was repaired, the seller should be able to show the repair, and (outside the Table 6 models) the permanent code will clear itself once the monitor has run.

Pattern D is not a used-car verdict, it is a price. A stored code is a fault the car has confirmed. Note that under 40 CFR 85.2222(d)(2) a car "shall pass the OBD inspection, even if DTCs are present" when the lamp is not commanded on: a dark lamp does not mean no codes, so always run the code read. What a light and its code usually mean is covered in what to do when the check engine light is on. 🟢

Which readers show these screens, in the makers' own words

Readiness is reported through the standardized Mode $01 PID $01 request, and the EPA's procedure requires the test system to be "capable of checking for OBD monitors and the evaluation status of supported monitors" (40 CFR 85.2231). 🟢 What varies is whether a consumer app puts the counters on screen. We checked three makers' own pages on 30 September 2026; this is not a ranking.

OBDLink's app page adds the caveat that applies to every reader: "Your OBD2 compliant vehicle most likely supports a subset" of the listed PIDs. None promises to label a clear as a clear; you draw the conclusion with the four patterns above. For the consumer tools compared, see the diagnostic app comparison and FIXD versus PulsCar. ⚪

Check your reader before you rely on it. If it has no permanent-code (Mode $0A) screen and does not show warm-ups and distance since codes cleared, it will show nothing rather than "none found", and you have not done fingerprints two and three. No reader, or a basic one? Ask the mechanic who does your pre-purchase inspection to include those screens in the written report.

What an OBD2 check cannot see

The fingerprints above are narrow by design. CARB's fact sheet defines the system's job: OBD capabilities "monitor virtually every component that can affect emission performance." 🟢 That is close to comprehensive for emission-related components, and silent on much of what a used-car buyer worries about.

PulsCar counted the vocabulary of that silence earlier this year, and we cite the finding rather than re-derive it. In the FIXD versus PulsCar analysis, we counted the vocabulary in 39,355 NHTSA owner complaints for ten high-volume nameplates, model years 2015 to 2022. Of the complaints in which an owner described a sound, 79% mentioned no warning light, no code and no scan; only 1,009 mentioned both. That is a count of words, not faults, and the analysis states its limits, but the direction is clear: much of what owners notice first is something a code reader was never asked to see.

Items on the FTC's defect list that, in our reading, a generic OBD2 scan is not designed to report include frame damage, rust, worn tires and cracked wheels, leaking shocks and damaged bushings, brake thickness, belts and exhaust leaks. That is why a clean scan is never the end of the inspection. More in sound versus scanner diagnosis.

When to pay for a professional pre-purchase inspection

The FTC does not treat the professional inspection as extra caution. Its consumer page: "it’s always a good idea to pay for an independent mechanic to conduct a mechanical inspection of a used car." It closes the loophole people reach for first: "A mechanical inspection is a good idea even if the car has been “certified” and inspected by the dealer, and is being sold with a warranty or service contract." And it separates this from a safety inspection: "A mechanical inspection is different from a safety inspection. Safety inspections usually only focus on things that make a car unsafe to drive." 🟢

What to ask for is short and worth copying: "Ask the mechanic for a written report with a cost estimate for all necessary repairs." The FTC adds "Be sure the report includes the car’s make, model, and VIN", and notes that the estimate doubles as a negotiating tool. If a dealer will not let the car off the lot, the FTC suggests a mobile inspection service or asking the dealer to bring the car to a facility you choose, and then: "If a dealer won’t allow an independent inspection, consider going to another dealer." 🟢 A private seller has no Buyers Guide, but the logic is the same.

This page deliberately does not price the visit: our car inspection cost guide covers pre-purchase inspection prices, shop versus mobile, and how to find an honest mechanic covers choosing one independent of the seller.

