Both of John’s turbos were “Junkyard direct” . Mine was also and I noticed considerable play while pushing up and down on the wheel. I’m sure at very high RPM the wheels get in some harmonic vibrations and it is matter of time before they snap at the root. Just imagine your Dremmel tool at high RPM with the cutting wheel able to play in the chuck? I rebuilt my turbo and there is no end play at the wheels.
Bulent

> ...
> ... It's the turbine wheel that failed, not the compressor wheel. On
> each occasion the weld broke right at the base of the turbine wheel, the
> wheel came off and blocked the exhaust outlet.

   The *weld* broke?  Does that mean that the shaft is
still intact?  

   Someone else mentioned the term "friction weld"; I'm not
familiar with it.  This is something different than
"interference fit"?

   Gotta wonder what kind of stresses would cause that weld
to break.  Is the shaft keyed?  Maybe it should be - although
that would complicate balancing it.  Perhaps balancing is
part of the problem.  Maybe at umpteen thousand rpm, +/- .5
gram isn't close enough.  

   Two senarios come to mind: (1) Turbine wheel heats up
faster than the shaft; develops a small amount of clearance
at the end opposite the weld; starts wobbling a little, too
much for the weld to absorb; (2) Turbine gets a little
wobbly at high speed; after sustained high speed operation,
the vibration overcomes the weld.  

   I don't quite understand why the only thing holding that
turbine wheel on the shaft is a *weld*.  Isn't the turbine
Iconel or similar?  Is the shaft also Iconel?  If it isn't, the joint is closer to being a brazing (in principle,
anyway).

My somewhat semi-random thoughts ...

Dale R.
COZY MkIV-R13B-NA #1254