Charles P. Kuss wrote:
Paul wrote:
I bought a 1991 3.8 L Taurus rad in the junkyard
today for $30 because it had a leak in the core
to header tank joint connection.
Normally a non-leaking version would cost $100.
That was OK since one should put new aluminum header
tanks on it anyway.
snipped
Paul,
While I feel that you usually stress that people not cut corners
building their firewall forward packages, I take exception here. I
certainly don't see the logic of spending 2,000 to 3,000 hours building
an aircraft, then using an eight year old radiator/radiator core in it.
You are absolutely right. Severe brain fade on my part.
sorry I implied that in my message. However Tracy Crook got
at least 800 hours out of a couple of junk yard evaporator
cores.
Cars are made to turn to dust on their tenth birthday now-a-days.
This is quite true. So why try to reuse a 9 year old radiator core? Is
the $170 saved (assuming $200 for new)on this critical part worth the
risk? Have you really saved any money in the long run? I doubt it.
Absolutely. Start with a new one.
A home made silicone rubber gasket should last
much longer. It looks quite feasible to machine
new header tanks out of two by two aluminum
bar stock and crimp them back on with silicone
gaskets. I will do a drawing on this soon.
The reason composite radiators fail is because of the constant cyclical
changes in temperature and pressure that they are subjected to. The tabs
being aluminium, they have a finite fatigue life. Why reuse a design
that WILL fail again?
It is not the aluminum tabs that fail. The tanks being made of plastic
and the gasket being made of cheap rubber harden with age.
That is what caused the leak. Old plastic tanks have
been reported on here to crack as well.
The core itself held up well.
The number of thermal cycles the core has seen is way below
what you would expect to see with a true metal fatique
failure. Twice a day for ten years is only 7300
cycles.
If something is vibrating it can see that in one minute.
Rads should be mounted as they are in cars. With rubber
vibration isolator mounts.
This core lasted for ten years with far more thermal cycles than
any privately owned aircraft would see. Ageing and corrosion is
much more of a problem in private aircraft service.
New header tanks could also probably be welded
on as the end plates look to be about 0.040
to 0.050 thick aluminum.
I feel that this would be the correct procedure. It will require a
skilled welder.
This might be tricky as the tubes appear to be furnace brazed into
the end plates. I say appear as I am not sure.
This poses another potential problem with modifying a used radiator.
They could be merely bonded in with epoxy
from the outside. If that is the case they could
easily be re-bonded after welding.
While the epoxy
will bond the parts, it's long term strength is suspect. The difference
in thermal expansion rates of epoxies and aluminium, will insure that
the joint will fail at some unknown point in the future. Epoxy expands
very little with heat. Aluminium expands considerably. This is exactly
the sort of thing that Allan Tolle (and others) is warning us not to do.
You are not going to have much choice in the matter. I suspect if you
buy a "new" radiator from Griffin it will use mass produced cores
that are made exactly the same way Ford, Toyota or Honda or whomever
make
their cores for them. Brazed aluminum joints can also fail. There are
no garrantees.
I suspect epoxied cores are now standard.
Check with Griffin on this matter. BTW a NASCAR race car has an expected
life of two to three hours. It is rare for a part on a race
car to last longer than about 100 hours. Given that, where are you
going to find an aircraft quality radiator core?
I want a powerplant with a reasonable TBO. That means that I must
insure that every part of the system will also have a reasonable TBO.
Each builder must decide for himself the definition of "reasonable".
It's your life!
Respectfully,
Charlie Kuss
Paul
The Aircraft Rotary Engine Newsletter. Powered by Linux.
http://home.earthlink.net/~rotaryeng/ http://www.linux.org