Upper-Lower Correlation

The upper-lower correlation tools are used with flip-chip bonders. Here, a chip must be placed exactly on a wafer. The position of the component is determined by the structure on the underside. The structure on the bottom side cannot be seen with a camera from above the wafer. Therefore, the offset (X, Y and rotation in world coordinate space) between the structure on the bottom side and a structure on the top side is determined in an upper-lower correlation station with two cameras. One camera (upper-camera) sees the component from above and a second camera (lower-camera) sees it from below. The offset is stored in the upper-lower correlation and can then be added back over the wafer (destination camera).

CorrelateCamerasTool

This tool is used to establish a relationship between the lower and upper cameras. For this purpose, a calibration mirror is moved between the cameras. Both cameras see the same rectangle. The CorrelateCameraTool performs the following steps:

  • Find the corner points of the rectangle in the upper camera image.

  • Find the corner points of the rectangle in the lower camera image.

  • Set the N-N Fixture tool to the points found.

  • Run the fixture tool.

  • Notice the transformation.

../../../_images/CorrelateCamerasTool.png

  1. Find Rect Upper: Settings for the Gefasoft InspectEdgeTool to find the rectangle in the upper camera image.

  2. Find Rect Lower: Settings for the Gefasoft InspectEdgeTool to find the rectangle in the lower camera image.

  3. FixtureNPointToNPoint: Settings for the Cognex N-N-FixtureTool to calculate the transformation from the lower to the upper camera image.

  4. Fashion:

  • SingleImage: Uses the images of the input terminals InputImageUpper and InputImageLower to determine the corner points of the rectangles.

  • MutiImage - MeanPointsForN2N: Uses the images of the InputImageArrayUpper and InputImageArrayLower input terminals to determine the corner points of the rectangles. The corner points found in the Upper and Lower images are averaged and then passed to the FixtureNPointToNPoint tool.

  • MutiImage - PassAllPointsToN2N: Uses the images of the InputImageArrayUpper and InputImageArrayLower input terminals to determine the corner points of the rectangles. The corner points found in all Upper and Lower images are passed to the FixtureNPointToNPoint tool.

  1. LastRun.Upper.LastRunImage: The red circle shows the center of the lower camera image, shown in the upper camera image.

See also
InspectEdgeTool

CorrelateTheTool

This tool creates or overwrites a correlation offset. A correlation offset stores the component position of the lower camera image relative to the upper camera image. A correlation offset must be created for each bond header. The tool requires as input the position and rotation of the part in the lower, as well as in the upper camera image. These can be found e.g. by the GetPositionTool. The positions must be passed in pixel coordinate space.

../../../_images/CorrelateDieTool.png

  1. CorrelationKey: Name of the correlation data to be filled. CorrelationKey is filled by the command parameter Bondhead.

  2. Delete: Deletes the selected correlation data set.

  3. AngleInDegree: Specifies whether the angles are in degrees or in radians.

  4. LastRun.LastRunImageUpper: The red circle represents the position of the component in the lower camera image, shown in the upper camera image. The yellow line represents the rotation of the component in the lower camera image in the upper camera image.

  5. PosUpperX/Y/R: Reference position in the upper camera image. The position must be passed in pixel coordinate space.

  6. PosLowerX/Y/R: Relevant position in the lower camera image. The position must be passed in pixel coordinate space.

  7. TransformWordlUpper: World coordinate space of the upper camera image.

  8. UpperLowerCorOffsetX/Y/R: Offset (X, Y and rotation in world coordinate space) determined by CorrelateDieTool between the relevant position on the bottom and the reference position on the top.

../../../_images/CorrelateDieUpperLowerPositions.png

  1. PosLower: Relevant position in the lower camera image.

  2. PosLowerInUpper: Relevant position in the lower camera image, displayed in the upper camera image.

  3. PosUpper: Reference position in the upper camera image.

AddCorrelationOffsetTool

This tool adds a correlation offset to the part position. I.e. one searches for the component position (reference position) in the upper camera image (destination camera) and receives the position in the lower camera image. The correlation offset must have been filled before by the CorrelateDieTool. The tool needs as input the position and the rotation of the component in the upper camera image. This can be determined, for example, by the GetPositionTool. The positions must be passed in world coordinate space or pixel coordinate space.

../../../_images/AddCorrelationOffsetTool.png

  1. CorrelationKey: Name of the correlation data to be used. CorrelationKey is filled by the command parameter Bondhead.

  2. Delete: Deletes the selected correlation data set.

  3. AngleInDegree: Specifies whether the rotation is in degrees or in radians.

  4. LastRun.LastRunImageUpper: The red circle represents the position of the component in the lower camera image, shown in the upper camera image. The yellow line represents the rotation of the component in the lower camera image in the upper camera image.

  5. TransformPosTargetToWorld: Must be selected if the position is passed in pixel coordinate space.

  6. PosTargetX/Y/R: Reference position in the upper camera image. The position must be passed in world coordinate space or pixel coordinate space.

  7. CorrelatedPosTargetX/Y/R: Component position with correlation offset.

  8. UpperLowerCorOffsetX/Y/R: Stored offset (X, Y and rotation in world coordinate space) between the relevant position on the bottom and the reference position on the top.

../../../_images/AddCorrelationOffsetUpperLowerPositions.png

  1. PosUpper: Reference position in the upper camera image.

  2. PosLowerInUpper: Relevant position of the lower camera image, displayed in the upper camera image.