BSSD Plots

 
Here is an excellent example of    Above is seen a tiny  metal  dot
a metal cross short between two    between  an  SVX  input  channel 
silicon  detector  strips.  The    trace and  an  adjacent  bonding
strips are 20microns apart.        pad. The dot is 3 microns across.

The types of shorts seen above result in two svx channels fighting
each other.  One channel will swing to one extreme while the other
swings to the opposite extreme.  When the channels are on SVXs on
opposite sides of the detector, as one gets when the flaw is on the
silicon detector, the large currents cause the sharp spikes seen
in the RMS noise as shown below.  When adjacent channels on the same
SVX are shorted, then one gets not only a large spike, but a noisy
area around the spike.  This is also clearly seen on the figure below.

The first type of short is easily repaired by removing the wire bond
between the silicon detector and the svx input pad.  The second problem
is more difficult to repair.  One must cut through a layer of "glass" to
reach the trace below, which you also cut.  A silicon carbide scribe, 
10 microns in diameter is drawn across the trace several hundred times.
Eventually, after wearing off part of the tip, one cuts the trace and
the svx channel is isolated as seen below.

Once all of the "bad" channels have been removed, the overall noise
in the good channels tends to drop  by about 10% and looks like the
following:

Station 1 of the beam silicon detector as seen in the E781 pit.

Stations 1 and 2 of the beam silicon detector system

Revised last by CRN  12-Feb-96