Coplanarity Test
Coplanarity testing system for electronic connectors and pins alignment quality control
1. Experimental Set up
The experimental set-up developed for this project is shown in figure 1.1. It consists of a refined lens which avoids optical aberration at the edges of the image and has a long working distance (14cm focal distance for the selected magnification 5.6/20x that enables easy access to sample). Zoom can be manually adjusted in case more magnification is needed for certain connectors.
A digital USB camera directly connected to a PC is used to record images and a very powerful LED ring light is used for sample illumination. This ring light can be manually controlled in intensity and orientation allowing a perfect illumination of the samples whatever their colour, characteristics and background illumination might be. This is very important since the binarization parameters used by the software are highly affected by light conditions.
The Stand is manually adjustable in height to focus image on the pins, regardless of the connector’s height. A diffractive glass is used as sampler to spread light homogeneously in the connector’s surrounding, convenient for binarization reasons.
2 Experimental procedure.
All co-planarity tests were conducted following the same methodology.
First the connector must be placed in the centre of the circular sample glass, horizontally aligned. As ring light is being used to illuminate the sample from every angle, having the connector on the middle of the diffractive glass will give the best possible homogeneous light. The horizontal alignment is arbitrarily chosen, but it is important to fix it for experimental reproducibility reasons. Meaning that it is needed to have a very precise sample position if comparisons between experiments must take place since, even though the linear regression coefficient (R2) used as quality parameter is not affected by the connector’s position on the base plate, the trend line is, and also the illumination characteristics that influence the binarization parameters, which translates into greater experimental error.
After placing the connector, a dimensional calibration must be performed for each set of experiments using a ruler and the software’s calibration feature (Figure 1.2)
In the calibration window, the exact value of a precise known distance must be entered so the software can assign a correlation between the image pixels and real lengths. As this technique is based on image analysis of the pin’s relative position to each other, all the calculations are affected by the calibration. Auto-size function must be selected.
In the former report the pixel calibration resulted in: 479 Pixel = 20 mm
Figure 1.2: Calibration
Once the calibration is done, the procedure for co-planarity test of all connectors using Smart-eye® tailored software is as follows:
1) An image of the connector is taken by using the feature Snapshot at the top menu of the screen.
2) Open the created image by clicking the icon with a magnifying glass on it, and then click the square at then bottom so you can select only the pins’ area of the connectors by enclosing it in a rectangle. Click ok button when asked to crop image.
3) Binarize the obtained image by using the button Particles on the top bar menu. A window will appear for binarization parameters control. Click manual, and modify the min. and max. tags until you can see clear uniform spots matching the pins. (figure 1.3)

Figure 1.3: image binarization for a connector with 4 pins
Note: binarization map varies depending on illumination characteristics and the connector’ colour (black/white) and if it has holes or other indentations. This is why a general value for min and max manual bianrization controls cannot be given for all the conectors. The strong point of this software is that all connectors can be inspected as a result of the binarization and illumination versatile controls. Thanks to the design of experiments analysis, optimum values for light and binarization parameters where obtained as shown on next section of the report.
4) Having the pins correctly binarized, the software can automatically calculate all morphometric features of each pin. Including their relative position, with pixel accuracy. For this click black particles in the binarization window and a table will display, showing the pins’ position and size and shape characteristics
5) After that click show on plots and a graph will appear. Select X and Y position the respective axis and the graph will show the alignment of the pins.
6) To quantify the results add a trend line and the linear regression correlation coefficient will show the deviation of the pins from a straight line. This parameter within a certain production tolerance can be used as critical to quality parameter.
Note: Smart-eye software for this project could be further developed to include the linear regression on its calculus and show it on the final report, if required by costumer.
Figure 1.4 shows the pins’ relative position (in mm) for the connector on Figure 1.3. It includes the trend line equation and linear regression coefficient with experimental error.

