Automated Production Machine Determines Printer Head Quality Using Modular Instruments from National Instruments

Author(s):
Kurt Hensen - ES International

Industry:
Machines/Mechanics, ATE/Instrumentation

Products:
PXI-5102, PXI-2530, Vision Development Module, LabVIEW, Arbitrary Waveform and Function Generators

The Challenge:
Developing a machine that operates in an industrial production environment and inspects the industrial inkjet piezoelectric PZT (Lead zirconate titanate)material before final assembly.

The Solution:
Using National Instruments arbitrary waveform generators, digitizers, and switches to develop a tester that applies a waveform to each piezoelectric cell and measures the piezoelectric crystal acoustic response, which is a qualitative measure for the operation of the nozzle.

Print Head Background

Industrial inkjet print head units are typically very expensive. The print head has more than 1000 microscopic nozzles made from a piezoelectric PZT material and built with precision. A thin, base carrier is bonded to the piezoelectric material to form a long, rectangular strip. The piezoelectric material is then cut into slices each approximately 70 µm wide. Then, electrodes are attached to the piezoelectric material. Small holes are made in the base material to form the actual nozzles. These strips must be tested before assembling the electronics nozzle control and ink container.

This type of printer head can emit up to eight subsequent subdrops. These small drops are emitted quickly after eachother and while airborne, form one larger drop. By emitting more or less subdrop the volume of the final drop is adjustalbe and affects image quality and resolution. The printer head is directly responsible for the quality of the print job; therefore, the printer head, especially its nozzle, must be made well.

Resonance Testing

When a voltage is applied to a piezoelectric cell, it deforms and generates pressure on the ink, pushing it out of the nozzle through the hole. To build high-quality print heads, each piezoelectric cell is important, as is its response in relation to neighboring cells. The best way to measure this quality is to test the resonance of the piezoelectric cell. This is comparable to tapping a finger on a wineglass. When the glass is of good quality, it produces a nice ringing sound. When it is poorly made, or broken, it produces a dull buzz.

In theory, the quality of a piezoelectric cell is determined by providing a swept-frequency ringing waveform to the cell and measuring its responding resonance pattern. To apply this theory to practice, we performed a feasibility study using an NI M Series PCI data acquisition (DAQ) board in a desktop PC.  The advantage of the M Series DAQ board  was that we did not need expensive LCR meters, oscilloscopes, and function generators because the DAQ board can deliver the same functionality and integrates seamlessly into the NI LabVIEW development environment. Once we proved the theory could be put into practice, we developed a full-scale test machine, including printer head handling.

Production Test Machine for Print Heads

We developed a test machine that completely handles the transport of the PZT actuator to the test probes. The complete machine consists of 15 motors, four cameras, and a PXI measurement system. We used the PROFIBUS protocol to communicate between the motion drives and our software. We developed the software completely in LabVIEW.

We placed trays with up to 32 PZT actuators in the machine. A pick-and-place unit integrated into the machine took an actuator and placed it in a fixture. This fixture slid the unit under a camera to check the 2D barcode that describes the product serial identification number. We also used the camera to locate the first electrode on the PZT actuator. Because the piezo cells are only 100 µm apart, the camera inspection had to be very accurate to ensure that the actuator’s position of the contact electrodes was known exactly. We used the NI Vision Development Module to analyze the image and find the offset of the first electrode in relation to the position of all neighbor electrodes.

The computer moved the fixture to the inspection station. Four sets of test probes then touched the nozzle electrodes. Each test probe contained 32 needles to make direct electric contact with the 32 electrodes of the piezoelectric cells. Four cameras ensured that the test probes were in position and that the product was correctly aligned. If needed, each test probe could move independently of the others in two planes of direction to provide better contact. Because the actuator had a large number of contact pads divided into two sides, the cameras were also placed on a motion system to slide them from the front to the back of the actuator under test.

We used an NI PXI-5401 arbitrary waveform generator to generate a signal between 0 and 3 MHz to provide a stimulus signal to the piezoelectric cells. The signal type (sine, block, or arbitrary waveform), sweep range, and voltage level can be set up by the end user and may vary depending on the nozzle type.

Using four NI PXI-2530 128-channel matrix switches, we transmitted the test signal to the four test probes. We quickly switched the test signal over these 32 contacts, and within three seconds we could generate up to 64 different combinations. This is done simultaneously for all four test probes. Four NI PXI-5102 digitizer modules simultaneously measured on four channels the response from the piezoelectric cell at 20 MS/s. Then the modules streamed this data to memory for analysis. When the first set of 4 X 32 contacts had been tested, the test probes were moved to the next set of contacts.

We used LabVIEW to control the motion and the instruments and to conduct the signal analysis, such as determining frequency response and resonance frequencies. Pass/fail limits were based on these characteristics. When a unit completed the test, it was placed in either a pass or fail tray.

The application had two operator monitors: the first showed the current measurement results and the second showed the status of the machine. With these monitors, an operator directly controlled and supervised the machine. Results were sent over Ethernet to an enterprise database and logged for future reference.

Conclusion

Print head quality depends on the quality of the actuator unit. In a feasibility study, we proved that correct analysis of resonance frequencies can be used as an indicator for actuator quality. We developed this into a full-scale production tester using NI instruments controlled completely by LabVIEW, which provided the flexibility to adapt the tests easily to new types of nozzle units.

For more information, contact:
Kurt Hensen
ES International
Diepenbekerweg 10
3500 Hasselt
Belgium

Tel: +32 11 35 25 48
Fax: +32 11 35 25 49
info@es-int.com
www.es-int.com

 

Browse All Case Studies »