Reading time ( words)
A Digital World Running on Analog Technology
Our world is a digital world, and it has been for a long time now. Personal computers have been around since 1980, The Internet has seeped into almost every aspect of our lives, and most of us now carry mobile devices that have the computing power to put a man on the moon.
Imaging has been taken over by digital cameras--a welcome change from the days when we had to wait to develop film and print pictures. We can now delete bad photos immediately, print good ones at lower costs, and share them within seconds using the communication platform of our choice. Even the old light bulb is being replaced with less power-hungry LEDs, which are all mounted on a PCB carrier.
So why are 95% of the circuit boards that drive all of these digital devices still imaged with film, the same way they have been for the past 50 years? This white paper will address this question by exploring the challenges of digital PCB imaging and the emergence of inkjet imaging, a technology that has the potential to revolutionise the industry.
Digital PCB Imaging: A Brief History
The first digital imaging in the PCB world was introduced 15 years ago and known as laser direct imaging (LDI). Since then, a very slow (and costly) revolution has been brewing, with LDI driving smaller track and gap resolutions. Most recently, advances in light-emitting diode (LED) direct imaging have threatened LDI’s hold on the market. But the basic imaging process is still based on a photo sensitive resist that is exposed and developed in a multi-step process requiring an incredible amount of infrastructure. If we draw an analogy with the graphics industry, we can see that digitisation also started with laser printing. This laser-based technology was subsequently replaced by LED and years later by inkjet, with direct digital imaging technology replacing all but very high-volume offset (with variable data also making a place for inkjet in this area).
A clear innovation pattern can be seen across the printing industry:The key drivers behind this innovation pattern include:
- The move to full digital data processing and printing;
- The need for short turnaround and lead times;
- Lower plotting costs and no waste of film and screens; and
- Lower running costs and less labor.
The advantages are clear. Why then has this revolution not taken place in PCB production? It’s not because companies haven’t tried. In fact, inkjet has penetrated the PCB world with legend printing, where the quality demand meets the technology available.
Inkjet for PCB Imaging
Imagine printing an etch resist directly on copper. You get to save material (the entire surface is no longer covered with photo sensitive resist) and most importantly, the development step is eliminated. You can even envisage a future in which conductive tracks are printed directly without the need for massive infrastructure and chemistry. It sounds great, doesn’t it? Why then are we not imaging inner and outer PCB layers with inkjet?
The short answer is that it’s really difficult.
The jetting process of inkjet, suitable for graphic applications, is not reliable enough for PCB imaging. There are also conflicting demands. The ink must be robust in the processing applications, it must adhere to copper, it has to be resistant to etching (alkaline and acid), compatible with tin plating, and of course it should be strippable.
Typical inkjet text on the left compared to PCB lines at the same magnification.
Inkjet also has a number of serious drawbacks which can lead to failure in the drop formation process:
- Air entrapment in the nozzle due to high-frequency jetting;
- Nozzle plate wetting conditions;
- Clogging of nozzles; and
- Drop deformation and satellite formation (small drops next to the main drop).
To put it simply, the biggest drawback is that inkjet jetting is not a 100% robust process. It’s not a question whether inkjet drop formation will fail, but when. This can be disastrous for an imaging process used for PCBs. It can hamper jetting, leaving areas open or unfilled, and it can spray drops at different speeds, resulting in drop displacement (spurious copper). These jetting problems also come without warning, leaving the printed substrate unusable.
If we manage to overcome these jetting problems, inkjet could replace 11 out of 15 process steps used with contemporary lithography. The whole process would then be reduced to four steps: A cleaning/light pre-etch of the copper substrates; imaging with inkjet; etch; and strip. This reduction in processes leads to higher yields, fewer registration issues, fewer costs, and full digital production with process time from CAM to etch in less than five minutes.
Inkjet could replace 11 out of 15 process steps used in contemporary lithography.
The advantages inkjet brings will drive industry adoption in time. The real questions are, how and when?
The jetting properties of inkjet today will not perform at the required level, even though current reliability is about one error per billion actuations. With more than two billion drops needed on a panel, there is a 100% chance of failure. Jet instability will therefore remain an intrinsic inkjet problem.
