Does anyone know of any website or other source to get general statistics about when most cars need repairs? Of course I suspect a bell curve around the 80,00 to 110,00 range - but can’t support that with numbers.

I’m unaware of any other than Consumer Reports, and that’s not really they type of data you’re seeking.

I don’t think it would be a bell curve. But I can only offer anecdotal experience…40+ years of it.

Typically, based on my experience, I expect to begin replacing chassis components like struts, bushings, etc. at about 150,000 miles. I expect to begin replacing components peripheral to the engine (alternator, starter, AC pump, power steering pump, stuff like that) somewhere over 200,000 miles. I’ve never worn out an engine even after over 300,000 miles, so I can’t offer that insight.

My experiences I’d expect to be wildly different from that of others. Make and model of vehicle, vehicle use, driving environment, driving style, and maintenance make all the difference in the world.

If maintained somewhat according to the schedule the 150k mark is when more expensive repairs mostly happen(some people call high mileage maintenance to make themselves feel better).

The 150k figure is for cars getting 12k-15k/year in mileage.

I tried to develop such a thing, but the results were all over the map.

In general:

  1. On good Japanese cars, with normal maintenance, repairs start well beyond 100,000 miles and coninue to as long as 400,000 miles when too many things wear out and you should scrap the car.

  2. On unreliable European cars with expensive parts, repairs will start at 60,000 or so and by 150,000 miles the car often is no longer worth fixing.

  3. In the past with such throw-away vehicles as the Hyundai Pony, Renault Le Car, and British imports, both the body and the mechanicals/electricals would bite the dust at 65,000-70,000 miles.

  4. On sturdy US trucks and large rear drive cars, repairs would start at about 60,000 but the cost would be reasonable. By constantly replacing parts these vehicles can go to 400,000 miles as well, as testified by many of these used in taxi service.

Some component life ranges:

  1. Shock absorbers 20,000-100,000 miles

  2. Struts 50,000-125,000 miles

  3. Ball joints 30,000 to 125,000 miles

  4. Transmissions 70,000 to 300,000 miles (before rebuild)

  5. Engines 100,000 to 600,000 miles (before rebuid)

  6. Starters and alternators 50,000 miles to a 150,000 miles

  7. A/C compressors 50,000 to 150,000 miles

You get the picture. Since cars last longer and warranties are longer, component life has increased over the years. The items listed above are from my actual observation on mine or colleagues’s cars.

There is no bell curve for the whole industry. There is a bell curve for each company’s parts, such as Volkswagen, Toyota, Chevrolet, etc. Parts on Russian and Chinese cars, for instance would last a fraction of those listed.

All items and figures listed are based on normal required maintenance.

Who knows, but it’s not a bell curve, it would be ever-increasing as the miles pile up.

I see what you are saying, but I do believe it is a Bell Curve as related to miles when they START to fail. If you look at total failures and when they happen, then it will be an upward curve over miles, which is what I assume you are thinking, right?

Last April’s CR had a chart on this exact issue, sort of when all repairs occurred, and that is what it showed. At first glance, one might assume it is when they start needing repairs; this is not so because the same cars will be reported each year. On a Bell Curve of the start of repairs, only at one point because a starting point can only occur once.

A Bell Curve of the START of failures will mean a very few cars start breaking right away, they are called, ahem, Yugos. As miles increase more cars start to break often, at some mid-range of miles, most cars have started needing occasional repairs, and as miles go way up, a very few never seem to start needing much repairs.

I don’t know of anyone who has ever done the Bell Curve chart of starting to need a lot of repairs, but that doesn’t mean it doesn’t exist. The CR chart is more relevant, and one can with care deduce from it what you want to know.

By the way, my first light bulb on my 2002 Sienna failed at around 85,000 miles. The first oxygen (called A/F on my Toyota) sensor failed I think around 120,000 miles. For most folks, it doesn’t matter, because if the CEL comes on with a bad sensor, you can take it to a local shop right away. In my case, it is a long, long drive to a place that would sell my sensors, so I replaced the other 2 at around 150,000 miles as a preventive measure. The cat converter is much more expensive than the sensors are.

Batteries perhaps around 5 years, except in Mcallen people tell me they go out after around 2 - 3 years. I don’t know why, unless they run them dry?

Wow, great posting, Doc! Good job!

No I would not expect a bell curve. It is not likely to be a even distribution.

Personally I believe you would find that repairs are most likely to be needed not long after the owner ignores maintenance.

Further complicating matters, it would need to be a function of both age and miles. Also, as you repair stuff, the potential pool of stuff that is in imminent risk of breaking goes down, slowing the increase in number of repairs. So I would propose that instead of a bell-curve, it would likely be a logistic function, where it increases exponentially (though probably in this case with a low constant), but where the rate of change begins to decline as it approaches a limit.

Try fitting 02 sensors,evap systems and catalytic converters in the list. These three items take up a lot of space on CarTalk and can be difficult to correctly diagnois.

