May 17, 2012, 09:37 PM Flying Zayin Thread OP Thank you everyone, I'm quite happy to get reasonable answers and learn a lot about this aspect of our hobby.
May 18, 2012, 04:29 AM
Registered User
Quote:
 Originally Posted by AndyKunz Good advice! And at every level it's a little of each. And just when you think you got it as one, go down another level and it changes again Kind of like "is light a wave or a particle?" Andy
Good answer - I like the comparison to light.
This thread reminds me of the old joke.
"There are 10 kinds of people in the World; those who understand binary and those who don't."

I hope somewhere along the line the original question hasn't been left unanswered.
May 18, 2012, 05:58 AM
Proud to eat Kraut ;-)
Quote:
 Originally Posted by AndyKunz Yes, you DID get that mixed up. The signal from a receiver to a servo is digital - it is either ON or OFF. The duration of the signal is proportional - NOT analog - with the time indicating the position that the servo should move to. This is called Pulse Width Modulation, or PWM. [] Andy
I'm sorry Andy, but you have that wrong here.

"Proportional" has nothing to do with being analog.
Quote:
 two variable quantities are proportional if one of them is always the product of the other
(Wiki)

Our quantities are the servo deflection and stick movement, which are proportional. If the communication between the stick and the servo is digital or analog, is irrelevant when discussing proportionality.

What we see when we look at the pulsetrain a reciever puts to the servo, is a repeated analog signal.
A digital signal would have to consist of binary values, like 1 and 0.
I do not see these values.
What I see is one value that is variable, and hence analog.
The fact that it is repeated, makes it look like a digital signal to the unknowing eye, but a digital signal never changes pulsewith.

I have some scope readings at home showing how the pulse with changes from one frame to another with the flick of a switch, thus proving that the information is contained in the width of the pulse, and therefore analog.
May 18, 2012, 08:02 AM
Quote:
 Our quantities are the servo deflection and stick movement, which are proportional. If the communication between the stick and the servo is digital or analog, is irrelevant when discussing proportionality.
The travel of the servo is proportional to the width of the pulse, Julez.

Quote:
 A digital signal would have to consist of binary values, like 1 and 0. I do not see these values.
If you look at your scope, you will see that the pulse has two values - 0V and 5V (or 3.2V, or 4.8V...). It is digital because the height of the pulse has no effect on the servo's travel. A servo will respond to a 3V pulse the same as it responds to a 5V pulse (assuming it is able to see 3V as a logic high, which most modern ones are).

Thus, the 40+ year old "Digital Proportional" is still valid. The term was coined back in the 1960's. Its definition has remained unchanged for well over 40 years. Trying to change it now won't work.

Andy
Last edited by AndyKunz; May 18, 2012 at 08:42 AM.
 May 18, 2012, 08:50 AM Registered User Ah semantics again rears it's head as nit pickers enter the game --- The only thing worth mentioning -that Isee is that the servos which have internal " high speed " capabilities - are truly different.
May 18, 2012, 10:25 AM
Flying Zayin
Quote:
 Originally Posted by richard hanson The only thing worth mentioning -that Isee is that the servos which have internal " high speed " capabilities - are truly different.
In fact everything (or nearly so) said here was interesting - and the best thing is that behind all these semantic(al) discrepancies the question has got its answer, which is pretty simple: get an oscilloscope and measure the impulse framerate... or, get this rate from the manufacturer if there is no oscilloscope around!
 May 18, 2012, 10:51 AM Many servo testers will tell you frame rate as well, and are much cheaper than a scope! Andy
May 18, 2012, 11:40 AM
Registered User
Quote:
 Originally Posted by AndyKunz Many servo testers will tell you frame rate as well, and are much cheaper than a scope! Andy
easier yet - twist n listen
May 19, 2012, 06:46 AM
Proud to eat Kraut ;-)
Quote:
 Originally Posted by AndyKunz The travel of the servo is proportional to the width of the pulse, Julez. If you look at your scope, you will see that the pulse has two values - 0V and 5V (or 3.2V, or 4.8V...). It is digital because the height of the pulse has no effect on the servo's travel. A servo will respond to a 3V pulse the same as it responds to a 5V pulse (assuming it is able to see 3V as a logic high, which most modern ones are).[]Andy
You are mistaken because you believe that the pulse height carries any information. It is the pulse lenght which is important, as only the pulse length carries the information.

Quote:
 The travel of the servo is proportional to the width of the pulse, Julez.
Exactly! And now please tell me, what is digital in a variable pulse with.
By stating that something is proportional to the signal we discuss, you say yourself that this signal must be analog.

