8 reasons to choose the cloud for semiconductor yield management
If you work in the semiconductor industry, you'll find a lot of growing companies are moving to cloud solutions for their yield analysis. Here are eight reasons why you should consider making the switch:
1) ROI
Just concentrating on the man-hours saved alone, our customers tell us that our cloud solution adds 20%+ to an engineer’s productivity versus using on-premise tools. That’s an extra day a week per engineer. So you should be seeing a tangible return within months of purchasing a cloud-based solution.
2) Accessibility
Your engineers will be able to monitor production from any device from anywhere. They won’t be restricted to their work PC. They won’t have to download any data (e.g. STDF) to analyse. Instead, they interact with the database directly using an attractive GUI on their browser.
3) Scalability
As volume grows, you should be able to analyse and stack hundreds, even thousands of wafers at the same time. The image shown is a stacked yield map of 443 wafers which generated 4.5GB of original data, stacked in less than a minute using our cloud product yieldHUB on a Chrome browser. No data had to be downloaded first.
Additionally, if you make complex mixed-signal products, you should be free to analyse thousands of tests per datalog, across hundreds or thousands of wafers. Try that using a high-end stand-alone computer with only 8GB of RAM running a desktop data analysis solution. On the other hand, cloud servers will have 128 GB or more of RAM and dozens of CPUs working in parallel. So with this power at the server side, you’ll be able to see the true patterns of bin and test failures. This will enable your company to make better decisions based on richer data sets supporting long term yield improvement.
4) Data Management
If you want to, you can let the cloud YMS provider take on all your data management needs. They can do so at less cost than yourself, receiving the data on your behalf into their secure servers directly from the subcons. Data can remain on-line for as long as you need it to. For automative and high-reliability chips, this could be for many years.
Your engineers will be able to collaborate on-line with your subcons, saving travel costs and speeding up decisions. No need to send data back and forth. All the relevant data can be analyzed centrally over the web using login credentials you can provide to your subcons. Decisions can be made quickly and awaiting weekly meetings for decisions impacting yield and lots on hold will be a thing of the past. Engineers will have to do fewer site visits to solve yield issues.
6) Correlation
Provide the genealogy information to your cloud vendor and they will be able to link data from fab with wafer sort and packaged test, in software within the database. This means you will be able to correlate hundreds of wafers worth of data online across manufacturing. Think of the optimisation that you will be able to do in collaboration with your fab as you ramp.
7) Accuracy
The cloud approach is also ideal for automatically matching up the re-screen data with the first pass data in software and thus providing your engineers with accurate analysis in wafer sort and final test for all your production lots. Free from manual drudgery with data, your engineers will add a lot more value through direct root cause analysis and improvement.
8) Automation
Your engineers will be able to receive alerts and daily reports covering yield, bins and tests, even on their smartphone. They don’t have to look at any data unless there are exceptions. Up-to-date information comes to your engineers via email once the reports are set-up. The alerts will be based on rules decided by the engineers to spot yield, bin and parametric exceptions.
The above are some of the main reasons companies are moving to the cloud for yield management.
Thinking about switching to yieldHUB? No need to take risks, we offer a free demo if you'd like to learn more. All you need to do is fill in the form. Otherwise, if you'd like to talk to a sales agent now, you can find contact details here.