Analog versus digital, what next?

Brad Maisto

Registered User
Oct 1, 2000
743
215
43
66
Grandview, IN
Country
Region
Not sure if this is allowed, but I “borrowed” this picture from the “word/genius” website and felt a few NAWCC members might get a chuckle out of this?
Analog (definition) - Relating to or using signals or information represented by a continuously variable physical quantity such as spatial position, voltage, etc. The picture below was part of the examples they used to demonstrate this word!
6186568C-AF2D-4611-8865-C1AEC3D84DE6.jpeg
Maybe pretty soon this is how vintage clocks will be referred to by current generations of humans? But this also reminds me of my coal-powered electric generating plant a couple years back and this facility’s control panels were still all controlled by “analog” systems as a way to prevent “digital hacking”! I was impressed.
Brad Maisto, KY Floral #44 Secretary
 
  • Like
Reactions: Royce

Dr. Jon

Moderator
NAWCC Member
Dec 14, 2001
6,544
938
113
New Hampshire
Country
Region
I good example of misunderstanding of analog when applied to clocks and watches. Unless the clock is a synchronous motor, with its rate controlled by power plant frequency, which they used to do, all of these clocks are digital. If spring driven they have escapements move their hands in discrete increments, iff battery powered and quarts regulated they have stepper motors. The only truly analog timekeepers I know iare the Seiko spring drive watch and synchronous motor clocks such as Telechron.

The ultimate time standard, a cesium fountain clock, is a hybrid since it counts cycles or optical radiation. The optical radiation is analog but it's time signal is digitized by counting cycles.
 
  • Like
Reactions: Brad Maisto

Tim Orr

National Membership Chair
NAWCC Fellow
NAWCC Member
Sep 27, 2008
1,628
305
83
Boulder CO
Country
Region
Good afternoon, all!

And then there are the physicists who insist that time increments are, in fact, actually discrete, and that there is no "analog" in time.

Best regards!

Tim Orr
 

Dr. Jon

Moderator
NAWCC Member
Dec 14, 2001
6,544
938
113
New Hampshire
Country
Region
Not sure what ball clock is but clock that tells time with balls falling into slots is obviously digital. Rolling ball clock is also digital.
 

zedric

NAWCC Member
Aug 8, 2012
1,646
295
83
Country
Region
A standard mechanical clock or watch, to me, is an analogue to digital conversion machine - the spring or weight line, balance wheel / pendulum are analogue. The escapement turns that into a digital signal, and then anything from there on becomes digital. There are other fully analogue timekeepers, eg hourglass, candle or water clock..
 

etmb61

NAWCC Member
Oct 25, 2010
2,859
295
83
Mascoutah, IL
Country
Region
I can see that a swinging pendulum is a continuously variable physical quantity, hence analog, but I fail to see how the operation on an escapement creates a series of digits in the form of 1's and 0's. Please explain.

Eric
 

zedric

NAWCC Member
Aug 8, 2012
1,646
295
83
Country
Region
An escapement converts a continuously variable swing from a pendulum or balance into a discrete sequence of movements (Tick-tick-tick) that are now no longer continuous but stepped in nature. While not digital in the sense of 1 or 0 it is no longer analogue movement. Those of us that worked with analogue computers before digital computers took over might appreciate the difference. Eventually quantum computing will do to digital computing what digital did to analogue....
 

jmclaugh

Registered User
Jun 1, 2006
5,373
253
83
Devon
Country
Region
You could say the difference is in the display, digital shows numbers only in the format HH:MM while analogue uses hands or one hand moving round a dial marked with hours and minutes or subdivisions of.
 

Dr. Jon

Moderator
NAWCC Member
Dec 14, 2001
6,544
938
113
New Hampshire
Country
Region
The pendulum, balance and spring, tuning fork, quartz crystal and similar devices are forms of oscillator and are analog but when they are driven by an escapement which also counts oscillations, the result is digital and the display is a digital count. This is also true for oscillators driven electrically or electronically such as quartz crystals provided the the display is a count of cycles as in a typical quartz watch or clock, with the Seiko spring drive as an exception since its display is driven continuously.

