Janine Love, SIJ Editor and Technical Program Director for EDI CON
Janine Love, SIJ Editor and Technical Program Director for EDI CON RSS FeedRSS

Janine Sullivan Love

Janine Love is the editor of Signal Integrity Journal, working closely with the editorial staff and Advisory Board to bring technical, high-value content to readers in digital, video, and print formats. In addition, she serves as the technical program director for EDI CON and a contributing editor to Microwave Journal.

EMC/EMI

Ansys Interview: More on HFSS Mesh Fusion

February 19, 2021

Recently, Ansys released its HFSS Mesh Fusion product for simulation of complex EM systems. Signal Integrity Journal (SIJ) followed up with Dr. Matthew Commens, Principal Product Manager, HF at Ansys, Inc., to find out some more details. (You can see the original product news and background here.) The conversation has been edited for brevity below.

Dr. Matthew Commens SIJ: Tell us about the idea for HFSS Mesh Fusion, how did it come about? 

Matthew Commens (MC): Driven by emerging technologies like 5G and automated driving, these technologies’ higher   data rates and frequencies plus tighter form factors mean system design via discrete component modeling can miss   critical component-to-component electromagnetic EM coupling that could be ignored in previous technologies.

 A need for “system-EM” simulation arose where engineers needed to understand EM coupling between components   such  as an IC to package, or a PCB with connector and cable or, at a larger scale, a Ku band satellite antenna array on   an aircraft. Such simulation could be executed prior to HFSS Mesh Fusion, however due to the complexity of the system   CAD, as well as the potential order of magnitude range in geometry scale for the system, generating a finite element   mesh (FEM) could be very challenging.

 
SIJ: How long did it take to develop? 
MC: This was a multi-year effort that brought together patented CAD, meshing, and solver technologies in a tightly integrated FEM simulation flow.
 
SIJ: Your product information says, “HFSS Mesh Fusion bypasses previous barriers by applying optimal meshing technology at the component level, parallelized across cores, clusters or within Ansys® Cloud™” Can you provide some technical details of this meshing technology? How does it differ from other EM system simulation products?
MC: When you cannot simulate such “EM-systems,” you end up with design rules that might negatively impact the system performance in order to ensure the system will operate. With EM system simulation enabled with HFSS Mesh Fusion, these design rules can be broken and allow the engineer to truly explore the boundaries of the design to drive to an optimal solution.

As for optimal meshing technology, it turns out different approaches to meshing can work better in one CAD type vs. another. An example of “CAD specific meshing” is for laminate structures such as PCBs, ICs and IC packaging in which the upfront knowledge of how these components are designed as stacked 2D layers can speed up and enhance the meshing effort.

Prior to HFSS Mesh Fusion, it was a one-mesh-fits-all paradigm, and in an EM system with a mix of CAD types, generating the mesh could pose a challenge. But with HFSS Mesh Fusion, the right type of meshing technology can be applied locally to a component. So, for example, in the PCB with connector and cable example, a specific meshing algorithm can be applied to each to optimally address their respective and unique CAD types.

To understand the parallelization aspect, it helps to understand that with HFSS Mesh Fusion, the components of the system are initially meshed independently then merged together via the solver for an uncompromised, fully coupled EM solution. It is the independent, initial mesh generation that can leverage HPC and cloud resources. Downstream, the HFSS solver leverages multiple cores, processors, and nodes to solve a fully coupled EM system, including adaptive meshing for accuracy, with its standard HPC technology.

SIJ: What technical or market challenges had to be overcome to release HFSS Mesh Fusion? How did you ‘bring it all together’ from IC level to PCB, to connectors, to antenna?
MC: The big technical breakthroughs are in the solver algorithm which can both generate and solve a fully coupled EM-system matrix for these designs but also now solve much larger designs which HFSS Mesh Fusion makes more easily achievable. If there is a market challenge, it would be one of awareness. In many cases, engineers will not realize, or even not believe, what is now possible. 
 
SIJ: Can you talk us through a particular application example/customer case study for this product? How much time did it save designers as compared to their previous method?
MC: We had one particular example where a consumer electronics customer needed to perform an EMI/EMC simulation for one of their new products. Now it turns out EMI/EMC is one particular application that really benefits from HFSS Mesh Fusion because this application tends to apply simulation at the most complex stage of a design, near its end when the system is assembled and close to completion.
 

Fig 1

This customer example (pictured above) in particular was a touch panel display composed of many thin layers of dielectrics and conductors comprising the capacitive sensors array of the touch panel. The simulation was performed with the touch panel inside the TV housing plus the external EMI test antenna all inside a model of an anechoic EMI/EMC testing chamber. Without HFSS Mesh Fusion, solving this problem was essentially impossible as an initial mesh could not be generated. With HFSS Mesh Fusion, the first simulation ended in success. So, in effect, the time saving was infinite and this highlights how HFSS Mesh Fusion will make it possible to simulate what was thought impossible before.

SIJ: What software packages from Ansys does an engineer need to own in order for this to work?
MC: Just one: HFSS. However, since the difficult problems addressable with HFSS Mesh Fusion tend to be large and complex by nature, our HPC products can be leveraged to enable more cores and networked nodes to solve a problem more quickly. In addition, just under two years ago we implemented HFSS in the Ansys Cloud hosted by Microsoft Azure which can provide engineers ready access to additional hardware for their most difficult design challenges which HFSS Mesh Fusion is particularly designed to address.

 

 

You must login or register in order to post a comment.