Re: radiator neck failure mode analysis. This happened to me, and I devised a temporary fix by using a 1.25 inch plumbing down pipe from a bathroom sink ( Chrome plated brass- about 6" long), expanding the end slightly so it fits snuggly over thr remaining stump on the radiator, I epoxied the metal tube to the radiator, then reinforced the joint with fiberglass / epoxy. This has held for several months and I suspect may be stronger than the original, as thr clamp now presses on the brass not the plastic. While I left the metal pipe at its original 6" length, I may shorten it to minimize the potential leverage should someone lean on the pipe and re- break the plastic tube . Mike Posner.

[Mike Posner <posne@cts.com>]

At 11:39 AM 5/4/98 -0400, David Kavanagh staff wrote:
>
>It happened to me!
>On the return leg of a Taco Bell run on Saturday night, my upper
>radiator hose slides off the radiator neck! I pull into a nearby plaza
>lot and find the upper hose hanging off and steam everwhere! I push
>it back on, and start to wonder why it came off. Then I remember the
>postings about this failure mode. I take if off and see the ring left
>inside the hose! I loosen the hose clamp and clamp the hose on the remainder
>of the neck, add water and drive home. 
>I have a used radiator from an 88 or 89 that I'll be putting in tonight.
>
>So, about the failure mode...
>Is it just that the plastic gets brittle after a while and this happens
>to be the most stressed area? I noticed the crack happened right under
>the hose clamp. Does this mean that the extra pressure of the clamp
>is a contributor to failure? If this is true, It sounds like a good idea
>to clamp as close to the bump in the end of the neck as possible. This give
>you room to play if it fails on the road. I know I'll be carefull to 
>tighten the clamps gently as well! Also, on the replacement radiator,
>I noticed the plastic is thinner near the end of the neck. Does that mean
>it is better to clamp farther in?
>
>Lastly, any helpfull hints about changing my radiator? (I'll look at the
>archives later).
>
>
>	David
>	
>