I consider an hour glass to be a digital device since it has grains of material, usually sand. Candles are analog. My view is molecules converted vaporized or dripped are analog processes.
 

etmb61

NAWCC Member
Oct 25, 2010
2,859
295
83
Mascoutah, IL
Country
Region
How does an escapement count? I can see that it divides the movement of a wheel. The information on the dial is presented by a continuously variable train of wheels.
 

Bruce Alexander

Sponsor
NAWCC Brass Member
Feb 22, 2010
7,579
858
113
Hanover, PA
www.testoftimeclocks.com
Country
Region
Interesting. I suppose a Sundial, Incense and Water Clocks would be considered "Analog"?.

I think that perhaps the human interface is a good demarcation between digital and analog. Not many folks outside of Horology are going to look much deeper than the dial/face.
 
  • Like
Reactions: Brad Maisto

Dr. Jon

Moderator
NAWCC Member
Dec 14, 2001
6,544
938
113
New Hampshire
Country
Region
Escapements do not count. They cause gear trains to move in discrete nearly uniform increments. The hands indicate the count of these. For an 18,000 beat per hour watch or clock one turn of the seconds hand for one minute indicates 300 counts. Some English center seconds hand watches had minutes divided into five blocks so you can see the beat count.

Sundial and incense "clocks" are analog IMHO, as are most water clocks; but Su Sung, in China, is widely believed (by historians of China and Science) to have invented escapements which were on a water clock.If so he built the first digital clock.
 

etmb61

NAWCC Member
Oct 25, 2010
2,859
295
83
Mascoutah, IL
Country
Region
I'm still at a loss here how the ticking of a clock can be considered digital at all. The way I see it a binary signal that caries information is digital, and a binary state that caries no information is not. The switch that turns on my lights is not digital no matter how accurately spaced I switch it on and off, nor is the escapement of a mechanical clock. The ticking of a clock only tells that the clock is running. 01000100 01101001 01100111 01101001 01110100 01100001 01101100 00100000 01110011 01101001 01100111 01101110 01100001 01101100 01110011 00100000 01100011 01100001 01110010 01110010 01111001 00100000 01101001 01101110 01100110 01101111 01110010 01101101 01100001 01110100 01101001 01101111 01101110 00101110

(Digital signals carry information.)

Eric
 

zedric

NAWCC Member
Aug 8, 2012
1,646
295
83
Country
Region
An analogue signal is continuously variable, a digital one is not. Binary is certainly one system that is digital, but there are an infinite number of other digital systems.

A binary signal, or a digital signal in any other format for that matter, does not need to carry information.

For a signal to be digital, it just needs to be a signal that has a discrete, or stepped characteristic, rather than one that is continuously variable.

In terms of information theory, an escapement samples a continuously variable signal and converts it into a digital one.
 
  • Like
Reactions: Dr. Jon

zedric

NAWCC Member
Aug 8, 2012
1,646
295
83
Country
Region
which says


Of signals, information, or data: represented by a series of discrete values (commonly the numbers 0 and 1), typically for electronic storage or processing.Such data is commonly represented by discrete values of a physical quantity such as voltage or magnetic polarization, typically in binary form.
 

etmb61

NAWCC Member
Oct 25, 2010
2,859
295
83
Mascoutah, IL
Country
Region
which says


Of signals, information, or data: represented by a series of discrete values (commonly the numbers 0 and 1), typically for electronic storage or processing.Such data is commonly represented by discrete values of a physical quantity such as voltage or magnetic polarization, typically in binary form.
So you agree that the series of discrete (individual, separate, and distinct) values of a physical quantity must represent data to be digital?
 

zedric

NAWCC Member
Aug 8, 2012
1,646
295
83
Country
Region
Hi etmb61. I think we are heading too far off topic here to be useful. But no, I don't agree with your statement.