The Work-around: Detect and Replace Failing Nozzles
Even when using state-of-the-art heads and inks, the only possible remedy for inkjet based PCB imaging is not to solve the jetting issues, but to work around them. This is accomplished by measuring the jetting performance and switching off nozzles with performance degradation before they can cause misprints. Mutracx’s Lunaris has this technology built into its system. The performance of every nozzle is continuously measured in every print-head channel. How is this done? The piezo used for channel actuation is also used as a sensor. Using this sensor response mechanism, the system checks the status of the channel prior to using that nozzle for jetting. The system can then anticipate which nozzles it can use at any given time. Nozzles not jetting correctly are then replaced, meaning that there is also redundancy in the system.
Lunaris has built in three heads per raster line. Mutracx has named this monitoring system “Predict,” as it predicts potential issues before they can cause failure. With 60 heads and 15,360 nozzles in total, the Lunaris printer jets approximately 20 to 50 million drops per second (depending on the image) and does half a million Predict tests per second, all in real time. The system also uses this information to monitor the overall nozzle condition in the heads and will trigger regular maintenance between jobs without operator interference.
Does this solve all problems for inkjet? Not completely. There is still the issue of creating an accurate image based on drops on the fixed raster of the print head. The distance between drops, known in the industry as the drop pitch, is physically determined by the nozzle pitch in the head, and is typically 120 microns. Lunaris has developed a clever print strategy that allows you to choose which raster drops will be fired while the stage makes steps perpendicular to the jetting direction. By applying different frequencies, Lunaris can also change the line width to achieve optimal contours.
The CAM image is analysed at a very high resolution (2.5 micron scale) and a print strategy model determines which nozzles should be fired to print any feature within a 5-micron boundary. The physical properties of the resist and resist flow on the panel are also taken into account in this model. Together with other system tolerances, within an image, we print every feature within a +/-15 micron envelope.
With this system, overall registration is very high due to the use of carbon fiber (with very low thermal elongation per degree Celsius), critical positioning components and dynamic calibrations based on Robax© encoder rulers. The machine itself is temperature controlled for thermal stability (clean machine). The result is a very high and stable overall accuracy.
There is another key difference with other imaging technologies and inkjet printing: The smoothness of lines and patterns. Every feature has to be built up from individual drops. There is always variation in the jet speed over time. While the system is calibrated within microns, a slight wave can be visible when printing lines perpendicular to the stage movement. To stay within a 10% line width variation, you need to drive the limitation in feature size to 100 microns post-etch (even though the wave does not alter the conductive width of the line).
Image Quality Inspection (IQI)
Lunaris has one more unique characteristic made possible by inkjet printing: Inline quality inspection. Typically, in an inner layer production line, optical inspection is done at the very end after all 15 processes involved in inner-layer production have been completed.
Because the system only prints resist according to the CAM image, it can also do a printed resist image inspection as an inline quality inspection prior to etch. A high-resolution scanner is used with optimized exposure to ensure maximum contrast with the resist. This image is compared to the CAM image and analyzed within the print cycle. In this way, 100% print quality is guaranteed into etch.
In case of a misprint, the panel can be stripped. This means the substrate (with the valuable copper) is not etched and can therefore be reused, eliminating waste.
Not If, But When: New Inkjet Technology Will Conquer PCB Imaging
As this white paper discusses, inkjet has proven its advantages in many industries. By addressing the weaknesses of the technology in a system like Lunaris, the PCB imaging industry can now harvest the benefits:
- A direct digital imaging process for PCB production, producing cores from CAM to etch in minutes;
- Full panel automation, making PCB production as easy as sending a print job to your graphic inkjet printer; and
- An advanced inspection system that paves the way for a new business model: charging only for prints that are well made.
There are still limitations in the current Lunaris system. For instance, it limits feature sizes to 100 microns post etch. But coupled with LDI, it can fully digitize a factory fit for all feature sizes. As heads and inks improve and drop sizes decrease, feature resolution will increase.
The system is setting a foundation for the adoption of inkjet in the PCB industry. The first fully functional systems are now being integrated into factory installs, and Mutracx is working on expanding its customer base across Europe, North America, and Asia.
Henk-Jan Zwiers is the chief technology officer at Mutracx. He is a strategic technology professional with a strong commercial and development record, boasting more than three decades of experience in the print and electronics industries. Before joining Mutracx, he led the development of several key projects at Océ, the global leader in the supply of print and document management products and services. His work and expertise have been crucial to the development of the Lunaris Project. Zwiers obtained a Master’s Degree in Mechanical Engineering from the University of Twente.