When is a professional pre-purchase inspection worth paying for, and where does PulsCar stop? Pay for one whenever the driveway checklist turns up something it cannot explain: monitors not ready with low counters, a permanent code, a leak, a weld, a noise or a shift you cannot name, a branded title, or simply a price high enough that a hidden fault would hurt. The FTC's advice is broader still: it calls an independent mechanical inspection always a good idea, even on a certified car sold with a warranty, and advises asking for a written report with repair estimates and the car's VIN. PulsCar stops well before that point. Its free sound check can give a second opinion on a thirty-second engine recording made on the driveway, and it never sees the frame, the brakes, the codes or the paperwork. Use it, if at all, as one more input into whether this car deserves a mechanic's time, never as a substitute for the mechanic's inspection.

Known problems by model: read before you visit

A model guide tells you what this engine, gearbox or electrical system tends to do, and at what mileage. If your car is below, check the known problems for this model before the visit, so you can ask about a known noise, leak or code by name. Each guide lists its own sources.

Use a guide as a list of things to rule in or out, not a prediction for this car: a documented weakness at a certain mileage is a reason to ask for the repair record around that mileage.

The traps: three ways a careful buyer still gets caught

Trap one: trusting a dark check engine light. The situation: the lamp is off, the seller says the car has "no codes", and a quick scan agrees. What is real: a clear turns the lamp off and erases stored codes in one step, and the EPA lists clearing codes just before a test among the reasons cars arrive with unset monitors. The evidence survives in the monitors, the counters and, on 2012 and newer cars, permanent codes. The price: buying a confirmed fault at a no-fault price. The defense: never stop at "no codes"; read the monitors, PID $30 and $31, and Mode $0A before you negotiate.

Trap two: treating "certified" or a clean history report as an inspection. The situation: a certified car, a clean history report, and an inspection that feels redundant. What is real: the FTC says a mechanical inspection is a good idea even for a certified car sold with a warranty or service contract, that a history report "is not a substitute for an independent vehicle inspection", and NMVTIS warns that a search may return a false negative. The price: a mechanical fault that no database was ever going to record. The defense: use the report and the certification to decide what to inspect, not whether to.

Trap three: assuming you can take it back. The situation: a buyer signs, finds a problem at home and plans to return the car. What is real: federal law does not require dealers to give three days to cancel, some states do and elsewhere only the dealer's own policy applies, an "As Is" Guide means you assumed the risk after the sale, and a private sale has no Guide at all. The price: the full repair, at your expense. The defense: do every check on this page, and get any return promise written onto the Buyers Guide before you sign.

Three composite situations, decoded

These are composite scenarios, not individual customers. They are built only from facts already on this page, written in the conditional, and contain no invented names, amounts or dates.

Scenario 1: the clean-looking commuter car. Suppose a sedan from a private seller has no lamp, no codes, a clean title report and a pleasant test drive, and the seller says it is driven to work every day. The reader shows three periodic monitors not ready and single-digit warm-ups since codes were cleared, and no permanent code. That is Pattern B. The honest next move is a question about the battery and recent repairs, then a week's delay or a pre-purchase inspection. Lesson: low counters on a daily driver are a date stamp, and a date stamp deserves an explanation.

Scenario 2: the dealer car with a permanent code. Suppose an SUV from the 2012 or a later model year on a dealer lot with an "As Is" Buyers Guide shows no lamp and no ordinary codes, but the reader finds one permanent code. Permanent-code support was required on 2012 and newer light-duty models, and outside the Table 6 models a permanent code without an ordinary one usually means a fault was active when the codes were cleared and the self-test has not yet re-checked it. The buyer looks the code up and asks for the repair record; with none, they ask for an independent inspection, as the Buyers Guide invites. Lesson: on an as-is car, the permanent code is the one sign a scan-tool clear cannot erase.

Scenario 3: the older car with nothing to read. Suppose a gasoline hatchback from the mid-2000s shows no lamp problems, all monitors ready, and no data at all for PID $30 and $31. BAR's slides put majority counter support on gasoline cars at 2007 or 2008, and permanent codes arrived in 2010, so the reader has done what it can. The decision rests on the walk-around, the test drive and the service records, since a 2010 or older car also needs no federal odometer statement. Lesson: an older car shifts the weight of the inspection from the computer back to the eyes and the mechanic.

