Oldest or Highest Mileage Timing Belt: Who Holds the Record?

Maintenance/Repairs
41

When my brother bought his brand-new Beetle in 1964, it was shod with Continental tires, which neither of us had ever heard of. But, we assumed that “fine German craftsmanship” would ensure good quality. Yes, they were actually fairly-decent quality… until they wore-out at ~17k miles.

Fast forward to 2010, when my new 2011 Outback was delivered with Continental tires. I quickly noticed high-speed vibration characteristic of improper balancing. The dealership used their Hunter GSP 9700 Road Force Balancer on the tires, and they were okay… for about 4k miles. They kept rebalancing those Contis, and each time they would seem okay… for a few thousand miles. For some reason or other, they wouldn’t “hold” a balance for the long-term.

I still had 5/32 of tread remaining when I decided to ditch those Contis, and the difference with Michelins was nothing short of dramatic. The Michelin replacements have never needed to be rebalanced.
:thinking:

42

The ‘oldest timing belt’ is kind of like ‘oldest living smoker’. Sure, some smoker lived to 100.

43

my comment was based in part on a recent timing belt change for a Toyota at about 14 years and 114K miles
the belt actually looked ok, but one of the pulleys was showing initial signs of bad bearing(s), so changing that became part of the job

44

Just curious, have any of you had a car with the 3.0 Mitsubishi V6 (used mostly in Chrysler products in the 80s and 90s) where the timing belt broke?

45

I had a 95 Dodge Caravan with the 3.0L V-6. It was a great van, and the 3.0L engine was rock-solid reliable. I had the timing belt changed at 100,000 miles, and again at 200,000 miles. I never had any major problems with that van, and I 'd still have it today if it wasn’t totaled in an accident.

1 Like
46

That’s why Gates sets complete kits since a timing belt change is mostly labor.

47

Why? Noise? Broken?

48

I guarantee you that Honda is NOT doing the tests. They may be requiring the belt manufacturer to do the tests for them.

49

All do, but… and it is a big “but” how do you test for 90,000 miles or 9 years, whichever come first? The 9 years is impossible on a new design with no field history. You can run accelerated environmental tests to dry-rot the belt but these tests are generally poor analogs of the real thing.

The much maligned Vega 4 cylinder would complete a 1500 hour run wide open throttle, max load test with no problems. Seemed like a success.

What killed the silicon aluminum bores was the “Grandma Cycle” of short runs to the store and church. The acids formed by moisture and combustion products would eat up the bores. It took a couple of years for those to start showing up.

3 Likes
50

I believe this Mike.

We get this kind of thing put on the company where I work all the time.
It goes like this- you guys make belts, you should ensure they will last the specified lifetime in our application. Our reply- you’re right, we make belts but we do not make car engines. And, your car engine and its application is not exactly like any of the others. So our testing will not necessarily be indicative of how long one of these belts will last in your particular application. You may end up with a false positive for lifetime. We push back and our specifications make it clear that the onus is on the customer to provide validation in their application. We will help if necessary but they own it.

It is astounding how many engineering based companies do not fully understand the difference between verification and validation. An easy way to put it- verification ensures we built the product right, validation ensures we built the right product. Swapping those last two words makes all the difference.

The latter part can be approximated outside of the actual environment but nothing beats the real thing. Modeling and other design analysis, verification testing etc are all well and good but in the end, it needs to be validated in the application. Even then, as Mustangman appropriately points out, real time testing takes too long. So all kinds of accelerated life testing is carried out looking for weaknesses. To be sure, many times this weeds out the major flaws and allows the product to go into production with less risk.

The better plan is to do validation testing in parallel with your suppliers testing.

If Honda is eliminating all validation testing and relying on the belt manufacturer’s tests, they may be in for a rude awakening down the road and then the phrase- “how could this happen?” gets airtime…

1 Like
51

seems to jibe with the bearings being recognized as important as the belt in “record” cases

52

Mike: If you make parts for Honda you make them to Honda’s specifications with a proven life cycle test as well as Six Sigma accuracy. The problems that always plagued Jaguar was the very poor quality of their outside supplied parts, not just those from Lucas Electric… When Ford bought them that was the first thing they tightened up on.

When Japanese manufacturers set up shop in the US they encouraged their Japanese suppliers , like Denso, to follow then since US suppliers could often not meet their quality specs.

53

That’s right…and they prove that by running tests for the auto manufacturer and then showing them the results…sometimes the auto-manufacturer may have a rep when the tests are run.

54

Explain how you arrive at this conclusion on a new design and outside of the actual application. Should be interesting…

55

Every new item has to be tested through a number of cycles to determine the failure point. I once worked for a company making chain saws. The starter pull rope assembly was immersed in sawdust and dirt in the lab and pulled many thousand times to see how it wore out.

Lab testing to simulate real conditions are done all the time. When Toyota brings out an entirely vehicle, the components have gone through elaborate testing cycles. This is one thing the 1957 Chrysler products did not go through; they were rushed into production and the result was some of the worst cars ever built! I had one!

