Webinar Recap: Fabscape Development and Data Analysis in Action 

In a world where semiconductor chip production is increasingly complex, data becomes much more useful when it can be collected and analyzed on a single platform.  

In our March 2023 live demo with Semiconductor Digest, Yuji Minegishi and Jacob Strock of Gigaphoton explained the benefit of the Fabscape open platform and demonstrated how simple it is for semiconductor developers to go from a blank slate to robust data visualization.  

Here’s a quick breakdown of what we covered, and you can also watch the full recording on demand. 

  1. The Fabscape Customization Structure
  2. Creating a Driver 
  3. Adding Data Visualization
  4. Integrating Data Analysis 

The Fabscape Customization Structure 

In a previous webinar we broke down exactly what Fabscape is and how it works: a modular open platform designed to be customizable, scalable, and collaborative to support the needs of both device manufacturers and equipment vendors. 

In contrast, our recent live demo was focused on showcasing the technical side and demonstrating how developers can create extensions in the form of plugins and drivers to achieve any fab’s data collection, analysis, and visualization goals.  

The two main building blocks to leverage the power of Fabscape are plugins and drivers.  

Plugins: A plugin is a piece of software that adds custom features and makes it possible to visualize data in a certain way. Without plugins, Fabscape is essentially an empty shell.  

Drivers: A driver is closely related to a plugin, with the difference that they are specifically tasked with collecting equipment data and making it available to the Fabscape backend. 

Both plugins and drivers are Docker containers and can easily be created by creating an account and downloading the Fabscape Plugin Generator and using an integrated development environment (we’re fans of Visual Code Studio).  

Creating a Driver and Visualizing Data  

In a fab environment, a driver would pull equipment data from the fab floor (agnostic of the type of equipment or vendor). 

For the purposes of our live demo, we created a driver that pulled from a large public API site containing COVID statistics. We then built the proto files using a protocol called gRPC and spun up a user interface in the form of a modal screen by leveraging reusable code components.  

Plugins can be generated by following the same process as drivers, then be used to manipulate data and visualize it however you need to. 

If part of a dataset needs to be secured, you can attach a private database to a plugin or driver to siphon off proprietary data while allowing the rest to flow into the general population.  

Integrating Data Analysis 

Once you’ve connected fab equipment by creating drivers, you can add predictive services that leverage the power of AI/ML. 

This is accomplished by pulling parameter data from the Fabscape backend and returning forecasting results. A predictive model can be built to analyze data and deployed as a plugin service and hooked up to the Fabscape frontend. For our own needs at Gigaphoton, we utilize a Jupyter Notebook plugin that we developed for Fabscape that allows us to model data and use it to conduct predictive analytics for our lasers.  

Explore the Power of the Fabscape Today 

You can easily explore Fabscape yourself by downloading our free Toolkit for developers. Want to take it a step further and work with us 1:1 to solve your fab analytics challenges? Apply for our Advisory Program

The above content represents just a small portion of the original webinar. Watch the full Semiconductor Digest webinar recording now!