CB Inspection in Detroit

Date: Mon, 23 Aug 1999 15:02:56 -1000 (HST)
From: Ian Pain 
To: jcmt_all@jach.hawaii.edu
Cc: Dean Shutt 
Subject: JCMT_ALL: JCMT central Bearing inspection 

Hi All

The central bearing is in Detroit at the moment, and I've dropped in
en-route to the UK to witness some investigative work. This is an interim
report, and will be supported later by the full SKF report.

 The first step of strip-down was removal of the radial bearing from the
tapered shaft. This was a little difficult: initial use of hydraulic oil
on the removal port failed to crack the taper-fit as partial leakage of
oil occurred around the periphery of the fit (indicating either a poor
original fit of some post-installation dimensional change due to corrosion
or casting stress relief). Use of grease in place of oil allowed
sufficient pressure build up to release the taper fit.

This allowed the assembly halves to be separated and the thrust races
examined. Light dimples were evident in the race even before the grease
was cleaned up. Also noticeable was the rather low amount of grease
present, with obvious signs on the cage and balls that grease churning had
not taken place : in many cases original bare metal was present which had
never seen grease contact. 

Upon cleaning the races the bearing defects were more clearly visible. A
full set of defects (44) existed on both upper and lower races, matched to
ball locations. The defect is rather complex and will require more
detailed analysis to pin down the true structure. Initial inspection shows
what looks like a fairly deep pocket in the race shoulder, with coppery
colouration from either corrosion or grease-breakdown products. A
shallower defect - almost elliptical - stretches across the centre of the
race groove and it is this which was most likely connected to the central
bearing 'glitch'. Radial scratch-type marks (either false or true
brinelling) are present in and around the defect. Ball-tracks were visible
on the race too in some locations, and the general appearance of the race
was of a highly polished glaze - this "crazed" surface is indicative of
low lubrication levels, but is not entirely unexpected at low speeds and
temperatures.

Closer examination of the race was required to establish precise natures
and likely causes of these multiple features. Removal of the race was
expected to be difficult given the heat shrink fit. The race was clearly
non-salvageable for later refurbishment and I decided to allow removal
even if this caused some damage.  Options were discussed and  the least
destructive method (to the original casting) was to drill from the
underside and tap the casting for pusher screws. The casting was inverted
for this operation and left unattended whilst tools were laid out. Shortly
afterwards a loud clang caused near-heart failure in three SKF
technicians. It appears that the original heat shrink fit was not as
effective as intended! As we have evidence that a heat shrink was
performed (procedure witness report in the JCMT documentation series) then
this loose fit can only be explained now due to either (a) casting stress
relief with time or (b) the original components were machined to a looser
size but not verified as such before the heat shrink took place.

The balls and cage were inspected also. The cage appears undamaged, and
the balls were in good condition to the naked eye; although discolouration
existed which would match the pocket defects in the race very well. 

Final examination for today was of the radial bearing and races. Again,
low grease travel bit otherwise no initial signs for concern. Rollers and
cage were left to de-grease overnight

Although the full set of dimple-like defects was expected, the complex
defect pattern's do not on first view point to a single cause of failure.  

At least the following remain possible: storage corrosion, transport
damage, impact overload (eg earthquake), electric arc strike(s) due to
ground loops/welding/lightning, servo-chatter, poor lubrication. The fact
that all features are concentrated in azimuth points to mutliple
occurences of problems in a single area (presumably thus either a
transport problem or indicative of the favourite 'parked' azimuth during
early construction or routine observing).


Many photo's taken, and when I figure out how from my laptop I'll try to
send some by email. Tommorrow's plan is to finish the initial visual
survey and discuss what additional metrology is required to ascertain
cause of defects well enough to ensure that we can be confident they won't
re-occur. As a first off, microscope examination will be able to rule some
of these in or out, and a metallurgical test (hardness) can be made to see
if the bearing steel was defective in any sense. A straight-forward
physical defect profile will also assist in correlating which feature was
responsible for the JCMT 'glitch'. More later.

Best Regards

  Ian Pain

------------------------------------------------------------------
  Senior Mechanical Engineer, Joint Astronomy Centre		
  660 N.A'ohoku Place, University Park, Hilo, Hawaii 96720. USA	
  Tel: 	US-(808)-969-6540	Fax: 	US-(808)-961-6516	
  Email:pain@jach.hawaii.edu	http:	www.jach.hawaii.edu	
------------------------------------------------------------------