- RFIC design involves a lot of analog design which is easy to miss when you are lost in RF/mm-Wave world. What we do involves 60% analog and 40% RF design.
- Skills you need in addition to RF/mm-Wave skillset to ace RFIC interview:
- Small Signal Analysis
- Current and Voltage References
- Single and Multistage Differential Amplifiers, OTAs
- Frequency Response of Amplifiers
- Noise analysis
- Feedback, Stability Analysis
- Inverters, Level Shifters, Muxes, Switches, Combinational Logic etc.
- Systematic Layout, Symmetry, Guard Rings, Shielding, Floor Planning
- We learn more Microsoft Excel than circuit design (maybe a sad truth) but data collection, analysis and automation is very important in time management and Excel would be a big help.
- In a big company, there are lots and lots of teams for every small task so that you focus on what you are good at:
- System team (budgets specs)
- Design team (that’s you most probably)
- Layout team (someone else does it, you guide them, you can also do if you have time…
- EM team (inductors, routings, baluns etc. you give them specs they design/optimize it
- Verification team (does RTL modelling, AMS simulations etc.)
- Reliability team (electromigration, aging, breakdown concerns)
- Process team (sets up PDKs, helps you with process)
- Bench team (sets up measurement bench and related automation)
- Test team (collects and analyze measurement data for you)
- Technicians (general purpose hardware support)
- And many more!
- Industry circuits are not descended from God:
- When we are student, we think what we do is plain basic, circuits in industry would be much more complicated. Not true.
- Everything starts simple and stays simple unless you really really have to improve upon it. Keep it stupid simple (KISS).
- Analog circuit design is not possible without insight:
- As a mm-Wave/RF IC designer, we rely lot on simulations. We lose insight and stop approaching things mathematically. DON’T DO THAT
- If you cannot meet specs, question the specs. System level solutions are more effective than heavily fine tuning a circuit.
- You work on your tiny block inside a module inside a chip. You are forever lost in optimization, debug and interface issues. There will be so many specs and corners to meet your performance that will keep your plate full. This is more true at senior to senior staff engineer level. You start taking more of a lead role after these levels.
- It’s important to learn how to use the simplest model for the job (instead of the most accurate one). Often times during design or debugging, you will be making models of things to understand. The writer itself used to make very accurate models and take pride in it, only to realize later that it had lost intuition, or it has become too intimidating that no one will ever make an effort to understand it. Keep your models simple and intuitive.
- Corporate life always come with some level of work politics. If you don’t like what you see, hate it but play the game, play with established rules within your moral confines, and change the rules when you have authority.
- You get paid for value you add, not the hours you put in. Don’t be a spice monkey. Don’t give all your time to employer. Invest sometime in yourself and focus on self-improvement.
- Quality over quantity, right? Yes, but quantity also matters unfortunately. When you leave your company, you only take a piece of paper (resume) with you. Patents are lame? Yes, but that is what will leave with you on the paper. If you have an opportunity to file a patent, and you are thinking there is no innovation in it, just file it. Show quality in your work but don’t ignore low-hanging fruits that may build your profile.
- Five years down the road, you will start to feel bored. Distinction between you (the smart guy) and peers (the avg folks) will start to dim. Big companies put us all in the same mold, gives us all the same work, and you start to lose your edge. Show with your work that you are different, and once you have established some repo, you can start asking for good/challenging/visible work. Some people like to “Rest and Vest” meaning they are doing almost no work and just resting while their stocks keep getting vested. You will reach a year in your life where you wanna be like them. Take it easy. It’s likely that you are de-motivated. Rest. Pause. Don’t completely stop.