Where a free sound check fits, and where it does not

This is our side of the page, so check it against the homepage. PulsCar's sound check asks you to film the engine running for 30 seconds; the homepage says two independent AI engines return a free health report in 10 minutes, on the terms "No payment, no card, no subscription." The emailed report carries an Engine Health Score, an AI second opinion and system-by-system status, and says "inconclusive" when a clip cannot support a verdict. ⚪

It cannot read codes, monitors, counters or permanent codes, or see frame, rust, tires, leaks or paperwork. It is built around engine sounds; the homepage also lists belts, exhaust, starting and charging, suspension and drivetrain sounds, but not brakes. It gives no verdict on whether the car is safe to drive or worth the price.

Should you record the engine on the seller's driveway, and what does PulsCar's free check actually add? Record it if the seller agrees: thirty seconds of the engine running, ideally from the cold start this checklist already asks you to watch, filmed on your phone near the engine bay, is the whole job, and the report arrives by email in about ten minutes. What PulsCar adds is an acoustic second opinion on the one thing this page cannot measure with a reader: how the engine sounds. It adds nothing else. It reads no codes, monitors or permanent codes, it sees no frame, brakes or paperwork, and it may come back inconclusive, in which case record again or rely on the other checks. Treat a worrying result as a reason for a professional inspection and a reassuring one as a single data point, never as permission to skip the rest of this list.

What we looked for and did not find

Each gap below was an actual search or request on 30 September 2026.

NHTSA's own recall and odometer pages answered our automated requests with an access block (HTTP 403), and we did not try to get around it. The recall lookup is described in the FTC's and the Buyers Guide's words; the odometer rules are quoted from 49 CFR Part 580 and 49 U.S.C. 32703, not from NHTSA's blocked page. NHTSA's VIN decoder and recalls API did answer.

NICB's VIN check page was blocked too; its database appears only as the FTC describes it.

A federal list of states that limit "as is" sales. None on the FTC pages we read; the FTC sends buyers to their state attorney general. We did not build one from secondary sources.

The SAE J1979 standard. We did not read it. PIDs $30 and $31 appear in California's Bureau of Automotive Repair material, the EPA memo and OBDLink's published table, and agree; each other identifier comes from the one source named beside it.

A statistic on how often sellers clear codes before a sale. We found no regulator that publishes one. A search of our copy of NHTSA's complaints database (780,071 complaints received 1 January 2015 to 21 September 2026) for narratives pairing a code clear or reset with a used-car purchase returned 165 candidate matches, which would need two independent readers to classify before any count could be reported, so we report none. California's permanent-code figures describe cars presented for a Smog Check, not for sale, and we say so where we use them.

Two small inconsistencies in the FTC's own texts. The regulation's dealer threshold is five or more used vehicles in twelve months; the Dealer's Guide says "more than five". FTC pages also call NMVTIS information both free and fee-based. We report both.

Your action plan: at home, on the driveway, before any money moves

At home, the evening before:

1. Get the 17-character VIN in writing and run it through NHTSA's VIN decoder; check year, body and engine against the listing.
2. Check open recalls with the VIN at safercar.gov, the NHTSA address printed on the Buyers Guide; if the seller says a recall is open, ask them to have it done, or ask a dealer of that brand whether the repair is available.
3. Pull an NMVTIS report from a provider listed at vehiclehistory.gov: brands, insurer total losses, latest reported odometer reading.
4. Read the PulsCar model guide if there is one, and note the known problems and their mileages.
5. Ask the seller not to start the car before you arrive, and to have the title and service records ready.

