I have a 95 Accord. I got it in '99 w/ 65K. In 2001, at about 75K, I had the timing belt and water pump replaced. So, here we are in 2008 and it has 98k. According to the manual I should replace the timing belt at 90k or 6 years. That is one year beyond the time limit. I’m wondering if anyone has seen failures of timing belts purely from age, not mileage. I know it’s past the time but it only has about 30k on the belt. Are the belts better material than the originals? Thanks for any input.
Mark.
Life expectancy of a timing belt can be plotted on a bell curve. There will always be some on the low side of the curve. Some will be high. The majority will be in the middle. Those at either end are a matter of chance (or, “luck”, if you wish). There will be people who have won, or lost, the game of chance. It’s simply safer to take the middle road (middle of the curve).
A crashed engine will cost you a couple of thousand dollars (or, more) to “cure”.
It’s far to risky to take the chance. This is a interference engine…so if it does break it’ll cost you THOUSANDS to get the engine repaired. Do you Feel Lucky??
Agree with the Bell Curve, but the belt (at least a Honda belt) has a 100% safety margin under NORMAL USE. Volkswagens, on the other hand, often did not even make it to 60,000 miles in the nineties, and broke before the warranty was up.
Having said that, I would observe the factory interval to be sure. Under exceptionally extreme conditions, Alaska North Slope, Death Valley, I would change the belt every 2 years, regardless of mileage. When I worked in the Middle East, it was normal practice to change all rubber under hood components, such as belts, hoses, every 2 years. Survival in the desert is very important.
If you had ever seen a timing belt, you would have wondered how it could last even one year. Too bad it costs so much to change but I will have to recommend the work be done at the proper time interval. You would feel bad if you started breaking things after so many years of being a good owner.
Cam belt’s life limits are defined by mileage and time. A bell curve or failure distribution plot’s vertical axis would be number of failures. The horizontal axis could be either mileage or time but I see no way to combine both time and mileage to arrive at a useful plot.