We see these failures way before 80,000 in many of the posts we recieve.

Then you have the problem areas like intakes for GM vehicles that often happen well before 80K and then you get the squeaks and rattles that are “in” the car upon delivery.

You do specify “repairs” which is very different from replacing a wear item.

yes, old school, those items can cost quite a bit.

  1. Sensors in good Japanese cars seldom give trouble if the car is well maintained. I have personally never had problems with any. Sensors on European cars are much more troublesome and we have had many posts on this. expect problems around 60,000 miles and over.

  2. Evaporators take quite a beating in the Southern states. I was given a company car (a Mercury Sable) with 55,000 miles on it, during an assignment in Houston. The A/C did not work and the repair shop told me the system couldnot be fixed by simply adding refrigerant, and the conpressor and evaporator both need replacement.

That’s the bottom end; for those of us in colder areas, the compresor could easily last 150,000 miles in a good car. The one in my Ford Granada failed at 80,000. miles, and the one in my Caprice was losing capacity at 130,000 miles.

  1. Catalytic converters in good cars will last the life of the car, IF the car is regualarly tuned up so the cat does not get poisoned. I suspect that when oil consumption starts the cat’s life will quickly end. From my experience, although I’ve never “worn one out”, 150,000-200,000 miles might be agood average life in those states that check emissions. All based on regular maintenance.

Joseph, don’t confuse a bell curve with “cumulative failures”. The Bell curve is close to a Normal distribition of failures around a certain mileage, say 100,000 miles. Manufacturers want an average figure of a certain miles and a deviation that is relatively narrow, since early failures mean warranty claims.

The secret is to have good quality control and design to minimize early falilures, and have the average high enough so that the “bottom tail” of the bell curve are affordable failures.

The distance between the tails would typically be 90,000 miles and 110,000 or 20,000 miles with good quality control.

What makes the actual spread so much larger is the quality of the driver and the care or lack thereof that the equipment receives.

The key difference between good Japanese, European, and American designed parts is the length of testing and the design life which is the shortest for European, medium for American, and the longest for Japanese. Korean design life is now catching up to Japanese.

Chinese, Russian and Indian design life and quality control is still in the dark ages by comparison.

Doc,for evaporators I was really speaking about repair and diagnosis of the evaporative emissions system. We have had posts from many people with trouble in this area along with tank filling problems (posts are so common it’s like H1N1)

My feeling is with any mandated emissions system the warranty period should be longer (with a maintiance factor for abuse included)

Speaking of statistics did you know 2/3 of people have trouble with fractions and 75% of people have trouble with percentages. (:

Sorry abut the misunderstanding. Yes, we have an operator problem here with most people still wanting to “top off” the tank. The kid down the block where my wife tanks keep wanting to do this.

Since the engine emissions are guaranteed for a very long time or mileage, I would agree that the vapor management system should be covered that long as well. “Stop at the First Click” should be in bold print just under the gas door!!

On my Ford Granada the vapor canister failed at 50,000 miles, in spite of good filling habits, and I did not bother to fix it since there were no engine warnings on that car.

Yes, that reminds me of a recent article I read about the quality of our education system. The article said the standards were so bad that 50% of the students rated below average!

Joseph, don’t confuse a bell curve with “cumulative failures”. The Bell curve is close to a Normal distribition of failures around a certain mileage, say 100,000 miles.

I was a professional economist and statistics was part of my job. I would not expect to see a bell curve result from a study of car failures. I might add that to be meaningful you would also want to consider different kinds of failures as they may be given different weight based on cost to repair, possible injury and inconvenience.

That said depending on what exactly what you are looking for you might end up with a study resulting in a bell curve. However the only thing close to that was a cumulative line graph on page 84.

I suspect that in reality we are both correct, but are looking using different perspectives.

Thanks Joseph; when I perform reliability analysis, we use both the “bathtub curve” which shows initial failures (infant mortalities) followed by a flat line of low frequecy failures, then towards the end of the life an upward slope of increasing failures. This curve nearly always applies to things that wear out as they age. It does not apply to electronics in general. With those we have sudden “catastrophic” failure, i.e. without warning.

When plotting the failure at the end of the component’s life it looks like a Bell curve close to the mean time to failure, with the right hand tail slightly longer and going to zero, while the left hand tail started from that very low ongoing rate. Since there are some cumulative failures leading up to the Bell curve, the Mean Time Before Failure (MTBF) point is ahead of the mean on the Bell curve.

In reality, failure analyis is more complicated, since many failures are induced by poor maintenance or not enough maintenance.

Agree that the curve is shaped like a bell curve, but may not be a statistcally perfect Normal distribution.

“The article said the standards were so bad that 50% of the students rated below average!”

Obviously, you do not live on the shores of Lake Woebegone…

Old cold war joke, The Americans and the Soviets were in a car race (just the two of them)The Americans won but the Soviet press reported “Soviets finish second while Americans finish next to last”.