0
\$\begingroup\$

Taking a circuit using transistors, CMOS dividers and monostable multivibrators, can it be considered as analog one or it becomes digital?

\$\endgroup\$
6
  • 7
    \$\begingroup\$ Some fanatics say all circuits are analog... \$\endgroup\$
    – TQQQ
    Commented May 22, 2023 at 2:33
  • 5
    \$\begingroup\$ @TQQQ:They're not wrong. Even when up to my eyeballs in a complex digital design on an FPGA, all of the thorny issues are analog -- everything from allowed PLL frequency ranges to signal integrity (both on-chip and off-chip), metastability, fanout and routing delays, voltage tolerances, noise margins, radiation effects, etc., etc. The actual digital design is quite straightforward in comparison. \$\endgroup\$
    – Dave Tweed
    Commented May 22, 2023 at 3:29
  • 1
    \$\begingroup\$ A circuit's intended usage my be digital but all circuits are ultimately analog. \$\endgroup\$
    – DKNguyen
    Commented May 22, 2023 at 3:57
  • 1
    \$\begingroup\$ Digital (my opinion) means numerical. This is a layer above electronics. Analog (my opinion) means simulation. This is also a layer above electronics. The rest of electronics is just electronics. Why divide it into two camps? \$\endgroup\$
    – RussellH
    Commented May 22, 2023 at 4:00
  • 1
    \$\begingroup\$ @PStechPaul:That would be wrong, too. While we normally talk about particle interactions as discrete events, there's certainly nothing "digital" about the particles themselves or their wave functions (probability distributions). \$\endgroup\$
    – Dave Tweed
    Commented May 22, 2023 at 11:41

5 Answers 5

3
\$\begingroup\$

When you respect all the assumptions that 'digital' chips require - 'fast' edges that don't linger in the forbidden region in the middle, and cross it once when intended, clean highs and lows, propagation delays allowed for, setup and hold times respected, fanout/in specifications met, clean voltage rails, and a few more I can't bring to mind right now, then designing with digital components can be as easy as manipulating a few Boolean equations on paper, and you're done.

When things go wrong, the logic doesn't behave as you expected, you get thrown back into the real world where everything is analog to debug that circuit, and see what assumptions you have not met.

\$\endgroup\$
1
  • \$\begingroup\$ Thank You for Your reply \$\endgroup\$
    – Ken90
    Commented May 22, 2023 at 11:59
2
\$\begingroup\$

Digital is a strict subset of analog.

All digital circuits are implemented with analog components, whether that be transistors, or levers and cogs.

It's not about what it's made of, it's how you use it!

A digital circuit is one in which the rules of digital logic apply:

  1. There is a definite input parameter range (e.g. voltage) for logic '0' and '1' (or other levels if multi-nary)
  2. There is a definite output parameter range which meets or exceeds the input range
  3. The input and output parameter units and ranges match, or can be conveniently interfaced
  4. There is a fan-out or fan-in rule (one output can connect to one or more inputs, or one input can connect to one or more outputs); and
  5. There is a propagation delay between any given input changing, and the output(s). (The exact rules for propagation delay may be quite involved, depending on which inputs and combinations are changing into a given component/function block, but the important thing is the logic operation is known to be completed within some period of time from an input change -- or if the time is well enough defined, it can be used for delay-based logic, but this is generally discouraged for more than simple circuits.)

A whole circuit, combining analog and digital features, might be termed "mixed signal".

There are certainly circuits that are useful to attach to a system, and which are wholly digital; a serial interface peripheral for example. (One might rightfully argue, if that interface contains a PLL or voltage regulator, it isn't entirely digital. We might constrain this example to older peripherals like the 16550 chip, in contrast to, say, USB devices which contain a 48MHz oscillator and/or PLL, and likely a 3.3V regulator as well.)

Conversely, we can trivially nullify the digital-ness of any given device, by observing that its inputs and outputs are continuous voltages -- analog by nature. This isn't as useless an observation as it sounds, as one can indeed define the input and output pins of a logic circuit as 1-bit analog-to-digital (and vice versa) converters, and thus we might ignore the analog levels within that circuit (assuming it is fully described by the above digital rules, and its logic function), but still acknowledge its interface manifests as analog levels.

An application of this observation might be in high-speed design, where the impedance matching and ringing and etc. of a transmission line is nearly meaningless to a digital model (the only directly applicable parameter is propagation delay), but if we consider the interfaces as analog, then we are immediately reminded of their interface to the surrounding world, and the need for impedance matching, reduction of noise and crosstalk, etc.

Conversely, we can "abuse" digital inputs for analog purposes, given some restrictions. Networking protocols can sometimes be made directly with logic pins (e.g. microcontroller GPIOs); while we might still have to provide external filtering and protection (for nominal performance in a noisy and often hostile electrical environment), the analog-digital interface can sometimes be as trivial as the pins themselves. 10BASE-T Ethernet has been demonstrated on MCUs as small as ATTiny85 for example, with only passive interface components.