The OED definition quoted above, lists signals, information or data as examples of things that can be digital, but this is not an exhaustive list. And as stated before, digital data can be coded in a number of ways, with binary and decimal being common, but with an infinite number of possibilities, and a digital signal (binary or otherwise) does not need to carry information (in the common sense of the term) to be digital

For example, a randomly fluctuating analogue signal that is sampled via a thresholding device to convert into a series of 1s and 0s (binary) is converted into a digital signal, but it does not carry information in the common sense of the term.

Another example is semaphore, which is a digital signalling code, certainly not binary. In this instance the flags represent "data".
 
Last edited:
  • Like
Reactions: Dr. Jon

etmb61

NAWCC Member
Oct 25, 2010
2,859
295
83
Mascoutah, IL
Country
Region
Analog vs digital is the topic. I'm searching to understand y'all's thinking, but if that's off topic I'm out.

Eric
 

Dr. Jon

Moderator
NAWCC Member
Dec 14, 2001
6,544
938
113
New Hampshire
Country
Region
I suggest the topic is digital as applied to clocks and watches. at least that is my view. Clocks and watches are devices to keep and display time for a viewer. Devices to control timing are timers and perhapps theor output as digital or analog is closer the OED definition provided here. As I read it the entry was a descrition of digital data rather than a general definition of digital processes.

One aspect of clocks and watches being digital is that the gear trains are alwys "loaded" and do not backlash excpet perhaps when being set backwards.
 

etmb61

NAWCC Member
Oct 25, 2010
2,859
295
83
Mascoutah, IL
Country
Region
I suggest the topic is digital as applied to clocks and watches. at least that is my view. Clocks and watches are devices to keep and display time for a viewer. Devices to control timing are timers and perhapps theor output as digital or analog is closer the OED definition provided here. As I read it the entry was a descrition of digital data rather than a general definition of digital processes.

One aspect of clocks and watches being digital is that the gear trains are alwys "loaded" and do not backlash excpet perhaps when being set backwards.
I've been searching for any published reference that offers the definition of digital as it applies to clocks that y'all are suggesting with this thread to no avail. What is the source of your information?

Thanks,
Eric
 

glenhead

NAWCC Member
Nov 15, 2009
1,224
242
63
64
Williamson County, Texas
Country
Region
Analog things present an analogous representation of other things. A clock with hands and a standard configuration shows a twelve-hour visual representation of the passage of time. Mankind decided that the representation should be divided into twelve slices, and refers to those slices as "hours". The twelve slices starting when the little hand is at the 6 and ending when the little hand is at the 6 again is roughly analogous to "daytime" if the sun is up or "nighttime" if the sun is down. Yeah, yeah, the sun is rarely up for twelve hours starting at 6 and ending at 6; it's not a pure analog. With an analog display you can tell at a glance if it's "morning" or "afternoon" without having to convert numbers.

In the context of timepieces, "digital" means the timepiece shows the "time" with numbers or some other representation that can be mentally converted to numbers. The time shown by a 1, a 2, and a 7 is a digital representation that the analog timepiece shows by having the little hand a bit past the 1 and the big hand a bit past the 5.

On a dashboard there are analog displays (with pointers) and digital displays (with numbers). From an ergonomic perspective it's far more efficient to show "normal" with an analog display. If "normal" is represented by straight up the eye can see at a glance whether the attribute being represented is "normal" or not. A digital display requires interpretation.

If you want to get into analog versus digital telecommunications we can. I taught that to telecom engineers for many years.

