Test Automation of Sensor Characterization | Test for Design Validation, Part 2

test automation of sensor characterization

In Part 1 of this series Test for Design Validation, we took a look at how to solve unwanted noise interactions between closely located haptic force motors.

In this next blog, we explore another aspect of troubleshooting through iterative testing to narrow in on the underlying causes for unwanted sensor interactions. We also reflect on the big impact that early testing has on saving time and money by fine tuning performance even before a final prototype is done.

Test automation of sensor characterization enables our engineers and clients to better manage the volume and complexity of iterative testing required to discover underlying design flaws or reveal opportunities to improve a product’s reliability. Let’s explore the value of test engineering and how AC can help you rapidly improve hardware and software testing for manufacturing success.

Test automation for sensor characterization

Travis Tonder | Test Engineer
AC Product Development

When we sat down with Travis Tonder, it was easy to see why he succeeds as a Test Engineer. He described how he enjoys converting manual processes into automated solutions:

“I guess that’s just how my brain works. I want to make things better within the scope of what I do working with clients that manage large enterprises. Manual processes often get overlooked because there usually aren’t teams dedicated to improving them. So, I look for inefficiencies and opportunities to automate. It ends up saving the client time and money, while making testing more accurate as well.”

Troubleshooting Sensor Characterization

sensor characterization

Travis described one such test project he was involved with that required repeated and variable testing and data gathering to solve a sensor characterization problem:

“Our client encountered some unexpected issues with one of their wearable GameTech devices. An optical proximity sensor was triggering at the wrong time, causing unexpected responses from the device. The developer needed us to identify the cause and other potential triggers so we could fine tune the sensor to perform correctly.”

Small digital camera sensor plate in fingers
Small digital camera sensor plate in fingers

To reveal how the sensor was responding to different inputs, testing required 50+ different measurements sensor readings with 10+ targets, one at time, as he describes:

“We would set targets precise millimeter distances apart, take readings, then incrementally increase the distance and repeat the process again. Each time we moved the target to the next specified distance, it was important to keep its orientation as close as possible to the initial measurement.”

Accurately manipulating the targets for each measurement was not only time-consuming, but also subject to minor imperfections – issues that can throw off the readings and make data analysis less precise.

Adding further complication, different substrates were also used to determine deviations in optical detection, as Travis recalls:

“Each new criteria, whether to test a texture, color tone, or light level, required the same manual repositioning, requiring that we set up a test coupon, take a reading, change the distance, and repeat the steps again for each of the 50+ distances.”

Automating manual testing with robotics and scripting

As you may expect, the repeated manual steps required to set up, reposition, and retest meant the process was slow and arduous. Travis wanted to speed things up, so he decided to develop an automation solution.

He designed a test fixture with a gantry robot and Python scripting to take readings, move to precise coordinate distances, re-orient targets, and capture data.

Here’s the automation configuration and test process he landed on:

FixturingAutomation Configuration
3-Axis Gantry RobotA gantry robot is programmed with inputs that request the current position of robot and trigger it to move to next measurement position.
DAQData acquisition is controlled through the test script interfacing with the robot.
Control HardwareA laptop computer is used to write the robot control program and run the test script.
Control SoftwarePython scripting interfaces with and controls the system; selected for its flexibility and ease of implementation with third-party hardware.
Gantry Robot Fixture
Customized Automation Configuration
DAQ Interface
Automated Testing Process
1 – Affix test article onto robot base.
2 – Attach target onto robot end effector using a custom fixture.
3 – Set initial target location.
4 – Start testing.

The script controls all remaining steps:
> Take readings of the proximity sensor measurement and the exact position of the target
> Data is saved to a .csv file.
> Robot moves to the next preprogrammed increment
> Repeat steps for a predetermined number of times and increments
TABLE: Reworked testing process within the automated process.

Travis describes some of the key improvements that were gained:

“Once the configuration was developed, modifying a test was a one-and-done edit of the code. For example, if you noticed an interaction happening at distances between 5mm and 15mm, you might want to drill down and measure at 3mm in that zone, or even 1mm to get more resolution about what’s going on. It was quick and easy to change those parameters in the code than manually reconfigure the test at those distances. With test automation, the improvement in speed and accuracy was huge, easily increasing efficiency by 90%.”

Automation enabling running of 100 test points at 1mm increments, otherwise unrealistic with manual testing, both in terms of time required and measurement accuracy.

With test automation, the improvement in speed and accuracy was huge, easily increasing efficiency by 90%.

Early Test Automation of Sensor Characterization Enables Seamless Handoff to Manufacturing

When it comes to other types of sensors, building in efficiencies becomes even more critical. We asked Travis about the value of testing early and in parallel with other product development stages:

“With capacitive touch sense, for example, we recommend getting a head start on the tuning process. We can work with the board and sensor flex before the complete product is built just to streamline the testing. This means we can start testing the bare sensor or design a fixture to mimic what the in situ performance would be.”

Oftentimes, the testing process reveals design issues or flaws with other components that may not otherwise surface until after manufacturing. He adds:

“With early testing, we’re able to identify if there are any challenges based on the geometry and address those early. We might also find assembly issues with the board, integration issues with the flex sensor, trace issues, and so forth.”

sensor characterization methods

By getting ahead of the curve, we help our clients quickly move their final prototype off the line and pass it over to UX fully tuned, so the actual performance testing is ready to go and has a much better chance of success.

When we asked how much lead time can be saved with this often overlooked early testing approach, Travis responded:

“It’s significant, saving potentially months of time. We’ve had projects with testing and fine tuning done five or six months ahead of when the build was supposed to be complete.”

How can AC support your testing requirements?

> Are you ready to get a jump start on testing ahead of your final design build?
> Do you have prototypes or components in need of iterative testing?
> Is there a roadblock in your design that you need explored through testing?

AC Test Engineering provides a full range of custom hardware testing services, at our test lab or yours, to ensure successful product design, verification and validation, and failure analysis.

How can we help?

Categories

Services

Client Industries

Tech Talk Series