If you've never designed an RFIC at a company, this article provides an overview of what the whole process looks like and the requirements you will have to meet.
In the circuit or front-end design stage, who decides the architecture for a specific circuit like an LNA or Mixer? Is it the VLSI circuit designer's responsibility, or do experience circuit architects determine the specs and architecture? Is the IC designer only responsible for optimizing circuit parameters to meet the provided specifications using EDA tools?
The answer to this really depends on the product and the company. Smaller companies are more prone to experiment because circuit performance is what might be a differentiating feature for their product. Larger companies are less prone to trying out new circuits because they need to know it will work across billions of shipped products.
Many times, designs are derivatives of a previous family of products. And knowing that it worked well earlier, there is no reason to change things. But every once in a while, a tough spec comes along where you will have to innovate as a circuit designer to overcome the problem. As long as there is a good justification for moving to new circuit architectures, companies will be open to it.
Otherwise most circuit design comes down to optimizing parameters to meet spec, and running an extensive battery of simulations to completely verify the functioning of the chip. This is what is ultimately reviewed in a design review.
Let us take example of LNA to be used in particular receiver chain. Who decides whether it is to be single ended or differential, common source or common gate, single stage or multiple stages. When a design at one node needs to be transitioned to lower node, i believe someone needs to make these decisions whether block or sub-block architecture needs to be updated or not.
Are these decisions made at end of circuit design engineer optimising and running extensive circuit simulations. Or is there someone higher up in hierarchy who takes these high-level decisions ?
There is usually a design manager or lead who is responsible to driving engineers in the team to work on products and deliver them on time. If the design choices are not clear, it is usually a team decision on how to proceed.
For example, the team will discuss the pros and cons of various topologies and decide to run a feasibility study on whether a particular approach will meet the specification. This may be driven by the request of someone higher up the hierarchy. The engineers on the team will have to do the work of simulations.
After some preliminary simulations, discussions in group meetings, back and forth, then they will decide as a group that this is how they should proceed. At least this is how it should work in good design engineering groups in my opinion.
It is not only an engineer or manager decision. A lot of minds are required before committing significant resources.
In the circuit or front-end design stage, who decides the architecture for a specific circuit like an LNA or Mixer? Is it the VLSI circuit designer's responsibility, or do experience circuit architects determine the specs and architecture? Is the IC designer only responsible for optimizing circuit parameters to meet the provided specifications using EDA tools?
The answer to this really depends on the product and the company. Smaller companies are more prone to experiment because circuit performance is what might be a differentiating feature for their product. Larger companies are less prone to trying out new circuits because they need to know it will work across billions of shipped products.
Many times, designs are derivatives of a previous family of products. And knowing that it worked well earlier, there is no reason to change things. But every once in a while, a tough spec comes along where you will have to innovate as a circuit designer to overcome the problem. As long as there is a good justification for moving to new circuit architectures, companies will be open to it.
Otherwise most circuit design comes down to optimizing parameters to meet spec, and running an extensive battery of simulations to completely verify the functioning of the chip. This is what is ultimately reviewed in a design review.
Thanks Vikram for clarification.
Let us take example of LNA to be used in particular receiver chain. Who decides whether it is to be single ended or differential, common source or common gate, single stage or multiple stages. When a design at one node needs to be transitioned to lower node, i believe someone needs to make these decisions whether block or sub-block architecture needs to be updated or not.
Are these decisions made at end of circuit design engineer optimising and running extensive circuit simulations. Or is there someone higher up in hierarchy who takes these high-level decisions ?
There is usually a design manager or lead who is responsible to driving engineers in the team to work on products and deliver them on time. If the design choices are not clear, it is usually a team decision on how to proceed.
For example, the team will discuss the pros and cons of various topologies and decide to run a feasibility study on whether a particular approach will meet the specification. This may be driven by the request of someone higher up the hierarchy. The engineers on the team will have to do the work of simulations.
After some preliminary simulations, discussions in group meetings, back and forth, then they will decide as a group that this is how they should proceed. At least this is how it should work in good design engineering groups in my opinion.
It is not only an engineer or manager decision. A lot of minds are required before committing significant resources.
The link which you have mentioned for 15 min call leads to a page which says Calendly URL is invalid
I’ve temporarily discontinued the free calls
Very helpful article as an introduction to RF IC design process! Thanks Vikram for sharing.