My mechanic has been working on the a/c evaporator, which is really a pain to get to. Anyway, he reports the following. (BTW, it seems that the problem could have been prevented with paint):
I am finally at the point that the problem is identified. I made some
test fittings so that I could evacuate and charge the evaporator core
while it was sitting on the bench. The leak was not visible with just
the top of the unit removed, so I had to remove the core from the
housing. The leak is caused by a corrosion pit. The corrosion is from
the outside. There is another deep corrosion pit about 1/2" from the
leak, but this pit was is not leaking yet. The question is why. Both of
the pits are in areas that had some foam tape glued on. I wondered if
the glue or the tape promoted the corrosion, but there was no evidence
of a problem in any other area. I wonder if the foam held atmospheric
moisture and salt next to the aluminum until a pit was formed, but there
is no evidence of any such reaction anywhere else on the core. Both of
the deep pits are right in line with rivets that hold the main casing
together and I first thought that the person that assembled the case had
slipped with the drill and nicked the tube, but there is no evidence of
mechanical damage. It is clearly corrosion. The rivets are very close to
the tubes at these points and the thought of some kind of dissimilar
metal electrolysis entered my mind, but I think this is kind of far fetched.
I am going to try to research some repair or replacement options. A
copper core would be the way to go, but I doubt that finding the right
size and fitting layout would be possible. I wonder if anyone makes
custom cores. I was hoping that I could, at least, do something to make
the unit removable without taking half of the car apart. That looks
very, very, unlikely given the space available and the component layout.
Another thing that makes me nervous is that the heater core, which is in
the same box, is also aluminum and it has the crimped together ends that
are a common source of problems.
The first picture shows the test setup. The second shows the green trace of the dye at the leak.
The third picture shows the site of the leak. The fourth picture shows the other deep pit. This second one is not leaking now.