Electrical Design Case Studies

[polytechnophile]

Can anyone recommend a place to find case studies in electronics design? I am looking for concrete examples of the engineering process from requirements definition through PCB layout.
As a physicist who designs instrumentation I would like to improve my improve my design process and employ more techniques that are taught in a formal engineering curriculum that I have not been exposed to from a physics background. While I have many texts on electronics fundamentals, specialties, and general lore, I have not seen enough cases that demonstrate how a professional EE approaches and solves a design problem.

 

[berkeman]

Welcome to the PF.

Unless things have changed a lot since I was in undergrad EE (which admitedly was decades ago), they don't teach much practical circuit design and PCB design in school. That comes at your first few jobs, and/or when you build your first few hobby circuits.

Do you have any particular circuits in mind? If you post an application and some example circuits you find through Google Images, we can certainly help you with tutorial information about how to implent them well in PCBA designs and finished product designs. Might be a fun thread...

 

[anorlunda]

It was a long time ago for me too. I learned a lot from Steve Ciarcia's Circuit Cellar articles. I see that you can still buy Circuit Cellar Magazine on Kindle.

 

[polytechnophile]

I have a fair amount of practical experience in both circuit and PCB design but I suppose I am looking to optimize the process as I get into more complex systems. I get by fairly well with my own systems but I feel like there must be some formal methods for things like requirements capture, error budgeting, etc. In particular I am starting to work on data acquisition systems with micro-volt signals for the first time and really thinking about noise budgets.

I can't really post specific applications from my work but I will try to think of some example circuits that I would struggle with. Thanks for the warm welcome.

 

[Tom.G]

Texas Instruments (http://www.ti.com/) has an excellent library of free on-line training videos, application notes, and design tips for both analog and digital.

Definitely worth exploring!

Cheers,
Tom

 

[berkeman]

I get by fairly well with my own systems but I feel like there must be some formal methods for things like requirements capture, error budgeting, etc.

For discrete analog systems:

• SPICE simulations including Monte Carlo analysis are mandatory
• Be sure to simulate using the parts from different vendors -- there will sometimes be small differences in datasheet max/min specs that make a big difference in whether you can approve the part for use in your circuit or not
• Explore the tradeoffs in partitioning your system between the discrete analog components and the analog ICs (off the shelf or your own ASICs) -- it can be complex and subtle sometimes where it is best to draw that line (especially for high-volume products)

For analog ASICs:

• Spend a lot of time evaluating the available geometry options and standard cell / IP options -- making the right choices can be very challenging, again, especially for high-volume applications, and if your design involves voltages above 3V, 5V, 12V, etc.
• Consider partitioning your circuits into FET and Bipolar ASICs, if applicable (one of our highest-volume products fell into this category)
• Be sure that the models you get from the vendors are accurate enough for your simulation and validation work (higher quality vendors like those qualified for TSMC usually have good models for you)

For digital systems/ASICs:

• There are very standard design methodologies for digital IC / ASIC design -- they involve source code control, simulation (especially using IP provided by your vendors), timing analysis, and circuit validation. I work more with the analog end of things, but if you want some links I can ask our digital ASIC design folks.

Paging @analogdesign for better advice...

In particular I am starting to work on data acquisition systems with micro-volt signals for the first time and really thinking about noise budgets.

Awesome. I'll look around to see if I can find some useful links for you. Also, do a search for picoammeter here on the PF. We've had a few threads discussing challenges in picoammeter design.

 

[berkeman]

In particular I am starting to work on data acquisition systems with micro-volt signals for the first time and really thinking about noise budgets.

Also, are you familiar with "bootstrap" techniques for such small signals to improve S/N? @jim hardy posted about them in a recent EE thread...

 

[jim hardy]

be careful with this link - it puts a pdf right on your C-drive without asking.

 

[Tom.G]

be careful with this link - it puts a pdf right on your C-drive without asking.

@jim hardy, that is a function of your browser settings, mine is set to "Always Ask". Other options are "Save File", "Use Downloader", "Always do this for files of this type".

Cheers,
Tom

 

[OCR]

that is a function of your browser settings

Yes... . ✔

Another option, if you use Firefox, is...

No downloading of anything... .

They are easy to read, also...

.

 

[jim hardy]

Thanks Guys ---

Firefox got cantankerous about loading Physics Forum pages
so i replaced it with Chrome

i just changed this to ON maybe that'll do it.

 

[polytechnophile]

Texas Instruments (http://www.ti.com/) has an excellent library of free on-line training videos, application notes, and design tips for both analog and digital.

Thanks, I will have a look at their training resources. I've already gone though some of their extensive op-amp references and have found the reference designs from TI and ADi very helpful.