Glen
 

zedric

NAWCC Member
Aug 8, 2012
1,646
295
83
Country
Region
I also come from the world of electronics - telecoms, signal processing, Analogue to Digital converters etc.. Mostly this was learned in University courses, and many of the standard texts cover the topic, although I'm sure they have been updated since.. But here is a website covering the topic in simple terms Analog vs. Digital - learn.sparkfun.com
 

etmb61

NAWCC Member
Oct 25, 2010
2,859
295
83
Mascoutah, IL
Country
Region
A mechanical clock movement provides a shaft rotation as output. The angle of that rotation is the analog representation of the period of the pendulum. The escapement converts the period of the pendulum to the angle of the shaft rotation as set by the wheel tooth counts when the movement was designed. If you put a pointer on the shaft you can see the analog representation of the passage of time based on the period of the pendulum as shown by the position of the pointer. Nothing digital in there.

Eric
 

zedric

NAWCC Member
Aug 8, 2012
1,646
295
83
Country
Region
The way I see it, and clearly there are a range of different interpretations depending on how fine grain you look at it, it this...

A pendulum or balance makes a set of continuously varying movements, from one direction to another. For a standard mechanical clock to show time, it must sample this continuously varying movement, and turn it into a set of discrete movements. The escapement is the thing that does this sampling, and from the escapement onwards, all movements are a series of discrete, stepped movements, which are no-longer continuously variable, but sampled. In other words, digital. (although you could argue that recoil escapements produce a set of continuously varying movements in the gear train, and so are analogue).

In effect, the gears help to count these stepped movements, and turn them into the movement of the shafts for the seconds, hour and minute hands. These hands in turn move in steps, even if it is not obvious to the human eye in some cases, and so are displaying digital movement, even though a clock face is often considered an analogue display
 
Last edited:

etmb61

NAWCC Member
Oct 25, 2010
2,859
295
83
Mascoutah, IL
Country
Region
The escapement doesn't sample. Sampling is the action of taking a small part or quantity of something for analysis, like when my wife makes cookies I always take some quality control samples. Sampling is a discrete event. Continuous sampling would imply that a number of samples are taken when the escapement is locked. Where do they go? The escapement simply divides the rotation of a shaft into angular segments in response to the movement of the pendulum. Those segments advancing don't indicate the period or amplitude of the pendulum's oscillations or the position of the pendulum in an oscillation or the direction the pendulum is moving, they only indicate that oscillations occurred. Simple division of a quantity into steps doesn't make the results digital.

All mechanical clocks contain the value of Pi built into their mechanisms. Does that make them irrational?

Eric
 

zedric

NAWCC Member
Aug 8, 2012
1,646
295
83
Country
Region
OK, I don't think we are getting anywhere with this - in the context above I used sample to refer to taking a measurement of, but I suspect that any verbal description I use is not going to work - so lets try graphs and a thought experiment.

Imagine a clock with a recoil-less escapement, and a one-second pendulum, and lets focus on the movement of the seconds hand. Assume the seconds hand is pointing to 60 (ie the top of the dial), and the clock face is in a vertical plane, much like most clocks are. Now, make a plot of the vertical position of the tip of the seconds hand on the y axis against time, in quarter seconds, on the x axis. You get something like the hand drawn graph below. If you zoom in close on that graph, it is not a continuous curve, but a stepped graph, in a similar fashion to the stepped graph in the link I shared in post 30 (Analog vs. Digital - learn.sparkfun.com) .

Now, imagine that same clock, but with a pendulum of 30 seconds duration. And imagine plotting the position of the hand on the y axis, against time in seconds on the x axis. You get a graph much like the hand drawn one in figure 2. This, I would imagine, is what you would consider to be a familiar digital waveform consisting of "1"s and "0s"s. It is similar to the waveform produced in Figure 1 in that it too, has a stepped characteristic.

IMG_3436.JPG

This to me demonstrates that the movement of the shaft on a standard mechanical clock is done in discrete steps, and that meets the OED definition of digital.

