Join Angela's webinar on automating CT workflows. In light of her webinar topic, I compiled resources and best practices related to repeatability this month.
"The definition of insanity is doing the same thing over and over again and expecting different results."
(Scroll to the bottom for the answer.)
A repeatability test can look like a manifestation of insanity. We might not "expect" different results, per se, but we are not surprised when seeing different results. That is kind of the point of doing a repeatability test.
New webinar series: Mastering CT: Advanced Techniques in Practice
If you are repeatedly scanning similar samples or need to consistently compare analysis results across batches, streamlining or automating your data collection and analysis process can be beneficial. In this episode, Angela will show you how to standardize and automate both CT data collection and analysis using practical tools and real-world examples.
You will learn:
How to set up automated CT scanning for multiple or similar samples
Tips for choosing mounts and sample handling tools that support repeatability
How to use software tools like VG STUDIO MAX and Dragonfly to automate data processing
Real examples showing how automation improves speed, consistency, and analysis quality
Join us to learn how to get repeatable results without the repetitive work.
Let's think about what these terms mean in CT analysis.
Repeatability: The consistency of results when the same operator uses the same CT system under the same conditions on the same part within a short timeframe. It is about how likely two or more analyses match when done under exactly the same conditions, except for the time.
Reproducibility: The consistency of results when there are changes in conditions. For example, a different operator, a different CT system, or repeating the measurement on the same part days later. It measures robustness across people, instruments, or environments.
Think of repeatability as "Can I get the same answer twice in a row?" and reproducibility as "Can my colleague (or another machine) get the same answer as I did?"
When we use CT for materials and life sciences research, we are just happy that we can see things and quantify them, and we tend to be casual about repeatability and reproducibility. However, these concepts become important when applying CT to analyze consistency or differences of many samples, process control, or quality assurance.
To be efficient, inspired, and informed.
Angela’s upcoming webinar will focus on automating CT data analysis workflows.
After repeating the same data collection and analysis, we tend to think we fully understand the process, and it should be easy to automate it. However, it is not always the case. Knowing common pitfalls in the automation process and how to avoid them can be useful.
Measurement system
Common mistake: Skipping measurement system analysis. Automation amplifies variability in results if the measurement and analysis system's capability is not established first.
Solution: Check the repeatability and reproducibility of the system, i.e., the data collection and analysis, by having multiple people run the analysis, recalibrating the instrument, and measuring the same sample multiple times, to understand your system's capability, repeatability, and reproducibility.
Sample treatment
Common mistake: Inconsistent sample preparation or sample mounting can lead to unreliable results.
Solution: Use the same sample mount and ensure the sample orientation is consistent. Making a sample holder tailored to the tested sample type and shape is recommended.
Missing information
Common mistake: Incomplete metadata and a lack of a file naming convention. Without standardized naming and file formatting, you might not be able to find critical parameters, and you cannot reproduce or audit the results.
Solution: Make sure to save all metadata (sample information, scanning conditions, voxel size, reconstruction parameters, etc.) and establish a rule about file names.
Overreliance on post-processing
Common mistake: For any type of measurement, improving the raw data quality beats post-processing. Overreliance on post-processing can cause an issue down the road if you find a problem with the analysis results that can be sorted out only by improving data quality.
Solution: Minimize and simplify post-processing to overcome noise, lack of resolution, etc., when designing the automation process.
Ignoring standards and regulations
Common mistake: If you need to comply with standards, such as ASTM, VDI/VDE, and ISO, these need to be considered when designing the automated process.
Solution: Check what standards you want to or need to follow before starting to design the automation process.
A collection of priceless and embarrassing moments curated by Sam Robles.
Unknown
This quote is often attributed to Albert Einstein, but there is no evidence he ever said or wrote it. The saying appears in a 1980s recovery and self-help literature, notably Al-Anon and Narcotics Anonymous meetings and literature, and it was later popularized by novelist Rita Mae Brown. (Source: Quote Investigator)
"The definition of insanity is doing the same thing over and over again and expecting different results."
That's a wrap. Please let us know how we can help you learn more about X-ray CT. We love to hear from you!
