Reason for the automation
The Fischbach Group is a global leader in the production and printing of cartridges for the silicone and adhesive industries. The 94 million cartridges that are annually injection molded must be checked occasionally to ensure that an optimal product is delivered to their customers.
To do this, Fischbach takes daily measurements per injection molding machine on their cartridge to compare with preset tolerances. When they see that a particular measurement of a set of cartridges is out of tolerance, they can adjust the necessary settings on the machine. In the past, all these measurements were done manually which frequently caused errors and was very time-consuming.
Fewer errors
Faster quality control
More user-friendly
Introduction video
Introduction video
Introduction video
Up to 11 different quality measurements can be made per cartridge. In consultation with Fischbach, we looked at which measurements are most important and most commonly performed. For this purpose, we have developed a system that accomplishes these measurements quickly and accurately.
All remaining measurements were optimized and digitized. When determining deflection, one inserts the cartridge into a fixture and rotates it 360°, with help of a measuring clock our software will calculate the deflection.
Up to 11 different quality measurements can be made per cartridge. In consultation with Fischbach, we looked at which measurements are most important and most commonly performed. For this purpose, we have developed a system that accomplishes these measurements quickly and accurately.
All remaining measurements were optimized and digitized. When determining deflection, one inserts the cartridge into a fixture and rotates it 360°, with help of a measuring clock our software will calculate the deflection.
Performed measurements
All of these measurements are taken automatically in one cycle.
Determining of the cavity*
Height of the shoulder edge
Nose tip height
Height of the shoulder
The internal diameter at three heights
*There are 4 types of cartridges produced spread over 7 dies. Each die is marked with a letter or double letter in combination with the cavity number (1 to 16). This reference is found on the shoulder of the cartridge.
Performed measurements
All of these measurements are taken automatically in one cycle.
Determining the cavity*
Height shoulder edge
Height nose tip
Height of the shoulder
Inner diameters cartridge
Performed measurements
All of these measurements are taken automatically in one cycle.
Determining the cavity*
Height of the shoulder edge
Height nose tip
Height of the shoulder
Inside diameter at three heights
*There are 4 types of cartridges produced spread over 7 dies. Each die is marked with a letter or double letter in combination with the cavity number (1 to 16). This reference is found on the shoulder of the cartridge.
Recognizing the cavitation
One of the challenges for this project was to recognize the cavities located on the shoulder of the cartridge. Based on the cavitation, the operators know what type of cartridge it is and what mold it comes from. As a result, they know which injection molding machine to adjust any settings to.
One of the challenges for this project was to recognize the cavities located on the shoulder of the cartridge. Based on the cavitation, the operators know what type of cartridge it is and what mold it comes from. As a result, they know which injection molding machine to adjust any settings to.
One of the challenges for this project was to recognize the cavities located on the shoulder of the cartridge. Based on the cavitation, the operators know what type of cartridge it is and what mold it comes from. As a result, they know which injection molding machine to adjust any settings to.
Since the cavitation (here “PP6”) is very difficult to see, we chose to detect it automatically with a custom vision system.
The images are processed and classified with our own neural network, this with a 99% success rate. We also built the vision algorithm so that the operators have a wide zone in which to place the cartridge.
Since the cavitation (here “PP6”) is very difficult to see, we chose to detect it automatically with a custom vision system.
The images are processed and classified with our own neural network, this with a 99% success rate. We also built the vision algorithm so that the operators have a wide zone in which to place the cartridge.
Since the cavitation (here “PP6”) is very difficult to see, we chose to detect it automatically with a custom vision system.
The images are processed and classified with our own neural network, this with a 99% success rate. We also built the vision algorithm so that the operators have a wide zone in which to place the cartridge.
Measuring the inner diameters
Measuring the inner diameters
Measuring the inner diameters
The most important measurement for the Fischbach Group are the inner diameters of the cartridges. This operation had to be fully automated without any contact. To solve this, we used a laser probe connected to a stepper motor. We attached these two components to a vertical guide which allows us to insert the probe at different heights.
When the vertical guide brings the probe to the correct height, the stepper motor will rotate 720°. During this rotation the laser probe will continuously sends all the measured points to our system.
Finally, our own circular regression algorithm determines the inner diameter of the tube.
When the vertical guide brings the probe to the correct height, the stepper motor will rotate 720°. During this rotation the laser probe will continuously sends all the measured points to our system.
Finally, our own circular regression algorithm determines the inner diameter of the tube.
The operator can choose when to perform which measurements by using our GUI. All results are visually displayed and compared to their tolerances. Finally, these are forwarded to Fischbach’s own database so that they can be revisited later.
When the vertical guide brings the probe to the correct height, the stepper motor will rotate 720°. During this rotation the laser probe will continuously sends all the measured points to our system.
Finally, our own circular regression algorithm determines the inner diameter of the tube.
The operator can choose when to perform which measurements by using our GUI. All results are visually displayed and compared to their tolerances. Finally, these are forwarded to a local database of Fischbach and will be used for further analyses.
The operator can choose when to perform which measurements by using our GUI. All results are visually displayed and compared to their tolerances. Finally, these are forwarded to Fischbach’s own database so that they can be further analyzed.
