Can You Have It All When Producing Displays?

Can you have it all when producing displays?

A case study examining display production that optimizes quality and throughput – and lower total cost.

By: Travis Schneider

Business Development Manager, Aerotech


About the Author

Travis Schneider is Aerotech’s business development manager for advanced manufacturing market segments, including electronics manufacturing, laser processing, medical technology, data storage and precision manufacturing. He has 13+ years of experience in precision automation and robotics, holding roles in applications engineering, field sales, product management and business development. Travis earned his bachelor’s degree in Mechanical Engineering from the Milwaukee School of Engineering. His expertise and passion for innovation make him an invaluable resource for partners seeking to push boundaries in precision automation for advanced manufacturing.

An old adage states: Cost, quality, speed — you can have two. This statement addresses the fundamental tradeoffs one often has to make when evaluating purchase options. For example, anyone who has shopped for an automobile knows that a durable, well-constructed vehicle with class-leading 0-60 time often comes at a high price point. This cost/quality/speed balancing act applies to buying most goods, and often the summation of these three characteristics translates to value. Buying decisions usually happen when consumers feel they’re getting good performance and good quality at a reasonable price.


Of course, it’s natural to want it all – performance (often tied to speed or throughput) and quality at the lowest cost possible. This is a significant challenge facing the world of electronics manufacturing. Consumers have become particularly sensitive in seeking high-quality products, demanding higher performance (e.g. higher screen resolution, faster update rates, etc.) while also being cost conscious. In short, they want the best value for their money.


How can a mobile display manufacturing company meet this pressure to deliver the maximum value to customers while keeping manufacturing costs as low as possible? The following case study shows how optimized motion control solutions and superior mechanical hardware design can significantly increase a laser cutting process’s speed and tracking performance. All of this results in a throughput and quality boost that can significantly impact the bottom line of any business cutting precision parts from a glass substrate.


Display Cutting Considerations

When cutting small and large format displays for mobile phones, tablets or laptops, manufacturers often use high-precision automation systems coordinated with laser scan heads to enable precise, high-speed cutting of the screen material. According to one source, an estimated 1.2 billion mobile phone displays were produced worldwide in 2020. This translates to roughly 38 mobile device screens being produced every second over the course of that year.



Meeting this application’s demands requires the highest performing automation systems. This includes not only mechanics and optical systems with high geometric and dynamic performance, but also high-performance controls that take high-precision feedback sources into account. The combination of these elements enables manufacturers to tighten part quality specifications and boost throughput without sacrificing process yield.


Optimizing display cutting performance as described requires taking a complete systems approach to integrating a machine’s motion elements. These motion elements, shown in Figure 1, work in coordination to move the substrate and the beam relative to each other and control the laser’s firing along the commanded path.


Figure 1. The four primary motion elements of a high-performance motion control system used in display cutting.

When cut speeds can exceed five meters per second, the proper design, set up and coordination of these elements significantly impacts process yield. Inadequate machine base stiffness,

under-sized motors, laser scan heads that lack dynamic capacity, and insuficient resolution at the intended optical working distance all lead to process inconsistency and reduced overall capability.


Because the selection and integration of each of these elements affects the machine’s total performance, a systems approach is essential for maximizing performance (throughput) while maintaining cut quality and lowering production cost.



Case Study: Mobile Phone Display Cutting


Design Tradeoffs

To explore the real-world implications of design decisions made when integrating components into a display cutting machine, the following example compares the effect of varying just one component in the machine’s design – the laser scan head.


This machine’s high-level goals are as follows:


1.             Cut as many displays as possible, as quickly as possible.

For any given setup, the machine should cut as many displays as possible to optimize its throughput of displays cut per setup. The machine should make these cuts as quickly as possible, minimizing the cut time per individual display. The key metric here is displays cut per unit time.


2.             Maintain a high degree of quality in the cuts performed.

Achieving mobile phones’ sleek and smooth features requires tight part tolerances. This typically means individual device components are tightly coupled to one another. The key metric for this goal is how well a certain tracking error limit (the amount of allowable deviation the laser beam can make when following the intended cutting path) is maintained.