Of signals, information, or data: represented by a series of discrete values

Now, of course no physical system can be entirely digital - we live in an analogue world after all. So the graph in both cases is not exactly as shown - the hand in the second figure for example takes a finite time to flip from one position (pointing to 60) to the alternate position (pointing to 30), so the waveform is not quite as square as is shown in the graph, and the downward and upward transitions have some slope. That is true for pretty much any signal - in modern telecommunications for example, signals are sent representing 1s and 0s. Accepting standard terminology, this is a digital signal, but the actual waveform has rise and fall times and so is not truly digital if you sample (measure) it at short enough time intervals. But accepting standard terminology, a modern telecoms signal is digital, and so is a mechanical clock.
 

etmb61

NAWCC Member
Oct 25, 2010
2,859
295
83
Mascoutah, IL
Country
Region
I love a good mental exercise. My problem with your explanation is you don't seem to take into account what's going on at the escapement. Let's just say you have some form of deadbeat escapement. For some period of the pendulum's swing the escape wheel (and every other wheel in the going train) is locked. Next for some period the escape wheel is unlocked and accelerates to provide an impulse to the pendulum. Next the escape wheel drops off the impulse surface and accelerates unchecked to the next lock. The resulting graph would look something like this:
escape wheel 1.jpg
There are not two discrete states, but one fixed state and one accelerating (variable) state.
 

zedric

NAWCC Member
Aug 8, 2012
1,646
295
83
Country
Region
Lets deal with the escape wheel another time. Can you address the question of whether, to your thinking, the seconds hand has multiple, discrete physical states, moving from one to another in a stepwise fashion?

Moving to philosophy, any real physical system above the quantum level is, at its core, analogue. That telecommunications signal flying along the cable in your street, bringing this message to you and supposedly consisting of 1s and 0s, does not, in fact, consist of 1's and 0's, any more than the signal flying across the processor in your PC consists of 1s and 0s. Electrical lines have capacitance and inductance, so it is impossible to instantly change a signal from 1 to zero. This means that just as a seconds hand will accelerate and then stop, so an electrical impulse will not reach its "high" or "low" state instantaneously. So while on a macro scale, "digital electronics" is digital, on a micro scale it is far from that - which is why a decoder for a telecoms signal is not a simple thresholding circuit, but includes things such as timing circuits to ensure sampling is carried out at the right time - but I digress.

For real, physical systems, whether you consider them analogue or digital therefore depends on perspective - if you are a user of a computer, that processor is digital, but if you are a processor designer, you need to deal with the analogue nature of real world signals..

My perspective looking from a macro scale is that a mechanical clock is digital, driven by discrete impulses from the escape wheel. But if you look at the micro level, as it seems you want to do, you can see that the escape wheel is undergoing physical processes such as acceleration and deceleration, gradual wear and so on, and these can be viewed as analogue processes.
 

John Matthews

NAWCC Member
Sep 22, 2015
2,950
1,357
113
France
Country
Region
I have followed this discussion with interest.

It seems to me that if you consider any object that moves from one stationary position to another, be that motion angular or linear, you can consider an initial stationary state, followed by a period of acceleration, then there may be a period of fluctuating velocity, but always culminating in a decelerating period to reach a new stationary state. As I see it you could not describe that as digital motion. However, if you had an independent observer counting the stationary states and ignoring the motion path, he/she might consider it to be a digital system - is that not what is happening here?

John
 
  • Like
Reactions: Tim Orr and etmb61

etmb61

NAWCC Member
Oct 25, 2010
2,859
295
83
Mascoutah, IL
Country
Region
The movement of the seconds hand is an interrupted rotation with two states, stopped or accelerating. But those states mean nothing without external information that is not contained in those states. If you have a sampling of one second states, what time is it? The states themselves don't convey any information but that a second has passed. The 60 separate stopped positions of the seconds hand convey the accumulation of those seconds but only when compared to an external reference.

The closest I can come to a digital representation of time in a mechanical clock is a striking clock with a count wheel. The count wheel is divided into a series of discrete states, high and low, that represent the hour in an encoded form. The states are tested by the striking mechanism. If the state is high the mechanism strikes, increments the wheel and tests the state again. If the state is low the mechanism stops.
 
