Reader Question: Is Analog Engineering a Dying Art?

With recent advancements around us to make system more robust and automated electronics comes into picture. These IoT, Industrial automation and rest of many applications won’t be possible without the medium which connects physical world to the electronics that’s where analog design comes into picture. Also software applications has not meaning if there is no hardware on top of which they can run which itself creates the lot of opportunities in different domains. There are other perspectives as well which keeps us motivated to keep working to achieve the target of creation of devices with as small size as possible with as many functionalities. With systems becoming more complicated it is also becoming important to keep reading the latest advancements happening this world and being up to date whether it’s related with the tool or it is with understanding of individuals. Remembering the fundamental of analog design also plays the important role here as well which setup the bed to think in terms of complex designs.

Strange question. I’ve always thought of digital electronics as a special case of analog electronics. They are both subject to the same laws of physics, and everything in the “analog design octagon” could be considered as trade offs for the design of digital circuits operating in extreme conditions. It’s just that under “normal” conditions (low speed and short distances), digital electronics design is a slightly easier task given that most components are functionally discrete at the application level.

So as someone who is not an analog designer, it seems that as voltages continue to decrease and clock frequencies increase as well as space between traces on PCBs, the analog elements become increasingly important. Not to mention high frequency (copper) network PHY particularly as we move beyond 10GHz… Yes, it’s nice to live in a “digital world” where physical constraints never come into play, but as we all presumably learned in Electronics 101, the real world does not look like those perfect circuits in textbooks. And whose to say that trinary or higher encodings per gate won’t be needed by digital systems in the future - again getting us into interesting analogue issues.

With the exception of quantum mechanical effects our world is defined to be inherently of continuous nature. Modelling analogue functionality in discrete space often introduces artefacts and error not seen otherwise. Although we have learned to quantify and deal with them, as frequencies go higher (or conversely precisions go finer) we continue to need to have sufficient understanding in continuously modelling our developments. Therefore, analogue engineering will have to stay with us! After all, we are simulating processes that in their nature are analogue.

What seems to be more pertinent is to whether “journalists” can continue to recycle stale stories, and interest people in wasting their time to read them… or as to whether they can be replaced by “digital” journalists programmed to do the same thing… perhaps much better.

Analog design from my experiences is becoming a fundamental knowledge as the technology process advances. In my group named product system architecture, most engineers are skilled with analog electronics meaning they are solid on OPAMP design, Power management buck/boost design, passive/active filter design, coil designs. The thought process is always around ohms’ law when dealing with complex digital/high-level block designs. That said, the analog design is must-have knowledge for every electrical engineer working in the digital, mixed-signal world. I feel the difference is just the title and project preferences from mixed-signal/analog counterpart position.

Without solid analog basic knowledge you’re prone to fail even when designing digital circuits of “medium complexity” (say an ARM microcontroller and some periperals). “Digital” is really a somewhat simplified special case of “Analog” 😊
If - as a hardware designer - you have a solid knowledge about software as well, you’re more or less “the king”.
My younger colleagues are quite eager to learn the tricks I’m constantlydemonstrating - and the design rules that are (unfortunately) not taught at the universities these days.

I see lots of totally uneducated youngsters around me here, barely familiar with Ohm’s law, but successfully creating digital (And even - mixed!) circuits and starting to suppose themselves even better engineers than old-fashioned (capable in most cases to replace 3000000+ digital transistors with just 3 discrete BJTs ) ones.

I guess, here is a difference - as between painter artists and mass owners of digital cameras, but who cares?

Camera works incomparable cheaper and faster, than those oldies with their outdated paints and easels.

The world is analog. The digital aids just to compress the interface between Input and Output.

Analog electronics will never be less important. Arduino isn’t needed for everything. ( no disrespect to Arduino ) In so many cases, a simple analog circuit approach will do what is needed. Take audio for example…the human ear does not respond to digital data. The human eye does not do much with digital video data streams.

This conversation has been going on within engineering circles for at least 30 years. Each year digital gets a little faster and expands a little more into previously analog applications. What has become evident during this evolution is the “digital” signal is not a simple on/off signal, but an on/off signal with very ANALOG rise and fall times. The digital engineer must deal with this very analog characteristic in any high-performance design - which sort of makes him an analog engineer.

Indeed the whole real world is analog, or at least, fairly linear.  those who doubt that should attempt to design a 3.3 volt power supply for their CPU board.  Of course analog designing of linear circuits takes more thought and probably even some calculations, but generally the circuits function as intended. The challenge gets bigger as data rates rise and every mm of board trace is a transmission line not isolated from other transmission lines. and in the gigahertz realm the lines are much thinner and much closer so the transmission line theory is far more applicable.
So the digital segment is a small and tedious portion of the analog world.

I wonder where the question comes from in the first place. I suspect that it’s from the software world. If you’ve worked in a digital ASIC environment, for example, you’ll be the last person to discount the analog engineering aspect, or your chip won’t work. I’ve always thought that digital is analog until it works, and then everyone forgets how it got to be digital. I am personally annoyed that the analog vs digital question is still being seriously asked, mainly for all the reasons everyone has brought up here in these comments. Semiconductor physics is a priori analog and even the quantum aspect of it is hardly an “on-off” reality but a mess of levels and energies that need to be carefully tuned in or out in silicon. Can we just agree that the world is round, lol?

Right, well if you think “analog” isn’t all that important, try designing the RF front end for a transceiver like a cell phone or cell tower without any specific knowledge, training, on-the-job experience or tools to get the job done! Bear in mind you’ll need to be able to design it so it not only can make and receive calls, but also so it can meet all the tests imposed by the FCC that are designed to find out that all radio frequency radiation can occur only on the frequency bands for which that carrier (for whom that device is being manufactured) holds the appropriate radio frequency licenses. Or how about the “analog design” of a power inverter for an electric vehicle which can provide reliable power at a level of hundreds of vols AND amperes reliably for the next couple of decades? Or for handling the control and power conversion issues for the next generation of solar or wind farms? Does anyone really think these projects don’t provide design issues sufficient to provide challenges for a whole generation of engineers? Of course they do, it’s just that for some people these don’t readily come to mind as “analog design” issues!

Claude Shannon laid the mathematical groundwork for digital systems to prosper, but as everyone knows digital systems depend fundamentally on their underlying analog foundations.  In my specialty area of expertise, geophysical instrumentation, the reality is that the world is ANALOG, and the top priority is converting analog signals from analog sensors into robust digital representations.  Doing this while minimizing cost, power consumption & complexity in field operations is at the heart of my analog designs.  Including a few A/D converters & an appropriately-sized microprocessor is the easy part…