Signals- Analog/Digital/Continuous/Discrete

[thomasmwilson89]

I decided to Google "Discrete Vs. Digital Signal" today figuring I'd get a fairly simple answer, but I was quite shocked to find so many differing answers. It was my basic understanding that digital and discrete are basically the same thing, just different terminology. What I discovered what quite different, and here are my findings:

Signals need to be thought of as either continuous or discrete, meaning...

Continuous- at any given time (x axis) you can measure your value (y axis).
Discrete- values (y axis) will only be available at pre determined times (x axis). i.e. every .5 seconds you take sample data, but don't know the value in between these snapshots.

From there you can then look at analog vs digital...

Analog: Values range is not defined, the value can be any number
Digital: Values are set at defined numbers. i.e. when a switch is activated, its 1. When a switch is not activated, its 0.

So assuming my understanding is correct (which hopefully it is), it's not fair to just say "an analog input" or "a digital input". Because it seems as though you have 4 options. Continuous Analog, Discrete Analog, Continuous Digital, and Discrete Digital.

So my questions are:
1- Is everything I wrote accurate?
2- Would inputs and outputs going to and from the PLC be considered continuous or discrete? Do scan times and RPIs play into this?

I've found A LOT of conflicting information online, so hoping to maybe get some more concise info.

Thanks!

 

[synchro]

Continuous- at any given time (x axis) you can measure your value (y axis).
Discrete- values (y axis) will only be available at pre determined times (x axis). i.e. every .5 seconds you take sample data, but don't know the value in between these snapshots.

Yes, there are continuous time and discrete time signals as you described. A discrete time signal consists of "samples"at successive times, usually along with a clock signal that establishes when the sampled signal provided is taken and/or is valid.

There is another concept that needs to be considered, which is quantization of the signal values (y-axis). Digital signals are quantized such that they correspond to integer multiples of a quantization level representing the resolution of the digital signal. The number of quantization levels and therefore the range of the digital signal values will be limited by the width of the digital word provided (i.e, its number of bits). Of course, the full range of the digital word may or not be available in practical equipment.

Some confusion may are arise because digital signal values are discrete in the sense that they represent specific defined values and not a continuum of possible values. But that is a result of the signal being quantized, and not directly from its occurrence at discrete times. There can be discrete time signals that are not quantized into a finite set of values, but still maintain a continuum of values. But usually such signals are internal to signal processing equipment, such as in switched capacitor filters. And so they are usually not important unless you are designing equipment.

Analog: Values range is not defined, the value can be any number
Digital: Values are set at defined numbers. i.e. when a switch is activated, its 1. When a switch is not activated, its 0.

Any practical implementation of equipment handling analog signals will have a limited signal range that it can handle.
The lowest signal it can reliably process is limited by a noise floor. The maximum level is determined by clipping or excessive distortion. And so there will be a "dynamic range" between these maximum and minimum signal values.
A significant advantage of digital is that after its resolution is established from quantization (for example by an analog to digital converter), then the resolution will not degrade as long as bit errors do not occur in subsequent equipment. Error correction and other techniques can prevent this.

2- Would inputs and outputs going to and from the PLC be considered continuous or discrete? Do scan times and RPIs play into this?

An analog input itself into the PLC will be a continuous time signal, but the scan cycle and/or RPI will establish valid sample times of this signal.
See for example:
https://www.kb-controls.io/learn-logix/what-is-rpi-in-a-logix-5000-system

 

[LarryFine]

Digital vs analog is a matter of the nature of the type of signal, like in audio/video systems, determined by the source, and which the receiver must be able to accept, and continuous vs discrete is a matter of the nature of the desired reading of the signal.

Audio/video signals are continuous, because we want to observe an unbroken stream of sound and image, while an instrument signal could be periodic, because a continuous signal might be a waste of bandwidth when a regular sampling is enough.

 

[zbang]

To muddy the waters-
All signals are analog in transmission (usually light or voltage/current levels). Whether to treat them as digital happens somewhere up the stack and even "analog" electronics do this. For example, a relay has some power level when the coil has enough to pull the contacts, might be 10v for a 12v relay, might be less, won't be 1-2 volts. The same goes for PLC inputs, there's a mid-range level where OFF becomes ON.

 

[recklessconduct]

To muddy the waters-
All signals are analog in transmission (usually light or voltage/current levels). Whether to treat them as digital happens somewhere up the stack and even "analog" electronics do this. For example, a relay has some power level when the coil has enough to pull the contacts, might be 10v for a 12v relay, might be less, won't be 1-2 volts. The same goes for PLC inputs, there's a mid-range level where OFF becomes ON.

So where or when "off" becomes "on" could become a whole other subsystem and it's importance determined by the need for fine tuned resolution and the equipment to detect i.e. "4-20ma" signal seems like things can get complicated and branch off into other fields and perception of reality and self awareness all together. I think it's easy to take one idea and completely turn it into a whole other idea and we get off track and off subject being confused, lost, twisted and burnt out because we didn't take notes of when this change happend. We got too deep and now we are in some padded room and can't move our arms because we pissed off some,,,,,,,,,

 

[zbang]

Whole other subsystem? Not really, it's inherently part of the devices and ought to be on the mind of whomever is designing with them.

 

[Carultch]

So assuming my understanding is correct (which hopefully it is), it's not fair to just say "an analog input" or "a digital input". Because it seems as though you have 4 options. Continuous Analog, Discrete Analog, Continuous Digital, and Discrete Digital.

You'd be correct. There are four options, because you have the option to have a continuous value or a discrete value, for both the domain and range. Since the option for the domain is independent of the option for the range, that adds up to 4 possibilities. The domain would usually be time (could also be frequency, a topic for another time), while the range would be the value represented by the signal.

Discrete and continuous are terms that describe the domain. I.e. do you sample once a millisecond (discrete)? Or is the signal always there, and always has the chance of being something different, no matter how small the unit of time?

Analog and digital are terms that describe the range. If the value represented by the signal could be any real number within a given interval, it is analog. If the value represented by the signal is limited to numbers along a "staircase" of possible numbers, that's digital.

 

[4-20mA]

snip

So my questions are:
1- Is everything I wrote accurate?
2- Would inputs and outputs going to and from the PLC be considered continuous or discrete? Do scan times and RPIs play into this?

I've found A LOT of conflicting information online, so hoping to maybe get some more concise info.
snip

#1, I can't answer #1 and say 100% yes, but you are well on your way

#2 , to me, the I/O on any PLC or DCS is just that, an input or and output. How you look at and what you do with those with the plc/dcs, is what makes them discrete or continuous or otherwise