Advertisement

Blog

Analog Design Techniques: Past, Present, & Future

When I was a kid doing electronics as a hobby, I tinkered with circuitry based on vacuum tubes. I constructed circuits by doing point-to-point wiring between tube sockets and terminal strips. I used a soldering iron that looked more appropriate for soldering copper pipes. The components were usually mounted in a metal chassis, though sometimes I mounted them on a scrap of wood (the literal breadboarding of circuitry).

A typical breadboarded circuit. (Source: Wikipedia)

A typical breadboarded circuit.
(Source: Wikipedia)

There was no formal design or analysis process going on here. I was either copying someone else's design or tinkering just to see what would happen.

When I was a much younger engineer working with transistors, I typically fabricated circuitry on perf board using a point-to-point wiring technique. That's the phenolic board (like the inexpensive PC board material) with a grid pattern of holes across the entire surface, usually on centers of 0.156″ or 0.1″. Some amount of design was involved here — simple calculations to set up a bias points for transistors, voltage dividers as part of a power supply, and power dissipation calculations in a Zener diode-based voltage reference. There was no computer simulation of the circuits ahead of time. I built it, double checked the connections, powered it up, and poked around with a scope or my Simpson 260 Volt-Ohmmeter. If I didn't like the performance, I would heat up the soldering iron again and make the needed changes.

A typical perf board circuit. (Source: Wikipedia)

A typical perf board circuit.
(Source: Wikipedia)

Following this era of prototyping, I moved on to plugging ICs and passive components into a solderless breadboard or plug board.

A typical plug board circuit. (Source: Wikipedia)

A typical plug board circuit.
(Source: Wikipedia)

There was still no circuit simulation, but at least it was much faster to experiment with different op-amps or component values and see the effect. The soldering iron could remain cold. Of course, these boards are trouble at frequencies above audio. Parasitic inductance and capacitance become increasingly troublesome. Further, at any frequency, contact resistance would cause reliability and noise problems.

Now that my prototyping involves surface mount devices, I am far more inclined to conduct a circuit simulation first. This is actually pretty easy now, since many IC vendors have versions of a Spice simulation tool available for free on their sites.

In the future, prototyping may get more difficult if vendors continue to shrink their package sizes. My guess is that reasonable package sizes will remain available indefinitely. I also expect simulation tools to improve. Expect them to be easier to set up to get the sort of results you need to see.

As more analog circuit functionality moves into integrated analog devices, circuit simulation will become all the more important. You will still need to prototype a design, but it may mean you have to spin a PC board to build your prototype, because plug board techniques will no longer be practical.

8 comments on “Analog Design Techniques: Past, Present, & Future

  1. David Maciel Silva
    January 31, 2013

    @ Brad,

    Some companies provide services at low cost prototyping:

    http://www.seeedstudio.com/depot/fusion-pcb-service-p-835.html?cPath=185

    It might be worth checking out.

  2. Clyde
    February 7, 2013

    Brad,

     

    You left out wire wrap.  The first company I worked for had their own version of a perf board but actually accepted soldered components or sockets.  These boards had two internal planes and were orginally designed for 14 pin TTL logic.  The opposing corner pins were wired to the plands for +5 V. and return.  We also built many an analog circuit ground, cutting the power and ground traces as necessary.

    Wire wound sockets were used for the ICs, and headers were used for passives and transistors.  Power transistors were inserted directl into the board with their heatsinks attached.  The piecs were then interconnected with 2-level wire-wrap.

  3. Brad Albing
    February 12, 2013

    I had completely forgotten about wire-wrapping prototypes. We did that too – and we mixed analog and digital just like you described. It was a pain to build, but it generally worked pretty well and pretty reliably.

  4. Brad Albing
    March 24, 2013

    Altho' a bit messy at times if you miswired something and had to find the error (difficult to trace out the wiring) and then repair it (even more difficult to undo the wiring).

  5. patrick_m
    March 25, 2013

    Yeah, making the wire-wrap circuits was easy, the unravelling part became a pain in the neck: but that only made us double – nay, triple – check that we were making the right connections!

  6. Brad Albing
    March 25, 2013

    Quite so. And s/o else mentioned that you would do well to not reuse a section of wire previously wrapped, as it became fragile. So even moving a wire over one pin on an IC socket meant that you had to (or should) extract that wire completely and redo.

  7. Brad Albing
    March 25, 2013

    And even tho' we got away with prototyping our analog cktry 30 or 40 years ago (and it worked), I wouldn't try it now – with our very wide bandwidth parts, the ckts would be ringing like the pizza kid at my front door.

  8. Brad Albing
    March 25, 2013

    And just one more comment on wire-wrapping. We used IC sockets (these were the days when ICs were in DIPs) with wire-wrap tales of course. And we generally used the same sockets with a header plugged into it onto which we could solder resistors, capacitors, diodes, and transistors. One of the cool tricks we could do if we needed to add a resistor or capacitor somewhere in the ckt where no spare position on a socket/header was located was to just poke the component leads (all thru-hole parts in those days) thru the perf-board that everything was built on. Then use the electric wire-wrap gun and wrap a wire onto its leads. And usually just a tiny bit of solder to make sure, since the component leads were not square.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.