editor's blog
Subscribe Now

A New 3D IC Approach

While 3D ICs have been sexy for a while, actually building them has been a bit more of a stately process than simply talking about them. The details house numerous devils that must be worked out in order for the process to be manufacturable.

One way that TSMC is approaching it is through what they call CoWoS: Chip on Wafer on Substrate. This is a 2.5-D process, really: the multiple dice are mounted on a silicon interposer. The catch is that this happens while the interposer wafer is still whole. This allows known-good dice to be mounted and tested together using a wafer methodology instead of assembling onto interposers that have already been diced up.

Once the completed stack is diced up, they are mounted on the package substrate, and packaging proceeds as normal.

The issues driving this approach are concerns that potential customers of 3D IC processing have ostensibly had about having multiple players in the process, with wafers from different sources being shuttled to packaging houses for further assembly. Some companies apparently view such a process as risky. Using this process, TSMC controls and does everything up to and including getting the structures onto the package substrate (making me wonder whether it was really customers or simply TSMC who was concerned).

Cadence’s role was in the front-end design portion of the project. You have multiple dice that have to know something about each other in order for the required signals to meet properly, and in order for traces that start on one die and end on another, passing through the interposer en route, to be modeled properly. This can have a non-trivial effect on the design workflow, since changes on one die may have implications for another die, and such events must be communicated. It’s still a partially-manual process, but they’re laying down the infrastructure for this paradigm change.

Modeling of the TSVs themselves (which communicate from the interposer redistribution layer down to the package substrate) is also new.

Cadence expects TSMC to go into full production on this in 2013; right now, come customers are working with it in a pre-production state.

You can find more info in Cadence’s release

Leave a Reply

featured blogs
Nov 5, 2024
Learn about Works With Virtual, the premiere IoT developer event. Specifically for IoT developers, Silicon Labs will help you to accelerate IoT development. ...
Nov 13, 2024
Implementing the classic 'hand coming out of bowl' when you can see there's no one under the table is very tempting'¦...

featured video

Introducing FPGAi – Innovations Unlocked by AI-enabled FPGAs

Sponsored by Intel

Altera Innovators Day presentation by Ilya Ganusov showing the advantages of FPGAs for implementing AI-based Systems. See additional videos on AI and other Altera Innovators Day in Altera’s YouTube channel playlists.

Learn more about FPGAs for Artificial Intelligence here

featured paper

Quantized Neural Networks for FPGA Inference

Sponsored by Intel

Implementing a low precision network in FPGA hardware for efficient inferencing provides numerous advantages when it comes to meeting demanding specifications. The increased flexibility allows optimization of throughput, overall power consumption, resource usage, device size, TOPs/watt, and deterministic latency. These are important benefits where scaling and efficiency are inherent requirements of the application.

Click to read more

featured chalk talk

RF Applications in Satellites
Sponsored by Mouser Electronics and Amphenol
In this episode of Chalk Talk, Maria Calia from Amphenol Times Microwave Systems, Daniel Hallstrom from Amphenol SV Microwave and Amelia explore the use of RF technologies in satellites, the biggest design concerns with connectors and cabling in satellite applications and why RF cable failure mechanisms can be managed by a careful selection of materials and manufacturing techniques.
Nov 12, 2024
1,880 views