Figure 1.4: Plot with pins positions for connector A1 (in mm)
All experiments were repeated 10 times for each of the measured connectors to give statistical value to the results and calculate the experimental error. The experimental error has been calculated as the standard deviation from the mean of the 10 experiments for each connector. None of the experiments had to be dismissed due to data incongruence. The result for the R2 is given as the average plus/minus the standard deviation. Trend line’s Slope and Y-intercept were also calculated with experimental error, although it is not shown in the plot because gives not additional information, and results for them are shown as average value with significant figures.
Note: the blue marks in the plot correspond to the calculated centre of the pins’ tip. As the calculations are done in relation to the image, it has 1pixel resolution which means that for a calibration of 479 pixel = 20mm the accuracy of the measured distances is 20/479 = 0,042mm. This high resolution is possible because the developed technology is based on microscopic inspection image analysis.
3 Design of experiments and optimum parameters
Following the previous experimental procedure, experiments were conducted in different conditions (Figure 1.5) to assess which are the optimal parameters.

Figure 1.5: Experiments in different conditions.
Table 1.1 shows the parameters affecting binarization results and the optimum value for each of them. The optimum value was calculated by comparison of standard deviation in R2 at different values of the same parameter and by qualitative comparison of binarized images. For one parameter, the optimal value is given by the set of experiments with the lowest standard deviation
|
Parameter |
Optimum value |
|
Illumination |
Light Control 1 |
|
Lens-sample distance |
Focal distance: 14cm for 5.6/20 magnification |
|
Position of connector |
Centre of the diffractive glass |
|
Angle of the connectors pins |
Horizontal (arbitrary) |
|
Manual binarization parameters |
Min: 20; Max: 150 |
|
Ambient light |
Consistent |
Table 1.1: Affecting parameters and optimum value.
The illumination control should be the lowest possible since it makes manual binarizaton parameters more sensitive. The pins should be at focal distance to get optimal binarization, and this distance depends on the selected magnification. The position of the connector should be centred in the diffractive glass so illumination from ring light is totally homogeneous and even though the angle is not affecting the R2 value, it should be consistent so experiments are comparable. For black connectors with Light control in position 1, optimum binarization parameters are Min=30, Max =190, but if ambient light or colour of the connector varies they have to be optimized again until the pins can be clearly identified with black spots (as explained in figure 1.3). Ambient light affects the binarization parameters, so it is important that it is consistent.
In dim environments, or when connectors are white, the contrast in the image is not so vivid and this affects binarization. Thanks to the versatility of the developed system, for inspecting under these conditions, light control 2 must be selected and manual binarization parameters change to Min= 30 and Max = 210. Figure 1.6 illustrates binarization map of the same connector of figure 1-3, but with light control 2, and how it is modified when manual binarization parameters are modified to these new values

Figure 1.6: optimization of threshold binarization
Parameters when light selection is 2 (dark environments)
The diagram in Figure 1.7 shows the absolute effect of each parameter in the results for linear regression coefficient R2:

Figure 1.7: Absolute influence of parameter in the binarization results.
4 Conclusions
The developed Connectors’ Co-planarity Inspection System proved to be a successful tool to determine pins alignment for electronic connectors’ quality control testing. Since it uses a technique to recognize the pins position with pixel resolution, a linear regression coefficient can give absolute value of the deviation from a straight line correlation.
According to the results a linear regression lower than 0.8 could be considered as a bad alignment of the pins, although that has to be determined by the client according to their quality standards and production tolerances. A data base could be created for each connector tip using the explained methodology.
The system can successfully inspect all types of provided connectors regardless of shape and colour, because of the system’s flexibility for illumination and binarization parameters control.
Design of experiments analysis showed which are the most affecting experimental factors and conducted to an optimum value for each of them. By using Standard deviation calculations, the experimental error of this technique was also calculated.
5 Forward Steps
This technique conveys al requisites for the connectors co-planarity `test stated by the client, but can be improved in the following ways on customers request:
Hardware improvements
· Provide a mar con diffractive glass for accurate position of the conectors
· Z-axis automatic adjustment by use of a nema motor, for automatic focus of the lens.
· X-Y positionin system to do automatic analysis of several connectors at the same time (Lab-Robot technology.
Software improvements
· Include linear regression features into the software, to complement the position recognition and X-Y plots already included, to give a R2 value on the same report
· Creation of a Data-Base library with the production tolerance values of each connector, so the program can make automatic reference to it, and by comparison, decide if the inspected connector fits to the desired quality parameters.

