Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Automotive What causes a car to be a "lemon"?

  1. Mar 11, 2017 #1
    we all know that some cars are lemons and some are great.
    but sometimes we can get 2 identical cars, bought, driven and serviced the same, but 1 is great and the other will be a lemon.
    how this happens sometimes baffles me. does anyone have any plausible theories?
  2. jcsd
  3. Mar 11, 2017 #2
    Oh, boy...a chance to rant.

    In a word: "humans." The best Six Sigma efforts and automotive quality programs are only as good as the humans that interact with them. Do a little research on Six Sigma indices (e.g., "how many defects occur if you had 99.9% quality?") and it becomes clear as to the "why...?"

    The other part of this "why...?" is determined by the field of Reliability Engineering in which such measures such as Mean Time Between Failure MTBF is determined.

    And keep in mind: all of those quality indices and MTBFs apply to each of the 1000's of components in a vehicle.

    Case History1:
    I purchased a 1986 VW GTI. What a fabulous vehicle. Satisfied my need for speed, was completely practical, fun to drive, ultra-reliable, suited my personality perfectly. Just the best dang car I've ever owned. Drove it 20+ years. I quit replacing the odometer after the 2nd unit failed, so I can only estimate the final mileage was somewhere north of 250K miles. It become difficult to find parts for it (even from the VW Stealership). It finally had a brain aneurism in a rain storm in the middle of a busy intersection. It never ran again, and a friend bought the carcass from me for pennies. I had gotten my money out of it. Still dream about that car.

    I needed a new car, one suitable for long commute. I found a used 2001 VW diesel Bug. At first, it was great. First experience with a diesel, fell in love with it. Then the quality problems started happening. Deep subsequent research revealed that VW hired a charlatan to run their Purchasing Department during this time this series of VWs were built. He was effective in reducing costs, but drove suppliers into the dirt so they cut corners. Initial quality was OK, but not robust and things would fail later. The Stealerships are famous for simply doing a Remove & Replace strategy, repeating this until the problem is finally resolved. Cost = thousands, sometimes more than the vehicle is worth. I read that my vintage of VW New Beetle had 48 known serious quality problems. I personally can account for 45 on my Bug. Example: on hot days, while driving, the alarm system will spontaneously self-arm. When I stop the vehicle, I must press the Open-Door button on my keyfob else I will initiate the security alarm by removing the keyfob from the ignition. Root cause has been found to be poor solder joints in the electronic door latch module, causing cracks. Temperature rises, solder cracks expand, false signals happen. Dealer can fix it by replacing the door latch modules (~$600-$700 each) or I can remove door panels & latches, re-solder joints. Or just put up with it. And the list goes on. With my personal experience, and the most recent diesel shenanigans, VW has lost me as a customer forever.

    Case History2:
    My wife owned a 2000 Chevy Malibu. Insufficient testing was done on the engine. At Warranty-Expiration-Mileage the intake manifold sealant (note: not "seal") started failing. On most, but not all, vehicles. Coolant slowly leaked into the engine compartment. Required a $600 fix. Not a big fan of GM, either.

    My manufacturers of choice now are Hyundai and Subaru. So far, so good: no lemons.
  4. Mar 11, 2017 #3

    jim mcnamara

    User Avatar

    Staff: Mentor

    You are using the word 'theory', @jfoldbar, in the sense of: I just pulled this out thin air. @tygerdawg is mentioning very well tested concepts. Engineers and scientists use the word theory in a different way. It confuses people. 'It means I know this stuff and how it got this way. If it needs tweaking I listen and test until my hypothesis works great in the real world.'

    What he is talking about: precise control of procedures, materials, constant measurements to make sure parts are in tolerance, and quality tests on assemblies, sub-assemblies and completed cars.

    This whole thing is a balancing act on the part of the manufacturer. An example:
    Do I make parts locally and have complete control over quality, at a greater price to me. Do I subcontract out locally and have less control?
    As control of parts manufacturing goes down, there can be a cost penalty for testing somebody else's parts. They do not test and measure every part, just some. How often to test and how evaluate results are mathematically determined -- and way beyond what is needed here.

    This falls under the topic of Reliability Engineering - lots of statistics and models of testing regimes. One goal is called 6 sigma. That is statistics jargon for 'a set of management techniques intended to improve business processes by greatly reducing the probability that an error or defect will occur.' Management speak, if you like. Clearly it does not always prevent problems. He mentions VW as a mjor corporate failure. VM circumvented vehicle emissions testing. They programmed their car computers to behave differently in terms of ignition and fuel consumption when tested. What the cars really do on the road is to dump out lots more pollutants and consume more gas. VW saved big bucks and broke a bunch of laws. His conclusion - and mine: So much for 6 sigma certification.

    A measurement of product quality is something called mean time before failure - MTBF. Lets take a part of a car: an axle bearing . They create a model of how a car is driven. Not everyone drives the same, so that is taken into account. Next they figure out how many revolutions the axle bearing has to endure, how many panic stops, speed bumps, and everything else. They translate that into a driving time in hours- the point at which 50 out of 100 of the bearings have failed. MTBF. So a part will operate reliably and not fail for most people up to some point before MTBF. So manufacturers warrant (pretend* to pay for replacements) of parts before say 50000 miles of operation or five years - whichever comes first. They are betting that because the MTBF is really 7 years, it won't cost them too much money or bad publicity (which is worse for excessive failures) for the warranty - which as an estimate, is in the sale price anyway.