Another application that's been demonstrated: an ambient light sensor, using an LED, and its (and the pin's) capacitance, and the pin's weak-pullup function. This works by reverse-biasing the LED, charging it either to the supply voltage, or fractionally by way of the weak pullup. In any case, pin voltage starts above threshold, and say a logic '1' will be read. Over time, the LED's leakage or photocurrent (proportional to ambient light) discharges the capacitance, and some time later, a logic '0' is read. The time is inversely proportional to illumination.

Indeed, since the LED emits light when forward-biased, this can be used as the basis for a bidirectional "no-interface" link! Here, two MCUs are equipped with LEDs, one blinking out a signal and the other sensing the minuscule shift in ambient light. The signal might be deep in the noise floor (due to the weak reception between LEDs, masked by external causes of light variation), but given a low enough signal bandwidth and much averaging, communication can indeed be established.

\$\endgroup\$
1
  • \$\begingroup\$ Thanks for the extensive explaination \$\endgroup\$
    – Ken90
    Commented May 22, 2023 at 12:00
0
\$\begingroup\$

Analog originates from electronic simulation of mechanical systems, Digital originates from number systems.\$256\$ has 3 digits, the ones digit, the tens digit and the one hundreds digit.

"Digital" computers operate using "binary digits".

Any two level, position device is called a "binary device". Not necessarily digital. Example, two light switches generally turn two lamps on and off. Switch A for lamp A and switch B for lamp B. Are these binary digits or do the just turn the lamps on and off? Certainly they are binary devices.

Over time, analog came to mean "not digital", thus implying that the light switch might be referred to as a binary analog device.

Some may argue that analog should be used only when referring to electronic simulators. The rest of electronics would then be referred to as continuous, discrete level, binary or some other term that accurately represents the intended usage.

There is this tendancy to divide the electronics world into two: Analog and Digital. It is just not that simple.

So the answer to your question is it depends on whether the device(s) are to represent numbers or not.

\$\endgroup\$
2
  • \$\begingroup\$ I see. Thanks for the answer. Well if a logical gate needs for example to DIVIDE a frequency of course will need to count it. But I can say that a frequency is coming from an analog signal even using a Flip-Flop to divide it would become digital? Then a BINARY counter will of course need to do some operations with numbers? hence digital? For my POW digital is representing something that manipulates the operation itself like the code inside an MCU. But still can a simple CMOS Counter IC be considered as digital? \$\endgroup\$
    – Ken90
    Commented May 22, 2023 at 4:00
  • \$\begingroup\$ Counting implies numbers, therefore there is a digital application. Don't over think the issue. The physcal device is still electronic so its usage must involve electronics even though the application is digital. \$\endgroup\$
    – RussellH
    Commented May 22, 2023 at 4:06
0
\$\begingroup\$

As already pointed out in another answer, is the signal(s) being processed that defines the nature of the circuit. Generally, analog signals carry information by varying voltage in continuous mode, where all the levels are in some way "valuable" and different. Digital signals have only two states - on and off. Of course no digital signal has only two voltages, because during transitions all the different voltages are produced, but the main purpose is to differentiate 0 and 1 (0 volt - 3.3 volts for example).

By this definition, a relay would be "digital", because it has only two states, on or off, both in its input (coil) and its output (contacts), even if the signal through its contacts is analog; but in that case (switching an analog signal), the relay would also be "analog" because the signal it operates on is analog.

A pure analog circuit could be one that has no signals which (by design) take only two values. A comparator can have analog signals as input, but its state (by design) has only two values, so it is not pure analog.

\$\endgroup\$
1
  • \$\begingroup\$ Thank You for the reply! \$\endgroup\$
    – Ken90
    Commented May 22, 2023 at 12:00
0
\$\begingroup\$

It depends on what those transistors, CMOS dividers and monostable multivibrators are used for.

In general, analog circuits deal with continuously variable signals, while digital signals have discrete levels and time steps. A digital circuit may use transistors as on/off switches, and monostable multivibrators to generate fixed time delays. CMOS dividers could be used to count pulses or produce lower clock frequencies. The signals are digital because they only have two valid levels, high and low, that are evaluated at fixed time intervals relative to the clock.

But of course transistors can also be used in analog amplifiers, and multivibrators can generate continuously variable time delays. A CMOS divider is always digital, but its outputs could be used for analog purposes.

As an example of how definitions can be vague, here's the schematic for an 'analog' pong game:-

enter image description here

Here we see a CD4013 CMOS digital logic IC being used to generate variable time delays with SR flip flops, a CD4060 divider producing a 3 bit digital output that counts from 0 to 7, transistors producing digital outputs from analog inputs, and another transistor taking a digital input to reset a capacitor charging circuit.

The complete circuit could rightly be called 'analog' because it works largely with analog signals. But some parts of it are digital.

\$\endgroup\$
1
  • \$\begingroup\$ Thanks for the extensive and example related reply \$\endgroup\$
    – Ken90
    Commented May 22, 2023 at 12:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.