Quote:
 two variable quantities are proportional if one of them is always the product of the other and a constant quantity
(Wiki)
By stating that the pulse width is proportional to something, this must mean that it is variable.
Digital signals, in contrast, are discrete.
Quote:
 A digital signal is a physical signal that is a representation of a sequence of discrete values
(Wiki)

A signal cannot be variable and discrete at the same time.

Thus, by saying that it is proportional, you exclude the possibility of it being digital.

Quote:
 In computer architecture and other digital systems, a waveform that switches between two voltage levels representing the two states of a Boolean value (0 and 1) is referred to as a digital signal, even though it is an analog voltage waveform, since it is interpreted in terms of only two levels.
(Wiki)

Quote:
 Originally Posted by Andy Kunz It is digital because the height of the pulse has no effect on the servo's travel.
You are mixing things up again. The height of the pulse, and its absence, do not represent the two states of a Boolean value (0 and 1). The signal is not interpreted concerning its absence (0) or presence (1). The time difference between the rising and falling edge of the pulse carries the information.

This analog signal is repeated every 8-20ms, depending on the system. This repetition is something completely different to a digital signal, where a code of binary values is transmitted.
So, the pause between the analog pulses does not represent the binary value "0", it is simply a pause.
The pulse does not represent the binary value "1", the information is carried by its length.

If, as you claim, the signal form the RX to the servo was digital, it would have to consist of binary values.
If you have a resoultion of 1024 steps, you would need at least 10 digits, or bits.
(2^10=1024)
This means, you would need to have to transmit 10 binary values (bits), and hence 10 pulses being either 0 or 1, for each frame.

Please, by all means, hook one of your recievers on a oscilloscope, and show me this binary code.
It does not exist. You will find exactly one pulse per frame, not 10, or any other number. This is the proof that the information cannot be digitally encoded.
When there is no dicital code, there is no digital information. It is that simple.

I have put a lot of stuff on a scope, and will show some examples.

This are 2 channels of the reciever superimposed. Think away the second pulse of each pair, and you have what is going from the RX to the servo.
You can see that I rapidly changed the "servo position" with the flick of a switch, and you can see that the chance of the pulswidth occurs from one frame to the other. No digital code here, just the analog value of the pulsewidth.

Here I compared the pulsetrains of two different transmitters. This is one frame of the analog signal going from the TX to the RF module. You can see, that the channels 3, 4, and 5 have a different pulsewidth on both transmitters. In the RX, this signal is demultiplexed and the individual pulses are assigned to the different channels.

Here, you have a digital signal. One PCM1024 frame for all channels of a Futaba transmitter. This is the digital signal going from the TX to the RF module. This signal actually represents binary values. Such a shape would bee seen in the RX output if the signal from the RX to the servos were digital, as Andy claimed.
As it is not seen, the signal is analog, as I claim.

This is a comparison between different PCM systems. Each frame has 100-160 bits. It is plain to see that such a communication does not occur between the RX and the servo.

### Images

Last edited by Julez; May 19, 2012 at 07:48 AM.
 May 19, 2012, 07:37 AM Registered User I thought the question was about "digital" vs "analog" servos and the problems which may occur if the "analog" were used with fast rate output to the servo.
May 19, 2012, 08:09 AM
Quote:
 Originally Posted by Julez You are mistaken because you believe that the pulse height carries any information. It is the pulse lenght which is important, as only the pulse length carries the information.
No, that's exactly what it would be if it were ANALOG, but because it's NOT analog (it's digital) it does not convey any information other than a logic state (digital).

Julez, you're just going to have to face the fact that 40+ years ago they coined the phrase "digital proportional" and it's still correct usage of the term.

Andy
May 19, 2012, 08:34 AM
Proud to eat Kraut ;-)
Wait a moment.
Are you claiming, that the pulse lenght of the pulses the RX gives to the servos is irrelevant?

Let me tell you of a little experiment I made some years ago. After discovering I could use my soundcard as an oscilloskope to capture, for examples, the readings I posted above, I wonder if I could use it to create pulses that a servo would be able to understand.

I created a 50Hz hum with an audio program, and programmed the rectangular wave form to have different pulse lenghts.

It was a success. When I hooked the servo's signal wire to the audio output, it made controlled movements determined by the pulse lenghts I programmed.
This proves that the servo movement depends on the pulselenght, and nothing else.

I tell you something: If I capture the pulses that my reciever sends to the servos again, and I make a video of how they change their length when I move the stick, and how the servo moves according to the pulslength, would you accept that as a proof that the signals are analog?