3.             Reduce the cost per display cut.

A persistent goal in the market is to deliver the highest performing, highest quality devices at the lowest possible cost. While many factors impact cost (energy rates, wages, etc.) this case study will focus on display unit cost as a function of the machine’s output without exceeding quality specifications as a key metric of interest.


This case study will consider the relative performance of a system that must meet all of these objectives when cutting a fictitious mobile phone display. While this exact display and its features are not real, the performance data of the profile this screen represents is, and it was collected directly from Aerotech’s Automation1 Visualize workspace.


This particular display has eight radii that must be cut with a tracking error requirement of < 12 microns (µm) peak to peak1 over the entire outline of the display (see Figure 2 below). From a throughput standpoint, the goal is to cut as many displays as fast as possible. This example starts with an objective cut speed of five meters per second (5 m/s).

Figure 2. An example of a mobile phone display and the eight radii cut during production.

Laser Scan Head Performance Considerations

A laser scan head precisely and quickly positions a laser beam in relation to the workpiece to cut various media – in this case a display. Minimally, a laser scan head includes a pair of galvanometer motors with mirrors attached. These are essentially rotary motors that deliver very high dynamic performance. These motors also use high-resolution digital encoders so their angular displacement can be controlled via a servo control loop.

The mirrors guide a laser beam through an F-theta lens that linearizes the angular position of the mirrors to displacement of the beam over the lens’s field of view, as shown in Figure 3.

Figure 3. A simplified image of a laser scan head’s working premise.

Figure 3. A simplified image of a laser scan head’s working premise.


1Peak to Peak (pk:pk) directly implies that the error of the entire system will not exceed this for any part made or 100% of all parts must fall within this spec.


This gives the laser scan head a variable travel range directly correlated to the F-theta lens’s working distance. For example, a longer working distance will result in a larger field of view, and a smaller working distance will result in a smaller one. The laser’s wavelength and the input beam’s diameter will also affect the working field of view of any lens and the resulting spot size. However, this case study is looking purely at the system’s field of view and its working distance to see how they impact throughput and quality.


As a laser scan head’s field of view increases, the beam’s travel distance for a given mirror movement will increase in distance and speed. Accommodating the spot size change can have significant positive impacts to a single scan head’s cutting capability in terms of display size and the number of displays per unit time. It also affects the number of scanners, stages, lasers and other equipment infrastructure needed to achieve a specific throughput target (i.e. a single scanner with a big field of view is more cost effective and has more throughput capability than multiple scanners with small fields of view).


However, there is a price for this, and it’s related to the following three factors that impact all laser scan head systems:


1.             All laser scan heads use galvanometers, which have limited travel and spatial resolution to quantify the angular mirror position.

A laser scan head generally has better positioning resolution the closer it is to the substrate. The encoders measuring the motor’s position have very real limits to how small a movement they can detect in a specific time period. When moving away from the substrate, one loses detection resolution and must detect that motion in a shorter time period. This metric is called dither, or the motor’s minimum incremental commanded motion. The smaller the dither, the farther away the laser scan head can accurately position itself.


2.             All laser scan heads have limitations on how quickly the position loop can be closed.

All servo motors rely on current being applied and removed in precise timing to drive the motor to its intended position. As the speed increases for a given path, the electronics and software performing this position loop tracking operation will hit a limit, meaning the beam is moving so fast that it struggles to know if it’s on its intended path. This metric, called the servo control rate, is directly proportional to the commanded speed. There are only so many available points per second at a specified speed to indicate if the laser’s intended position is on the commanded path or not.


3.             All galvanometer motors (and mirrors) have mechanical limitations to how quickly they can position before exciting resonant frequencies, which causes position error.

As the speed of a specific distance move increases, the amount of time allotted to move its load must decrease. This translates to faster accelerations and decelerations, and just as in any mechanical system there is a limit to what the load can weigh and still be light enough to make the move in the time allotted. This may be due to limitations in the mechanical system’s physical rigidity or the motor not being powerful enough to carry the load without


excessively overshooting its intended target. This limit is often referenced as a ratio of motor torque to inertia. The mirror can make higher performance moves as this ratio increases. This is critical in the example depicted below because making faster, longer moves in a shorter time period will cause the motor to approach its natural resonance and produce larger positional errors.