On the driveway: 6. Dealer sale: read both sides of the Buyers Guide and photograph it. 7. Match the VIN on the windshield, the door-frame label and the title, and (our advice) check that the name on the title matches the seller's ID and whether a lender is listed; your state DMV's site says what a valid transfer needs. 8. Walk around with the FTC defect list, plus flood signs and tire date codes. 9. Key on, engine off: the check engine lamp should light (sometimes only briefly), then go out once the engine runs. Record the first thirty seconds if you want an acoustic second opinion. 10. Test drive with the radio off: straight-line braking, steering play, full-lock turns, every gear. Look under the car afterwards. 11. Ask before you plug in; then, engine running, read codes, readiness monitors, PID $30 and $31 and permanent codes. Screenshot every screen.

Before any money moves: 12. Anything failed or unexplained: book an independent pre-purchase inspection with a written report naming the make, model and VIN (what it costs). 13. Put every promise in writing, on the Buyers Guide for a dealer sale. There is no federal three-day return. 14. On a 2011 or newer model year, read the odometer statement before you sign.

For the neighbors of this page: what a diagnostic should cost, the check engine light is on, sound versus scanner diagnosis and the AI sound tools compared.

Sources and documents

Each page below was opened and saved as text on 30 September 2026.

1. FTC, Used Car Rule — Buyers Guide duties, state-law override, dealer definition: 16 CFR Part 455
2. FTC, Buyers Guide form (English) — the printed boxes and the defect list: cfr_buyers_guides_english.pdf
3. FTC, Buying a Used Car From a Dealer: consumer.ftc.gov, and its Buying a Used Car booklet
4. FTC, Dealer's Guide to the Used Car Rule: ftc.gov
5. FTC, Used Cars resource page — NMVTIS, recalls, NICB: feature-0040-used-cars
6. FTC alerts — flood 2022, flood 2021, buying online 2024
7. Department of Justice, NMVTIS — consumer page, approved providers
8. NHTSA, Odometer Disclosure Requirements: 49 CFR Part 580
9. Odometer tampering statute: 49 U.S.C. 32703
10. NHTSA, VIN, certification label and tire date code: 49 CFR Part 565, 49 CFR 567.4, 49 CFR 574.5
11. NHTSA VIN decoder: vpic.nhtsa.dot.gov/decoder
12. EPA, OBD test procedure (performance warranty short test) and readiness exceptions: 40 CFR Part 85, Subpart W
13. EPA, OBD checks in I/M programs, June 2001: EPA420-R-01-015
14. EPA memo transmitting a workgroup's diesel OBD readiness best practices (not official EPA guidance), March 2013 — permanent codes, Mode $0A, PID $30 and $31: EPA memo
15. EPA, I/M guidance index: epa.gov
16. California Air Resources Board — OBD II fact sheet, 13 CCR 1968.2, 2006 amendment text
17. California Bureau of Automotive Repair — OBD test reference, permanent-code workshop, 2018, ARSC News, Fall 2019
18. Makers' pages — OBDLink app, OBDLink MX+, BlueDriver, FIXD FAQ
19. PulsCar — FIXD versus PulsCar (the 39,355-complaint count), homepage (sound check terms)
20. Attempted, access blocked (HTTP 403), not quoted: nhtsa.gov/recalls, safercar.gov, NHTSA odometer fraud page, NICB VINCheck

How this page was built: every legal statement was read on the issuing body's site, or the eCFR and govinfo editions of federal text, on 30 September 2026, and quoted as published; state rules are California's own documents, labelled as such. Scan-tool features are makers' descriptions of their own products, marked ⚪, not a ranking. The 39,355-complaint figure is PulsCar's earlier published count, cited, not recomputed. Time estimates and the four reading patterns are our interpretation. The scenarios are composite and contain no invented names, amounts or dates.

Prices verified: September 2026 (this page quotes no inspection prices; see the car inspection cost guide) · Technical documents verified: September 2026 · Reviewed quarterly

Saw monitors, counters or permanent codes behave differently from this page? Email [email protected] with the make, model year and what your reader showed.