Quote:
 No, that's exactly what it would be if it were ANALOG, but because it's NOT analog (it's digital) it does not convey any information other than a logic state (digital).
Ok, lets assume that the signal is digital. So you claim, that the one pulse we see per frame, conveys enought information to tell the servo where it should go? One pulse is exactly one bit. You would need, as I previously explained, at least 10 bit to achive a 1024 step resolution of the servo movement.
But there are no 10 pulses per frame, only one. How do you put at least 10 bit into one digital pulse?
Last edited by Julez; May 19, 2012 at 08:49 AM.
 May 19, 2012, 08:49 AM Julez, Not at all - the duration of the logic high and low signals (ie, digital signals) contains all the information that is to be conveyed. "Analog Electronics" generally deals with signals of varying amplitude (voltage). This is where you find capacitors, resistors, inductors, opamps, and such. The height of the varying voltages as well as the shape and duration all convey information. "Digital Electronics" generally deals with signals with voltages conveying two logic states (0/1), where there is often a hysteresis, and with pulses of varying durations. This is where you find things such as timers (which can be driven by analog (RC - resistor capacitor) circuits (part of the "is light a wave or a particle" comment)), transistors (generally saturated), etc. Because the amplitude of the pulse train carries no information other than logic state (0/1 or absence/presence of voltage), and the duration of the pulse varies in proportion to the information to be conveyed (but amplitude does not), the servo PWM (hey, another digital concept!) is "Digital" and "Proportional." "Analog" deals with voltage variations - "Digital" deals with pulse variations. Andy
 May 19, 2012, 09:05 AM Registered User I wondered when the basic differences (as we see em) would be noted. For most of us clodbusters , amplitude modulation is "analog" or analogous to analog -- PCM /digital - is bits of fixed amplitude and varied length (and rate ) info. Close enough for folk music? Play on, brother ---
May 19, 2012, 09:43 AM
Proud to eat Kraut ;-)
Andy, you are still mixing it up.

Quote:
 "Analog Electronics" generally deals with signals of varying amplitude (voltage).
Not only the amplitude can convey information. Tilt your frame of reference 90° clockwise, and you see that the "amplitude" is now similar to the pulselength. And in the same way, as an analog amplitude can convey information, the analog pulslength can do this as well.
""Analog Electronics" generally deals with signals of varying values."

The signal can be there or not, that is true. But the presence of the signal does not convey any information.
As you correctly state, the pulse length (duration) "contains all the information that is to be conveyed".
This pulse lenght is variable, and can therefore not be digital. If it was digital, the pulse lenght would need to represent the states "1" and "0".
But it has a large varienty of lenghts depending on the control information, not a fixed set of 2.

Quote:
 []"Digital" deals with pulse variations.
This is the core where you are wrong.
"Digital" never deals with variations, it only uses the discrete values "1" and "0".

If a signal is digital or analog, does not depend on the carrier of that signal.
Voltage can both be a carrier of digital or analog information, and so can pulselenghts. If a voltage-based signal represents discrete values (on "1" or off "0"), it is digital. If the voltage is continously variable, this is an analog signal.

If a pulselength-based signal represents discrete values (for example, 2ms for "1" and 1ms for "0"), it is digital.
If the pulselength is continously variable, this is an analog signal.

Your misunderstanding is based on the misconception that the signal is voltage based. The signal output of the reciever has discrete voltage levels. But they happen to be a neccessary byproduct of the continous repetition of the pulses, and do not carry information. As you state yourself, all the information is conveyed by the pulswidth.
So what you mistake for ones and zeros (pulse present/pulse absent) are in fact no ones and zeros, as noone attempts to convey any information this way.

Quote:
 A digital system[1] is a data technology that uses discrete (discontinuous) values. By contrast, non-digital (or analog) systems represent information using a continuous function.
(Wiki)

The pulse width is a continous fuction of the stick deflection. It does not use discrete values.

http://en.wikipedia.org/wiki/Analog_signal

I will quote from Wiki in the following text:

Quote:
 An analog or analogue signal is any continuous signal for which the time varying feature (variable) of the signal is a representation of some other time varying quantity, i.e., analogous to another time varying signal.
The variable we have is the pulsewidth, and it is the representation of the stick deflection.

Quote:
 An analog signal is one where at each point in time the value of the signal is significant, whereas a digital signal is one where at each point in time, the value of the signal must be above or below some discrete threshold
The value of our signal is the pulsewidth.
If it was digital, the pulse would have to have discrete thresholds, which is does not have.
Last edited by Julez; May 19, 2012 at 10:37 AM.