AR Lifetest Fixture
Another project that I worked on during my time as a Hardware Engineering Co-op at Amazon Robotics was a test fixture to determine the life cycle of one of the station components. It is important that these stations survive the stress demands of repetitive use and that the anticipated replacement times for components are known in advance of component failure.
Background:
One of the AR stations utilizes stop pins to absorb the impact force of incoming racks. These pins get impacted on all sides and it is important to conduct extensive failure testing to make sure they can withstand these impacts for the duration of a full life cycle and to forecast a recommended replacement schedule for these pins to prevent component failure.
Stop Pins on Back Plate
Project Goal:
I was tasked with designing a test fixture that would repeatedly impact this stop pin in a similar manner to how the pin would be impacted in the field. For this test, the pin would be impacted from the side as shown.
Calculations:
Prior to actually designing my fixture, I had to calculate the criteria of the test profile to ensure that the fixture would closely resemble the actual environment that the pin would experience during its lifetime application in the field. I began by calculating how many impacts would need to be tested. Based upon team testing prior to my involvement, it was determined that the pin was impacted laterally on about 15% of the total expected life cycles. Knowing the induction rates of each station, I was able to determine the total inductions within the defined lifetime and calculate how many of those inductions would cause a lateral impact to the pin. Next, I utilized known information about the speeds and accelerations of the use case to determine the speed of the pin and using the tolerance stack up of the system, I calculated the deflection of the pin.
*Note: Due to confidentiality reasons, I cannot give these exact values.
Pin Motion
Design:
The next step was to design the fixture for the motion shown to the left. For my design, I utilized a tie-rod air cylinder mounted perpendicular to the pin to pneumatically push the pin. I designed an end piece for the arm that would create a larger surface to contact the pin and that would allow the pin to slip past and oscillate back to its equilibrium point similarly to how it would perform in the field. It was designed to work bi-directionally to optimize the number of impacts we could simulate within the duration of the test. I also designed a mounting structure to be able to securely mount the pin assembly by itself for testing without requiring the full setup of the station.
Once I had all the materials, I began building the physical components of the fixture and ensuring that everything was properly aligned and tightly fastened in place. With some guidance from a mentor on the hardware test team, I fully wired the system together to be able to run a built profile so that the system would be able to operate automatically without the need for manual inputs. I worked cross-functionally with our software team to define the parameters needed for the profile to run properly.
Test Fixture Assembled
Wiring of Electrical Components
Challenges Encountered:
During this process, there were a few challenges encountered with the fixture as well as some lessons learned for similar projects in the future.
-
Increased Lifecycle Count: Further testing revealed lateral impacts on 100% of cycles instead of 15%, increasing cycle count over six times
-
Rotation of Piston Arm: Torque on the piston arm caused it to rotate around creating different impacts and deflections on the pin component. An 80/20 frame was built to compensate, but due to the nature of 80/20 and rectangular frames this structure tended to rack significantly. Wood block was added to stabilize, but this solution was limited in how close it could be placed to the arm
-
Pin Sticking to Block: The pin would often get stuck on the block instead of slipping past it. This stopped the pin from snapping back and oscillating to equilibrium as it would in the real world. Time delays were added to allow for more time to snap back as well as fillets on the end block to decrease the surface area that the pin could get stuck on; however, these remedial measures only worked as temporary solutions
​
Final Design:
I was able to collaborate with some engineers on the hardware test team and eventually we decided to use a large circular washer for the contact point instead of the block to maintain the right contact distance on the pin and to eliminate any effects caused by the rotation of the arm. The circular shape also helped to address the slipping as the washer was much thinner so it could more easily pass by the pin and because it was also rounded, the pin could more easily slip below the washer in passing.
Piston Arm Rotation
Filleted Block with Wood Path Guide
Pin Sticking on Piston Block
Final Fixture Running with Circular Washer Attachment