This document contains mechanical
information on 3,6,9-chip ladders as well as explanation of relevant terms,
notations and brief description of measurements and processing data.
OUTLINE
· Push direction and ladder
types
· Zeiss measurement of
completed ladder
· Fitting procedure
· Quantities that summarize
mechanical ladder information
· Plots
Due to peculiarities of the construction procedure push direction during the construction of a ladder and during its installation into a barrel is not the same for all three(3,6 and 9-chip) ladder types.
To avoid confusion it is important to keep in mind that the terms “left-pushed” and “right-pushed” refer to the construction push, not the installation push.
RIGHT-PUSHED ladder (be it 3,6 or 9-chip ladder) is pushed TO THE RIGHT on the CONSTRUCTION FIXTURE looking from the active side towards the passive side.
LEFT-PUSHED ladder (be it 3,6 or 9-chip ladder) is pushed TO THE LEFT on the CONSTRUCTION FIXTURE looking from the active side towards the passive side.
What makes a distinction in terms of push direction between 3-chip ladders on one hand and 6 and 9-chip ladders on the other hand is specific placement on the construction fixture. 3-chip ladders are build in a rails-down position whereas 6 and 9-chip ladders are build in a rails-up position
In contrast during the installation a ladder of any type is inserted into a bulkhead channel in a rails-down position
Therefore during the installation
right-pushed 3-chip ladders are pushed to the right
left pushed 3-chip ladders are pushed to the left,
whereas
right-pushed 6 and 9-chip ladders are pushed to the left.
left-pushed 6 and 9-chip ladders are pushed to the right,
The ultimate check for ladder push direction in the barrel can be the following :
Any ladder regardless of its type and location in the barrel must be pushed
COUNTERCLOCKWISE.
Completed ladder is measured on Zeiss machine out of fixture. Locations of fiducials on silicon detectors and edges of beryllium notches are measured. For 6-chip ladders only beryllium notches are measured.
16 fiducials are measured for 3-chip ladders, 8 for 6-chip and 14 for 9-chip(2 of the existing fiducials are covered by HDI). On each beryllium notch 4 points are measured : 2 on the X locating edge and 2 on the Y-locating edge. Eventually only most prominent point on each edge is kept. Measurement points on a completed ladder are shown in this picture. For double-sided ladders both sides are measured and the average of two measurements is used if both measurements are available.
The coordinate system in which measurements are taken is established as follows: first the XY plane is established on the silicon surface, then coordinate system is established so that the X axis passes through the central leftmost fiducial which is chosen to be the origin and the central rightmost fiducial.
· Fitting procedure
Since these measurements are not perfect the mechanical picture of a ladder that we get from such a measurement cannot repeat precisely the real mechanical picture of this ladder.
To make the resemblance of the two as close as possible we implement 2-D fit procedure
This procedure involves two steps:
1.Exclusion of the points that are found to be poor measurements from a data set
2.Fitting of a ladder
These steps are described below in some detail:
1.First of all we exclude from consideration the fiducials whose measurements are either clearly accidents (absolute value of X or Y residual > 100 microns) or missing.
Then we look at each detector separately. Because of good manufacturing quality the fiducials within a single detector match the nominals fairly well (within 2.5 microns).We can use this fact to separate out poorly measured points.
For each of the two detectors we make a fit by allowing rotation and translation and minimizing the following quantity:
Sum of squared 2-D distances between measured and nominal fiducilal locations. The sum is done over the kept fiducials within a detector.
As a result of the fit we find the most appropriate coordinate system. In this system we calculate the residuals, keep only those points for which both X and Y residuals are within 3.5 microns and discard the rest as bad measurements. Thus we end up with a reduced set of data that consists only of reasonable accuracy measurements. This set is then used in step 2.
2.What we are ultimately interested in is relative positions of silicon fiducials in the coordinate system which makes a ladder look best. The coordinate system that was used during the measurements(with X-axis passing through the central leftmost fiducial and the central rightmost fiducial) may not necessarily be such a system for two reasons:
-the two fiducials may not have been measured precisely;
-misalignment of two detectors with respect to one another can bias a ladder picture in this coordinate system
Therefore we make the same kind of fit as in step 1for a whole ladder (silicon fiducials on both detectors) this time. Then we recalculate locations and resudulas of silicon fiducials and beryllium notches in the coordinate system found through the fit. These are the numbers that we analyze and present. There is an exception for 6-chip ladders: since misalignment is not a problem for a single detector we use measurement data without fitting.
· Quantities that summarize mechanical ladder information
Below is the list of quantities that we use to describe mechanical quality of a ladder. Note that as it was pointed out at 03/06/00 SMT Assembly Meeting this list is not entirely sufficient. Currently work is being done to complete it with misalignment information for combined push types.
Beryllium notches:
· Y residuals for beryllium notches on relevant side
In plots appear
as Active Y Residual and Passive Y Residual
· X residuals for beryllium notches on active side
In plots
appear as Relevant X Residual and Irrelevant X
Residual
· Y misalignment for beryllium notches on relevant side, i.e. the difference between active notch Y and passive notch Y
In plots
appears as Y
Misalignment = Active Y - Passive Y
· X misalignment for beryllium notches on active side, i.e. the difference between relevant notch X and irrelevant notch X
In plots
appears as X
Misalignment = Relevant X - Irrelevant X
Silicon fiducials:
We use only one quantity to describe a quality
of the alignment of silicon fiducials:
![]()
Where DU is Y residual of silicon fiducial and the averaging is performed over all fiducials on a ladder.
In plots
appears as Silicon Y Misalignment
· Plots
Two peaks on pages 1 and 2 of 3-chip are related to different push types.
Two distinct peaks on page 1 of 6-chip are related to different push types.
Two distinct peaks for Irrelevant X Residual and X Misalignment we couldn’t
associate with anything (e.g. construction time, push type)