High B+ regulation in standby.
The margin for error is very small on the hv in the splitter.
As the capacitors fail in the power supply, the standby voltage will
drift over time up to as high as 165V, versus the 140~145V that is
suppose to be there.
Since that is a 15% increase in initial B+ at start up, which is beyond
the 10% margin for the insulation in the hv block, the hv is too high
initially at start up and this starts internal arcing at start up.
Over time these carbon tracks finally cascade and the hard carbon arc
short is started. The problem arises in that the regulator kicks into
run mode in about one second after turn on. The regulator is then
running off the horizontal feedback pulses for regulation. Since this
happens so fast and is only slightly above the x-ray trip voltage, the
x-ray circuit never triggers. Each turn on with standby B+ running
high will cause an arc over. The big problem we found was that if the
smps was not repaired and was the cause of the failure, it most times
would take 6 to 9 months before we would get a call back on the set
being dead. So shop policy had to be verify the smps secondary
voltages on ALL ctc169 sets, cr4118 and cr4116 must be in proper range
and clean looking on the o'scope. Extra 5 minutes of work to guarantee
no future problems.

The cr4118 failure can also cause the B+ to run high as it starts to go
leaky and fail. However I have never seen it fail to the point that
the B+ will run over 150V before the 15V line was too low for the set
to attempt to start up.

It is mainly the 2.2uF capacitor on the standby feedback line of the
smps that goes open that causes this problem. It is a simple matter to
measure the standby B+ voltage on the cathode of cr4116 with the tv
turned off to verify that there is not a problem.