The degree of pre-production testing and the quality of the design determine how long a trouble-free life the product will have,. Accelerated stress and wear testing is commonplace and practiced by all quality focused manufacturers.

The wings of the Douglas DC 3 were tested by having a bulldozer drive over to determine their strength. On the other hand the wings of the British Comet did not have repeated stress cycles applied, and they failed form metal fatigue with disastrous results.

Since you are not a design engineer, I hope you will find this INTERESTING!

56

Well, I think I know which one of us is NOT a design engineer…

I don’t wish to get into any protracted discussion on the merits of design VERIFICATION. I think I was clear previously on the merits of such. The issue I was pointing out was your use of the term PROVEN. There is no such thing, especially when something is tested outside of the actual application.

I found your example kind of humorous- driving a bulldozer over a wing to determine its strength. There’s a tenet in design world regarding quantitative evaluation and relevance. The example has neither quality and should not be used as a measure to define application suitability or reliability.

57

In 1940 when there was a high demand for DC3s. Douglas Aircraft had to do something to test the strength of the wings! I did not imply that this was to way to test wings, but it was better than nothing which apparently De Havilland did for the Comets The post crash tests of those Comets were done in tanks of water and after a number of stress cycles the wings started fatigue cracking. That was a lesson for the entire free world aircraft industry. These simulated tests are now standard procedure.

To try to simulate real world conditions is done all the time; If you don’t do it and do not subject the item to a sufficient number of load cycles you tend to produce crappy components.

In the 50s car companies bragged about going 100,000 miles at 100 mph on the test track without breakdowns… Packard was one., and they had no breakdowns. That was no real world testing since the car ran at steady pace in a wide oval. No demand on the starter motor, power steering pumps, and other components.

So, Japanese companies tend to test their parts extensively and verifiably and lo and behold produce products that outlast other’s products and components.

The idea of testing is to simulate as accurately as possible, the actual operating conditions and then accelerate those to get a large number of cycles, preferably 200% of expected product life.

Nothing I am writing here is new; Dr. Edward Deming was assigned by the US government to initiate statistical quality control and testing during WW II. Industry hated him because they thought they knew better. US military equipment turned out to be the most reliable and easiest to fix in WW II.US army trucks and Jeeps were in high demand after the war.

After the war, Deming was invited by the Japanese government to teach the Japanese these principles. He succeeded beyond his wildest dreams, and the hallmark for quality in Japan is now called the Deming Award.

Not sure what your actual point was, but extended testing closely simulating actual conditions and then making sure you achieve 200% of expected design life is a good way to deliver quality products. This process is understood by many, but practices by much fewer. The Russians don’t do any of this to consumer products. Fiat and its suppliers likely don’t do much of it.

Many Americans were appalled in the 50s when they learned that in Sweden unmarried couples lived together for many months before tying the knot. If they were not compatible they would know soon. It also reflected the resultant low divorce rate in Sweden at that time compared to the US.
That was Actual Conditions for you in determining life expectancy of an arrangement.

6 Likes
58

This is the crux of the problem.

closely simulating actual conditions
making sure you achieve 200% of expected design life

You can’t make sure using closely…

I’ve been doing product design for more than 40 years, we do this stuff day in/out. It’s not something I read about. No one is saying it isn’t beneficial. Design verification is no substitute for validation. They go hand in hand. Using the terms you’re using and being definitive about outcomes is akin to drinking the koolaid…Margin is directly related to product reliability. No question. But if you’re not doing application based testing, you have no idea where the cliff edge is and therefore do not really know how much margin you actually have…

59

Companies that do all this end up with good products. However, a car has many parts and even a new model Toyota may have some teething problems with a new model. Good companies correct these shortcomings quickly at their own expense.

Examples of “over-design” could be the Eiffel Tower, the Brooklyn Bridge, and others. The Tacoma Narrows bridge is an obvious example of operating in the dark.

The first NASA moon flight was a tough one with a lot of unknowns. Yet it was very successful!

Clark Gable, the movie star was the first California customer for an XK-120 Jaguar.in 1949. It overheated in LA traffic. It took Jaguar till 1963 with the XK-E to put a bigger radiator with 2 fans in the car to prevent overheating…

My last design project was an oil processing (600,000 barrels per day) and loading operation in Africa involving an undersea pipeline to load 3 supertankers simultaneously.
The owner wanted 98% overall system availability year round, including any maintenance downtime. I had to design enough redundancy in the system to take daily fluctuations and the effect of tropical storms. To get the assured availability we had to convince the owners to spend an additional $125 million to twin the aging pipeline to the offshore loading platform. To get 98% overall availability in a system like this you need 99.98% reliability of each of the 250 or so subsystems. That meant buying the best components available on the market.

Today the system is built and operating as designed.

I’m happy with thorough testing at long cycles. This is my last comment on the subject.

60

I’m pretty late for this thread. Either way, 2012 jetta original owner. I just changed the timing belt and water pump at 254000km.