    Cars have computers now. Those computer parts and sensors have an MTBF. So do individual drive train parts, and the whole assembly does, too. Everything.

    So, a lemon is a car for which people have expectations of perceived quality, the quality violates that perception. It can be anything from transmission to dome light repeated failures that trigger the lemon response.
    1. made from parts for which reliability testing is poor by design or delusion.
    2. MTBF estimates for some parts have problems. Those often result in a recall.
    3. Like Toyota did, pretend there is no problem because admitting some is not as advertised or implied is "bad". Or they simply are dishonest. - this was/is a corporate culture problem.
    4. What VW did, in essence, produced large numbers of lemons. They violated consumer perceptions. Not to mention laws all over the world.

    *pretend in the sense that consumers pay up front (in the purchase price) for a part the expected failure repairs, whether or not they have any.
    Note: you can buy extended warranties, which gives you an idea of what car makers think the MTBF really is.
  5. Mar 11, 2017 #4
    mayby im on another planet or something. so i'll try to go into more detail with question.

    imagine i buy 2 identical toyota corollas. on the same day. they both have the same build date. came from the same factory. they both get driven by the same person and serviced at the same mechanic and are refueled with the same petrol.

    car A and car B are exactly the same

    now, these cars use many parts that were not made by toyota. perhaps the lights came from some factory in china. and the gearbox from mexico. the bearings from another japanese factory, and the alternator from another.

    car A, after 10000 kls, starts to have light problems that car B does not. this must be a problem with the chinese factory.

    then at 20000kls the gearbox develops a noise that car B doesnt. this is a problem with the mexico factory/quality. but that mexico factory sourced parts from russia for all of their gearboxes. under warranty toyota put a new gearbox. it has the same problem after 20000kls. car B still has no problem.

    at 17000kls car A devolops a brake problem. car B does not. the brakes need to me sent to the brake developer in thialand for evaluation.

    at 25000kls, car A windows all start to leak. car B does not. the rubber came from india factory. made in same machine on same day with same materials.

    do you see where im goin with this? or should i go on?

    how could coincidence line up so perfectly that the human error is a factor in every factory of the world simultaneously in the production of the parts for 1 car. theres more chance of winning lotto.
  6. Mar 12, 2017 #5


    User Avatar

    Staff: Mentor

    But that's the key issue here. The cars are not exactly the same. It is impossible to create two identical parts, let alone two identical cars. Sure they may appear to be identical, but there are always small differences which are the basis for failures. These small differences can be decreases with increasingly detailed testing, development, and manufacturing techniques, but they can never be brought down to zero.

    While human error certainly increases the failure rates for parts, it is not the root cause of most failures.
  7. Mar 12, 2017 #6
    """These small differences can be decreases with increasingly detailed testing"""

    what do you mean by this?
  8. Mar 12, 2017 #7
    Part and process variation. Not one of the cars that rolls off the line will be identical. The goal of quality is to control parts and processes to reduce failures.

    The way variaions stack up can cause a massive variety of anticipated and unanticipated failure modes.

    You immediately went to parts quality/ supplier fault in all your examples. Why?

    eg Gearbox Failure
    Car B is fine. Car A killed 2 boxes after the same distance. So you investigate and find that engine A is slightly misaligned due to toleranes on the engine mounts. This misalignment wore the input shaft of the gearbox.

    The chain of root cause can be extremely difficult to find.
    Last edited: Mar 12, 2017
  9. Mar 12, 2017 #8

    jim mcnamara

    User Avatar

    Staff: Mentor

    @xxChrisxx point is really clear. The root cause could be anything - unknown contaminants in one part that cause another part downstream in the circuit to go nuts. So, you keep replacing the wrong part - the one that failed. There is also a systemic effect. All of the parts do great on their own, but less great when they are all working together. Or in the sleet in Santa Fe County NM, where they use unique materials to melt ice on the roads (in fact, they do this), which later causes a corrosion/short/damage somewhere.Six months later.

    There is what they call a constellation of effects - a fancy term for an ungodly number of things that may upset a system. Some people like to call that Murphy's Law - 'If something can find a way to go wrong, it will go wrong'.

    I tried to simplify this mess in my first post - didn't work well. But @jfoldbar your assumptions are plain wrong. @Drakkith nailed it simply: no two cars are completely alike to start with. Not ever.
  10. Mar 12, 2017 #9
    What a loaded question. First off, I for one an grateful for the quality of the vehicles we have today. Wasn't long ago a car with 100k miles was pretty much wore out and almost unsellable. That's the mileage of the first tune-up now. Some of us remember how it was when planned obsolescence was a strategy.

    !00% QC testing of each and every part would improve quality and lower "lemon" rates but nothing will completely stop lemons from happening. Nobody wants to see what 100% testing would do to the price of vehicles.