Last edited:

Tim Orr

National Membership Chair
NAWCC Fellow
NAWCC Member
Sep 27, 2008
1,628
305
83
Boulder CO
Country
Region
However, if you had an independent observer counting the stationary states and ignoring the motion path, he/she might consider it to be a digital system - is that not what is happening here?
Good afternoon, John!

Wow! I just flashed on the Paradoxes of Zeno!

Best regards!

Tim
 
  • Like
Reactions: John Matthews

etmb61

NAWCC Member
Oct 25, 2010
2,859
295
83
Mascoutah, IL
Country
Region
One more note. Digital as it applies today is a computing term first proposed by a mathematician at Bell Labs in 1942. History of Computers and Computing, Birth of the modern computer, Relays computer, George Stibitz
Prior use of the term digit meant the fingers on your hands or the numbers 0-9. Prior use of the term digital meant of the digits, as in using the fingers to perform a task, like playing an instrument or giving a prostate exam. Digital in computers refers not to the discrete states themselves, but to the information encoded in a series of those states.
 

zedric

NAWCC Member
Aug 8, 2012
1,646
295
83
Country
Region
I have followed this discussion with interest.

It seems to me that if you consider any object that moves from one stationary position to another, be that motion angular or linear, you can consider an initial stationary state, followed by a period of acceleration, then there may be a period of fluctuating velocity, but always culminating in a decelerating period to reach a new stationary state. As I see it you could not describe that as digital motion. However, if you had an independent observer counting the stationary states and ignoring the motion path, he/she might consider it to be a digital system - is that not what is happening here?

John
Hi John

In the real world, digital signals often look anything but the sharp "1"s and "0s" of theory. In telecoms, for example, signals will have rise time and fall time, they will have experienced non-linearity in their path which can lead to timing jitter, they will have overshoot and undershoot, etc etc. If you plot a long series of "1"s and "0s" as received at the end of a transmission line, one on top of each other, you end up with what is called an eye diagram. In theory, the centre portion of this should be a perfect rectangle (the transition from 1 to 0 should be sharp, and every 1 and 0 should have the same amplitude). In practice, it looks something like the plot below.

Eye-diagram-and-eye-mask.png

Even with the blurred edges, this is still very much a digital signal, just one that has interacted with the real world! To recover the signal to a series of 1's and 0's you would sample it - ideally at the sampling point shows. Any signal at the sampling point that is above the threshold of the eye mask is considered a 1, and anything below a 0. For a clock, you would do the same to demonstrate the digital nature - looking at the hand at some point after the acceleration/deceleration event, and measuring the angle of rotation, will get you to the digital signal being sent by the clock.
 

John Matthews

NAWCC Member
Sep 22, 2015
2,950
1,357
113
France
Country
Region
Hi Zedric - I think ..

For a clock, you would do the same to demonstrate the digital nature - looking at the hand at some point after the acceleration/deceleration event, and measuring the angle of rotation, will get you to the digital signal being sent by the clock.
is equivalent to my

an independent observer counting the stationary states and ignoring the motion path, he/she might consider it to be a digital system
:) John
 

Brad Maisto

Registered User
Oct 1, 2000
743
215
43
66
Grandview, IN
Country
Region
I have found this discussion interesting so I forwarded it to my electrical engineer who gave me the tour of our local coal powered power plant. Here is his take on these matters:
“Those guys get pretty obsessed with what's digital and analogue. I know in
the old days AEP controlled the entire Eastern United States frequency and
it had to cross 60 HZ cycles in the positive and negative direction so many
times by law. I think they still control it in Columbus, OH. So your
clocks could run slow or fast if they were the electric plug in the wall
type.”
I believe he worked from the 1970’s until just a few years ago.
Thanks Again for this lively discussion, Brad Maisto
 

Forum statistics

Threads
167,147
Messages
1,456,506
Members
87,334
Latest member
fredericdb
Encyclopedia Pages
1,057
Total wiki contributions
2,914
Last edit
E. Howard & Co. by Clint Geller