These three factors will be considered in the following example, which correlates their impact on performance with real data produced by two Aerotech laser scan heads.


Mobile Display Cutting Example

This case study compares the relative performance of two Aerotech laser scan heads

(AGV20-HP(O)-2 and AGV20-XPO-E2) according to the threshold and objective goals established for cutting the desired display. For easy reference, these goals have been translated into Table 1.


Table 1. Threshold and objective goals for a display cutting application.


Goal typeCategoryValue
Threshold (must achieve)QualityPath errors of <12 microns
Objective (maximize)ThroughputDisplays cut per unit time

(initial target of 5 m/s cut speeds)

Objective (minimize)CostCost per display cut


Given the required cut profile, radii 2-7 present the greatest dynamic challenges to the machine (see Figure 4). This study focuses on these radii, assuming a focal length of 250 mm and a targeted cut speed of 5 m/s.

Figure 4. An example mobile phone display showing eight radii that must be cut. Radii in the highlighted region are the most challenging to cut.

Both laser scan heads used in this comparison deliver industry-leading accuracy and overall performance. However, the AGV-XPO configured for this application has been optimized for precision and dynamic performance. Its low-inertia motors enable high acceleration, and its extremely high resolution feedback minimizes dither. Table 2 provides a brief comparison of these two products.

Table 2. A relative comparison of the performance of the two laser scanners in question.



Feedback Resolution0.012 µrad (25 bit)0.00016 µrad (32 bit)
Dither (minimum incremental motion achieved)10.4 µrad0.02 µrad
Peak Acceleration2,380,000 m/s288,000 m/s2
  1. Without -AC air cooling
  2. Typical performance with f = 160mm F-Theta
  3. Based on the motor’s maximum rated


The graphs in Figure 5 display the real peak to peak error observed when cutting this example profile at a speed of 4.36 m/s with both the AGV20-HP(O) (blue) and AGV20-XPO (orange).

Figure 5. The resultant error in µm witnessed in X and Y dimensions across the two laser scanners examined at a commanded cut speed of 4.36 m/s.

This data indicates the AGV20-HP(O) exceeded the threshold limit for error of 12 microns and would have to be further refined to meet the precision requirement. Alternatively, further reducing the commanded cut speed could minimize following error and meet the threshold limit.

However, the AGV20-XPO met this requirement – with margin to spare – at a cut rate 4.36 m/s. Given that the threshold requirement includes some overhead, examining the objective requirements shows just how far it’s possible to push throughput and minimize cost.

Increasing the focal length can increase the speed at which the beam traverses the work piece. This increase in dynamic performance will come at the expense of precision, but again there is additional margin on the precision performance that can be exploited with this laser scan head.



Increasing the laser scan head’s focal length makes the beam move faster over the workpiece per a given angular displacement of the mirror. This allows the laser scan head to address a larger field of view. Again, this will have an impact on the precision of the beam. However, if the cuts’ precision can still fall within the threshold requirement (<12 microns), the advantage of this longer focal length allows for processing more parts at a faster rate. This improves the machine’s throughput, and minimizing the time required to produce a given display lowers production cost.

Using the same cut profile as outlined above, a different optic was selected to extend the focal length from the laser scan head from 250 mm to 500 mm. The system’s field of view and speed were increased, and its performance was documented. The speed was increased by roughly a factor of 2, from 4.36 m/s to 8.73 m/s, with the resulting error data captured in Figure 6.

Figure 6. The resultant error in µm witnessed in X and Y dimensions across the two laser scanners with a longer FOV optic and higher commanded speed of 8.73 m/s.

The Takeaways: Quality, Performance and Cost Delivered

So how does this all translate to the three core goals outlined earlier? Using the data above, some assumptions about the machine’s performance will help to project the potential impact of optimizing the system design. Assuming it takes 100 milliseconds to cut a single phone display at

4.36 m/s, this machine’s time costs $0.05 per display. Table 3 provides a summary of these requirements along with baseline performance.


Table 3. A summary of the baseline system performance by laser scan head.


RequirementsLaser Scan HeadResults

Path errors of <12 microns

AGV20-HP(O)Failed: 46.4 micron path error
AGV20-XPOAchieved: 9.3 micron path error


AGV20-HP(O)Not Applicable – did not meet quality threshold requirement.
AGV20-XPO10 displays per second


AGV20-HP(O)Not Applicable – did not meet quality threshold requirement.
AGV20-XPO$0.05 per display


Given the relative performance of the two laser scan heads, the AGV20-XPO was the only one capable of producing displays within specification at the desired cut rate. Optimizing the machine to use an optic that allows for twice the focal length brings a number of advantages. First, doubling the focal length from 250 mm to 500 mm quadruples the scan head’s work area. Second, the processing speed was doubled from 4.36 m/s to 8.73 m/s, meaning that twice as many screens could be processed per unit time.

As a result of processing twice the displays in the same, the cost per display as a function of machine time is also halved. Table 4 shows how this basic optimization compares against the baseline performance captured in the previous table.


Table 4. A summary of the baseline system performance as a baseline and throughput optimized configuration using the AGV20-XPO.

RequirementsLaser Scan HeadResultsImprovement

Path errors of

<12 microns

BaselineAchieved: 9.3 micron path errorMeets specified requirement
Throughput optimizedAchieved: 10.5 micron path error


Baseline10 displays per second2X increase in display throughput
Throughput optimized20 displays per second


Baseline$0.05 per displayHalf display cost relative to machine time
Throughput optimized$0.025 per display


Granted, there is a catch here. The AGV20-XPO is a higher performing product, and it’s priced accordingly. However, the manufacturing cost advantage the AGV20-XPO brings will very quickly make up a differential in purchasing cost. The proof is in the data – calculating the number of displays produced by the baseline system relative to the throughput-optimized version with the AGV20-XPO clearly shows the advantage.

If the goal is to produce 250 million displays per year and operate these machines 2080 working hours (7,488,000 seconds) per year, using the baseline assumptions above it would require four machines to meet this goal. However, using the throughput optimized version of this machine would halve the cost to produce displays (saving $6.25 M following the cost assumptions presented above). Moreover, the capital expense of two machines would be saved, along with the energy and staff required to operate these machines.


The Aerotech Advantage: You Can Have It All

Through careful analysis and attention to detail – yes, you can have it all. Designing a precision laser cutting machine can be challenging. There is a delicate balance to strike when maintaining quality, maximizing machine performance and lowering cost. It requires an understanding of multivariate design tradeoffs that exist within demanding applications. How is a small design team supposed to manage this in addition to day-to-day operations?

Aerotech is here to act as an extension of your team, ready to bring 50+ years of motion control experience to the table. We obsess over the hardest design problems and don’t stop until we’ve optimized our customer’s outcomes. Our talented engineering teams work collaboratively to provide data-driven recommendations for achieving your goals and requirements. These teams understand the technical nuances of complex applications such as the one presented here, and they have deep expertise in many applications. Our teams work across all disciplines (mechanical, electrical, software) with a systems-focused approach to ensure we’re delivering solutions that are optimized for your needs.

Aerotech’s motion control products are world-renowned for delivering the highest performance and quality available. Our products are used in the production of the most mission-critical manufacturing operations across the globe and are the standard when precision performance is non-negotiable. We often leverage our trusted standard products – including multi-axis motion controllers, drives, motors, stages, laser scan heads and software – to produce complete motion solutions that are controlled in a single environment. Our products are tested and verified according to stringent standards to guarantee performance. When further optimization is required, we work alongside our customers to produce tailored solutions that meet their unique requirements.

If you have a challenging precision automation application and haven’t found the right partner to serve your needs, we’re ready to help. In addition to our in-house design and manufacturing teams, we have field sales engineers located worldwide. No matter where your machine ends up, there is likely an Aerotech expert in close proximity.






Braskem continues investing in the development of its biopolymers in Asia and just started its Tokyo Representative Office
JTC anchors two A*STAR research institutes at Jurong Innovation District to grow advanced manufacturing ecosystem and launches new Sectoral Job Redesign Initiative to support businesses in I4.0 transformation