Issue link: https://iconnect007.uberflip.com/i/1511625
12 PCB007 MAGAZINE I NOVEMBER 2023 the facility data lake. In this way, at imaging we can use the offsets assigned to each code from the error between the CCD camera and the mechanical or laser process to gain a few more microns as well when placing the image. Measurement of Fiducials Fiducial quality, design, and placement are critical to ensuring that the measurements are representative of the PCB. As a general rule, each fiducial position should have a minimum of four targets which can be analyzed, as well as a back-up set of four in complementary offset positions in case of fiducial damage detected by a low quality score. Many shops still use one hole or one fiducial and this is like designing a process with a single point of failure. Of critical importance is the camera/sensor used for target acquisition. White light, dual- wavelength, and X-ray measurements are all currently used in both dynamic-scan and fixed camera modes. Each sensor/camera has a known error, and the sensor type/quality must be matched to the product tolerance require- ments. It is a common issue in the industry that the wrong sensors are used for the wrong applications because fab process engineers fail to convey the detailed product requirements to the suppliers to make sure the appropriate Here are the general design considerations which process engineers should consider for each category. While this is not an exhaustive list, it does cover the primary variables of con- cern. Also included is a list of the eight most common best-practice hacks which the vast majority of fabs have not adopted, yet which have been proven to offer high ROI in the improvement of registration system perfor- mance. Note: ey typically will cut registra- tion error by >50% if all are incorporated. 1. Compensation and Measurement Steps Compensation and measurement process steps consist of three general sub-process steps. Environmental stabilization of product Stabilizing the product prior to compensa- tion and measurement is an oen-neglected topic. Failure examples here typically include processing through drilling or imaging when the panels are still warm from either prepreg lamination, dry film lamination, or in-pro- cess baking/drying. In the newest factories, we typically include timed buffers in the pro- cess kit aer each heat excursion to facilitate stabilization. Aer plating, it is important to do a copper-anneal bake since there is a risk that, if it is not done and the job is moving quickly through the facility, that the anneal- ing/de-stressing of the deposit will occur in the lamination press under pressure, leading to unpredictable scale factors. Hack: CCL serialization/pre-tooling of imaging targets One popular hack for addressing stabiliza- tion controls for CCL, as well as providing for individual traceability of each core, involves the drilling (by mechanical or laser) of through- hole 2D codes and de-scaled imaging fiducials in each CCL. By so doing, the fiducials act as go/ no-go gages to confirm whether the material is at nominal position prior to imaging. Some sites also measure the fiducials aer placement on the same machine and save them to the 2D code in