    Then again maybe some drivers should quit riding the brakes and learn how to downshift. Or maybe don't park under sappy trees or leave the car in the sun all day or at least cover the window seals. If every seal went I'd first consider the QC of the cleaning product I used or quit thinking the car was made for drifting and stop racking the body.

    Just a thought.
  11. Mar 12, 2017 #10
    so the answer is,
    we dont know, but no 2 cars are ever exactly the same to begin with.
  12. Mar 12, 2017 #11


    User Avatar

    Staff: Mentor

    Pretty much.
  13. Mar 12, 2017 #12

    jack action

    User Avatar
    Science Advisor
    Gold Member

    In your example in post #5, I don't think it is very realistic unless you could swear it actually happened. Having a single car having many different problems that its "sister" cars don't have should be an extremely rare case.

    The most probable cause of such a case would be buying unknowingly a salvaged car.
    I don't think this could be true except, of course, for the extremely rare case.

    A more probable definition of a lemon is one where a single particular problem appears out of nowhere and it comes back again and again.

    This has 2 possible sources:
    1. The repair shop doesn't do the job well;
    2. There was an uncaught problem at the factory which is hard to find with the usual diagnosis methods known by the repair shop.
    Find examples of these here. Let me give you other real examples I witnessed.

    Once, I was working as a mechanic in a dealership. There was a recent car that was taken apart when I arrived, covered in dust. I was told that the car had suffered from a fire in the dash. That included water damages from the firemen, such as rust on the dashboard frame. They exchanged the vehicle for the customer who bought it new. But the car company declared the car reparable, so the dealership had to resale it. The car was first dismantled by the guys in the body shop about a year before. When their workload increased, they move the car into the mechanic' shop. Nobody was touching it. Then, one day, a guy from the front desk entered the garage and said: «The car is sold. It's now a priority to put it back together.» I can assure you that no clients ever came to check that car ... and that no one would have bought it, have they seen its condition. Remember that we're in an authorized dealership for a well known car company.

    The guy who took the job (who wasn't the one who took it apart a year before) had to go fetch some disassembled parts in the body shop and order new ones. I remember that electrical wiring that went from the dash to the side doors (electrical locks and windows) weren't available, even for a dealership. The guy ended up cutting of the melted portions of the wires and welded extensions protected by electrical tape. I don't know if the client ever knew what he bought, but I highly doubt it.

    But that kind of thinking also happen at the manufacturing level. I was working at a reputable, international, well recognized, computer company, where I was on the assembly line for computer chips. My job was to put a cap with a sealant on the substrate, put them in a jig that applied a certain pressure on the assembly and cure them in an oven. I then had to do a "bubble leaking" test by plunging them into an aquarium to see if they were sealed properly. One day - distraction, I guess - I forgot to apply the pressure before putting them in the oven. I told my supervisor and it was decided to just do the usual bubble test. There was an incredibly higher-than-normal failure rate, but the ones that passed the test were sent down the assembly line. Were they rejected later on? Were those subjected to a shorter lifespan? We will never know. And I'm pretty sure that the decision was taken without higher management or the engineering department knowing. And I've seen other examples where errors were corrected before sent to clients. It doesn't mean it wasn't good, but they were not identical to others for sure.

    This was an OEM. Imagine what happens in smaller companies that are working for OEMs and don't want to loose their contracts when things go wrong.
  14. Mar 13, 2017 #13
    then why dont they just say that?
  15. Mar 13, 2017 #14
    To who? For what reason?
  16. Mar 13, 2017 #15
    thankyou jack action. finally an answer written in plain english that makes sense.

    i know personally of a few cases where this has happened, and ive read dozens in forums(which i do realise are not nearly as credible). so i experienced that its not as rare as i used to think.

    i get that a more probable definition of a lemon would be one with the same recurring problem, as xxchrisxx said, if the gearbox keeps dying, perhaps gearbox or engine mounts are the fault, not the umteen gearboxs you shove in.
    but i have experienced first hand that there is a far worse kind of lemon.

    jack action, your example about the faulty chips totally makes sense. so any machine that has the faulty chips could well have electrical problems relating to the chip. and the owner of that machine would be driven crazy by that and wonder if he got a lemon.
    however, if that chip is a car ecu, then it would not affect the faulty crank machining, the damaged seal inside the fuel pump,
    or the fault in the glue used to hold a dash cover on.

    but i guess, as the answer i finally got, we dont know.
  17. Mar 13, 2017 #16


    User Avatar

    Staff: Mentor

    We did. But just about everyone who comes to the forums wants answers with a little more elaboration and a few more details than just, "No two cars are the same".

    Please realize that good communication is hard. It is almost never helpful to blame the other person if you don't understand something since it takes two people to have a conversation. If you don't understand what someone says, ask them to elaborate, to simplify their answer, or something else. Statements like this makes others feel like they've wasted their time trying to help you. Even if someone's answer doesn't seem helpful, keep in mind that they at least tried. That's far more than many people would do.
  18. Mar 14, 2017 #17


    Staff: Mentor

    Since this thread has run its course and answered the OP's question, we will close it with thanks to everyone here who participated in this discussion.

    Closing thread.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook