AC Manufacturing Automation | Material Handling & Vision
Over the years, as Valve expanded their Steam product line, AC had the privilege of a front row seat. Through our Integrated Engineering Team (IET) approach, we supported the development of Valve’s Steam Controller and Linux-based Steam Machine console. On the heels of these projects, Valve was also preparing to launch its own first generation Index VR headset.
When they were ready to dive into full development of the Index VR system, our R&D Accelerator and Product Development teams jumped at the opportunity to compete and help them realize their vision within an accelerated one-year product development and market launch window.
While the headset hardware and firmware prototypes were taking shape, our Manufacturing Automation engineers were brought in to develop solutions for the headset assembly. Managing the assembly meant solving clear challenges with the flexible circuit sensor bonding process to achieve greater speed, accuracy, consistency, and quality assurance.
The Challenge: Transform an Arduous and Meticulous Manual Assembly of Flexible Circuits
Each molded VR headset included a front and rear case with multiple flexible printed circuits (FPCs) and numerous location sensors. The sensors in the headsets would detect precisely emitted laser light from external base stations (light towers) positioned proximate to the user. This sensor data enabled the VR system to process the time and position of the detected light and use it to triangulate real-time positions and to track user movement in three-dimensional space.
Each flexible printed circuit assembly included a set of up to twelve (12) sensors each that required highly precise bonding onto the inside of a front or rear headset case using two extremely small pressure-sensitive adhesive tabs per sensor.
Manual sensor bonding was a painstaking process which yielded too much variation in the part positioning of the finished assembly. A technician, equipped with fine tweezers, would painstakingly prepare each FPC and its individual location sensors by carefully removing the adhesive tabs from each sensor and maniuplating them onto the mounting locations throughout the inside wraparound of each molded front and rear case.
A completed assembly required three (3) different FPCs comprised of 32 sensors per headset. Needless to say, the process to manual place and bond each sensor was arduous, slow, and resulted in too much variability due to inherently imprecise manipulation.
The result was a high degree of difficulty coupled with an inefficient pace to ensure consistent and successful bonding of all sensors equally with a tight tolerance of <0.75 mm from the center of each mounting position.
Adding to the cumbersome assembly, was a critical lack of quality validation. The technician could not easily confirm that all sensors were accurately positioned. Precise positioning was essential to ensure that downstream manufacturing would not encounter sensor issues with the required headset’s VR coordinate data.
Have a Complex Manual Task You Need Automated?
The Solution: Automate Sensor Bonding to Control Part Position & Validate for Accurate Flex Circuit Sensor Bonding
Working in parallel with our PD team, AC’s Manufacturing Automation group identified key opportunities to transform the painstaking manual sensor placement and bonding process. By employing simple and effective mechanical constraints and mechanisms, our teams enabled the headset sensor assembly to be completed more accurately and efficiently. This also enabled a reduction in the number of technicians required to manage the process.
Flex Sensor Placement & Bonding Tool
AC’s design solution was a compact, multi-axis, and force-adjustable tool enabling the precise material handling and positioning of assembly parts.
Flex Sensor Placement & Bonding Tools were developed for each distinct front and rear molded case. The tools would enable a technician to quickly secure a headset case into a carriage assembly (A) with the corresponding FPC and sensors secured onto a lower nest (B) where the sensor bonding sequence would be carried out.
The lower nest would utilize an array of vacuum mounts and pneumatic actuators – a straightforward mechanical solution to enable precise energy control, actuation of the bonding sequence, and fine adjustability of the position of any given sensor if found to be misaligned because of minute deviations in manufacturing tolerances.
Control of FPC & Sensor Positioning
To address the most complex part position challenge, AC designed a convex nest with an array of vacuum and pneumatic actuators to achieve some critical requirements:
> Enable the technician to consistently and efficiently place the FPC across the nest matching to the geometry of the interior of the case
> Hold individual sensors securely to the nest using vacuum and two datum pins in exact positions within a tight tolerance enabling the technician to remove the adhesive liners before operation
> Enable simultaneous pneumatic actuation of each sensor’s pressure-sensitive adhesive tabs onto the molded case part to ensure full contact and consistent bonding of all FPC sensors
Coordinated Vacuum & Pneumatics of Sensors
A simple, clever approach to the coordinated vacuum-to-pneumatic sequence was developed by AC’s Senior ME, Peter Van Tamlen. From an early prototype (shown), he developed the control of both energy mechanisms by using a brass mounting pad for each sensor which was actuated by an inner vacuum tube nested within an outer pneumatic tube.
This enabled our controls engineers to configure exact vaccum to hold each sensor securely in place in its individual position on the nest before the pneumatic extension of the brass mounting pad bonded the sensors to the interior of the molded headset case.
Adjustability of Flex Circuit Sensor Positioning
Because of potential manufacturing tolerance variances, the vacuum housing was designed with two (2) adjustable hex screws at each sensor position to enable a technician to fine tune the X- and Y-axis placement of each sensor. Furthermore, the XY position of the tool base and the angle of the nest were all designed to enable fine adjustments. Valve’s technicians could ensure ongoing accuracy of the position of the tool nest and the sensors as their assemblies evolved.
IO & Energy Control through CODESYS
To monitor and supply precise vacuum to all 12 positions across the bonding nest, our team leveraged CODESYS automation controls platform to configure the system. AC Controls Engineer, Chace Hutchins, recalls:
“Any loss of vacuum would fail the bonding of one or more sensors. This led to time-consuming manual inspection, failed product, and rework. So we needed to really dial in the dwell for each sensor position on the nest. CODESYS enabled us a develop a scalable IO solution to control the timing of each step in the operational sequence at a much more affordable cost than a conventional PLC Control System.”
Control of Case Positioning
For precise positioning of the molded cases of the headset, tool design also included an upper carriage assembly built to slide along linear rails. The case could be quickly placed into the carriage (A) and clamped securely before lowering it down the rails and over the nest. To protect the delicate FPC and sensors in the lower assembly nest from connecting with the case at it was positioned, the team designed a horizontal cylinder (B) for automatically splaying the case open once the carriage clamps are engaged.
The technician would then disengage dual locking pins to slide the carriage down to a preset stop height. (C) The carriage was designed to cam out at 90º rotation (D) and remain splayed open slightly to avoid contact with the sensor bonding head and FPC before returning to its natural shape for the bonding sequence. The case is clamped down at a precise distance above the sensor head. (E)
With the case secured over the nest at the preset position, the bonding sequence was initiated. The tool would sense the correct position of the carriage and nest, actuate the brass sensor housing, and press all sensors simultaneously into the interior of the molded case.
After the specified time, the tool would disengage the pneumatics and the technician could release the carriage to remove the completed case.
The completed carriage, linear rails, and splaying mechanism provided effective mechanical solutions that enabled consistency, repeatability, and accuracy for the handling of the headset’s front and rear case parts to enable successful bonding of sensors to the FPC on the nest.
Completed Headset Cases with Flexible Circuits and Bonded Sensors
Completed case parts were then visually inspected by the technician to confirm the part conformed to key criteria:
> All FPC sensors are fully attached to the interior of the case with both adhesive tabs secured
> All FPC sensors are centered within the tolerance window of its position on the case
> No sensors are protruding outside of the tolerance window
Vision Inspection & Validation of Bonded Assemblies
With a sucessful and repeatable part placement and bonding process completed, the next challenge was validation.
Quality assurance meant the ability to measure the accuracy of the positions of the bonded sensors, which were required to be within ±0.30 mm of the average for all six related case parts.
Keyway Case Positioning for Precise Inspection
The team determined that sensor bonding should be followed by a rapid part inspection enabling a technician to quickly load the headset subassembly into a nest and initiate an automated visual inspection sequence.
The tool would need to enable precise placement and keyway positioning of each front or rear case.
> Nest Assembly (1) included locating pins for fixed positioning of cases (2)
> Cover Assembly (3) latched fully to enable vision inspection within the interior of the tool (4)
Lean on AC Engineering Experts to Develop
Custom Manufacturing Automation
High Sensitivity Cameras for Validating Flexible Circuit Sensor Position
AC’s Automation Controls team devised the Flex Sensor Inspection Tool with a vision inspection system that would analyze and validate the position of the bonded sensors attached to each completed case.
Principal Controls Engineer, Steve Frankovich, who led the team’s development of the tool, recalls how the design decisions emerged:
“We considered using a robot to point a single camera at the multiple positions of the bonded sensors. But it turned out to be less costly and more efficient to connect seventeen (17) high sensitivity cameras through USB3 using the OpenCV vision library for image processing. We could measure the location of each bonded sensor at once and analyze the xy position of each sensor almost instantly. It was a better fit for our client all the way around.”
The highly sensitive FLIR Backfly CMOS cameras were configured inside the inspection tool to capture the precise positions of each of the sensors bonded to the case.
By leveraging the cameras light sensitivity in the IR range, the tool’s vision inspection system could accurately capture and analyze the position of each FPC sensor.
Any XY deviation from center that exceeded the allowable tolerance would be flagged in the composite image to identify a match score.
The inspection tool could then accurately validate the results to ensure the quality assurance criteria was met:
> Each sensor must have a score greater than 0.5 for the technician to validate the accuracy within the bonding window.
> Any score less than 0.5 is flagged by an orange outline on the image to enable the technician to further review the image before qualifying the entire assembly.
Overcoming Flexible Circuit Automation Challenges
Through a highly collaborative approach, AC’s Automation engineers devised a novel solution to this uniquely challenging sensor bonding material handling and vision inspection project.
Our engineering teams were able to deploy tools for Valve’s VR headset assembly that would effectively:
> Improve the handling and assembly speed of the parts
> Reduce the number of technicians
> Reduce the cycle time
> Ensure reproduceable consistency
> Improve placement and bonding accuracy
> Enable fine adjustments of tooling
> Validate sensor alignment accuracy
Valve definitely planted their flag in the VR industry. VR headset use has grown year over year, with the number of monthly-connected headsets on Steam passing 3 million last year.
AC is excited for Valve’s ongoing success and continued technological leadership. But theirs isn’t the only success story to tell. What about yours?
The future is happening now with VR. AC’s at our best solving complex product development and manufacturing challenges. We can help you lead with your technology and innovation in this dynamic market. Connect with one of our engineers today.
> Do you have a manual process in need of transformation?
> Are you looking to reduce labor and assembly time?
> Does your assembly require validation for greater quality assurance?
Have a Complex Manual Task You Need Automated?