Electronics > Projects, Designs, and Technical Stuff

RF Experience and Analog Design in General

(1/3) > >>

Cigarsnob:
Howdy,

I am wondering if anyone can recommend to me some articles, app notes, sites, etc etc pertaining to wifi front end design? I am a graduate student working towards my masters degree in analog design and I am taking my first RFIC design course. I asked my professor what I can do to gain more experience (project wise, reading, etc etc) to be more attractive to employers once I finish up with school and he recommended that I become very familiar with Cadence and to design a wifi front end.

As I am new to designing at the transistor level and new to RF,  can any of you recommendation projects, tips, and/or anything else. Anything and everything would be much appreciated. Maybe simple to advance RF discrete projects I can do at home and maybe transistor level projects I can do on campus using Cadence? Both RF/RFIC and analog/analogIC information would be great.

Thank you
Phil

amspire:
Phil,

The great thing about starting in this work in 2011 is the amazing amount of refined technology you have to work with.

The difficult thing is it is really hard to understand anything. Your design work is often sticking a few ICs to a small board - each IC with documentation like a telephone book. Where is the understanding in that? There is lots of "what" things do in mind numbing detail, but very little on "why" it is done that way.

I do not know what your ambitions are, but if you really want to become an expert, the best way is to track back to the start.  You know those boring bibliographies at the bottom of papers and in the back of text books? Just start tracking those down, and you will find two things - new references going further back, and you will also find key documents that everyone continually refers to.

If you do this, you will find something unexpected. In many cases the very early papers and the key documents are extremely readable, and are often extremely far reaching. Also when you start to look at papers in historical sequence, it all starts to make sense, and that 5 page block diagram in a data sheet for a modern device all starts to look obvious. It is often that case that the real discoverers had extraordinary clarity into the possibilities of their new technology, even if the hardware at the time was very limiting and crude. All this sounds like a lot of hard work, but if your mind starts getting interested, it stops being hard work.

I think you will find this true in learning about WiFi and RF transistor design as well.  I think you will find some books and papers from the 1960's into RF transistor design quite easy to follow, even if they are only talking MHz rather then GHz. Remember, you are accessing the brilliant minds of the people who invented how to design this stuff, not the person who a few years ago regurgitated for the 20th time his take on how to design RF. It is likely that you find that many of the brilliant people in the field would be happy to talk to you directly as they will be surprised to find a modern student who understands and respects what they achieved decades ago.

Most people don't go back to the past like this, and most people become very average designers.

Now in terms of projects, they can be good, but with RF you just have to develop a feel for it and I am not sure how much you will get from building a working design. I think you to need to start breadboarding simple RF circuits and test them. A single transistor RF amplifier - try and maximize the gain. A tuned circuit. A filter. A mixer.  Impedance matching. An antenna. It will be really hard to work in RF if you do not develop a natural feel for it.

In terms of a project that will teach you about RF, look at some of the DIY spectrum analyzer projects around, often built from things like old TV tuners. The spectrum analyzer is a very modular device so it is easy to get something that works -  probably really badly, and then improve modules one by one.

Cadence tools are very powerful, but they are useless if you don't know how to design. You can probably learn more playing with your own ideas in one of the free Spice versions in your own time. When you know exactly what you need to do, then Cadence is great.

Richard.

Conrad Hoffman:
No reason to go part way! I'd suggest getting a copy of the old 1940s Radio Engineering book(s) by Fredrick Terman. I think everybody who came after just copied his stuff and rephrased it. They shouldn't have; clearer writing can't be had. The books were reprinted and revised many times. Some of the best fundamentals in the early ones were removed to make room for modern developments in the later ones. A search of used book stores should certainly turn up something for a reasonable price.

Cigarsnob:
Richard,

Thank you for such a great response.

As for as my ambitions, I'm still figuring out where I want to go in terms of analog design. Be it at the transistor level IC design or discrete design or RF. I'm assuming they all share similarities so maybe picking one aspect of analog design may not be that important. Agree? Disagree? I just know that I don't want to be an average designer. Maybe it's the fact that I'm studying in the Silicon Valley, or the fact that I'm living (academically) in the shadow of Standford that makes me have a chip on my shoulder and the desire to become as successful designer. That and I do enjoy the material. Albeit it does kick my ass at times.

Since I have access to IEEE Xplore through my school, I will take advantage of that and follow what you recommended by reading papers further back and work my way forward. And that DIY spectrum analyzer project does sound fun. I'll have to see if the project is within my budget. Either way, you're right about starting with simple circuits to get the feel of things.



Conrad Hoffman,

I did find some of Fredrick Terman's books online for just a handful of dollars, so I might as well check it out. Thank you.



-Phil

amspire:

--- Quote from: Cigarsnob on November 20, 2011, 01:02:41 am --- And that DIY spectrum analyzer project does sound fun. I'll have to see if the project is within my budget. Either way, you're right about starting with simple circuits to get the feel of things.


--- End quote ---

As I said a spectrum analyser is a modular design, and you are wanting to learn, so don't get sidetracked into thinking you need to spend $$$$$ on that great IF filter, or mixer, or spectrally pure oscillator.

If you make a lousy mixer from four 1N4148/1N914 diodes worth about a cent each and some badly wound transformers you made using junk part torriods, then you might only get 20 to 30dB dynamic range and a disappointing upper frequency range.  But this is exactly how you learn. If you instead buy a $40 MiniCircuits mixer module, it will work really well, but you do not learn much.

What happens if your oscillator that goes into the mixer outputs a crude squarewave instead of buying an ultra stable YIG sinewave oscillator assembly with harmonics down around -60dB or lower? It is really worth trying it. So you can start off with a very simple and very cheap voltage controlled squarewave oscillator.

To learn, you actually want to see why things are not working.

Richard

Navigation

[0] Message Index

[#] Next page