An electrically TRUE square wave possible?

fs/2 is the so-called Nyquist Frequency in your example, right?

And, by using a hardware signal generator, I'm assuming that is something much more specialized and accurate than a run-of-the-mill Realtek ADC/DAC, of lesser capability, trying to generate the square
Yes and yes. A hardware signal generator will have bandwidth well beyond the audio range. The waveform has not been sampled.
Ed
 
I think the point has already been made, but I can offer an interesting example. I worked on a gigabit ethernet chip in the early 2000s. Each pair of cat 5 carries 250 Mb/s and at the transmitting side there are 21 different levels that are sent, not just low and high. At the receiver side, after the signal is equalized it is sampled (at 125 Ms/s) and converted from analog to digital. The resulting digital word represents the analog value of the signal at the sampling time (sampling at the correct times is very important). Then DSP is used to sort out the bits that were most likely to have been sent. Other blocks accomplish crosstalk cancelation and maximum likelihood sequence detection. It's surprisingly complicated.

Tom

This is the kind of commentary that keeps me coming back here.
 
I think the point has already been made, but I can offer an interesting example. I worked on a gigabit ethernet chip in the early 2000s. Each pair of cat 5 carries 250 Mb/s and at the transmitting side there are 21 different levels that are sent, not just low and high. At the receiver side, after the signal is equalized it is sampled (at 125 Ms/s) and converted from analog to digital. The resulting digital word represents the analog value of the signal at the sampling time (sampling at the correct times is very important). Then DSP is used to sort out the bits that were most likely to have been sent. Other blocks accomplish crosstalk cancelation and maximum likelihood sequence detection. It's surprisingly complicated.

If only a low or high was sent, and the receiver had to wait for the signal to settle to something close to its sent value, the data rate would need to be much slower.

Tom
And I was the guy who, sorry to say when you did all the hard work!, that took advantage of available functionality like "IEEE 802.3ad" in easily managed via web interface (or laptop directly connected to) switches where I would run FOUR gigabit lines between them and servers and even some servers to the switch and then to workstations (some users did work with quite large files even back then and me at home especially with the 'side business' using huge databases, full compact discs in just one ISO file (650MB+), and then HDTV sports events to edit video files 20GB and larger). You did awesome for people like me who just followed the instructions in the Cisco book. Like I said earlier in this thread, I wish I had learned much more electrical engineering. I will try. 🙂
 
Read the Wikipedia article entitled Rise Time. You'll enjoy their derivation of the formula (Bandwidth x Risetime) = Constant.

Since a circuit which produces an "electrically TRUE square wave" is required to have a risetime of zero femtoseconds (in order to be perfect), it must also have a bandwidth of infinity. Which is not possible currently with the devices we have available.

_
 

Attachments

  • rise.png
    rise.png
    4.6 KB · Views: 63
I remember arguing with a co-worker who wouldn't back down from his claim "CDs are 100% perfect in any way." My "boss" had exactly the same misconception.

Marketing is just as corrosive to intellect as religion. Marx was spot on when he said "religion is the opiate of the masses."

I just got out of the hospital. They fed me a steady cocktail of drugs "the good stuff" through the IV. I could literally feel my intellect fading as the drugs coursed through my system (not an unpleasant feeling, especially when it's quelling the searing pain of post surgery.) I can see the same thing happening to people when exposed to marketing lies and religious drivel. "YOU MUST BUH-LEAVE." It's like free and legal morphine. Repeated exposure will permanently dim your intellect. Exhibit A; Murica.

View attachment 1118110
I remember the later 90's when the bands that made actually great music would somehow be convinced to 'loudness brickwall' their final product. The dynamic range was then down to around 3 to 6dBfs then. It was horrible. One of my best examples is Radiohead's "OK Computer". They brickwalled that one to the point of distortion and gave away all of the dynamic range. I'm hoping that a company like MFSL (Mobile Fidelity Sound Labs) gets that master tape and remasters it properly. That outfit has done a great job with other works released on CD some or many years ago and cleaned the sound up.

As for those who complain about the opposite, too much dynamic range, and especially with classical music that starts out quiet but then builds to a thunderous crashing, I tell them to get more powerful amplifiers and better speakers. And, as far as I know, the rule of thumb still is to provide a level of continuous amplifier wattage that is greater than what the speakers (quality ones) are rated for to eliminate electrical clipping and only have to worry about mechanical clipping.
 
Read the Wikipedia article entitled Rise Time. You'll enjoy their derivation of the formula (Bandwidth x Risetime) = Constant.

Since a circuit which produces an "electrically TRUE square wave" is required to have a risetime of zero femtoseconds (in order to be perfect), it must also have a bandwidth of infinity. Which is not possible currently with the devices we have available.

_
I will do so. That is again exactly along the lines of what I was trying to explain to my friend. But, discussing 'zero femtoseconds' and an 'infinite bandwidth' wouldn't have gone well. (grin)
 
In practice a 'digital' system is more robust and will work properly down to lower received carrier levels than an analogue system simply because the data can still be extracted cleanly from the noisier low level received signal... up to a point... and then it all falls apart.
In the late 1990's Dutch fire brigades, police and ambulances went from using narrowband FM to a digital system, C2000, because the government had decided so. It was a complete disaster: previously, when the reception was poor, the signal to noise ratio was below spec, but they could still understand each other. With the digital system, they would all of a sudden lose contact. Some fire brigades temporarily went back to narrowband FM. Nowadays the C2000 base stations have huge masts for the aerials to get acceptable coverage.
 
  • Like
Reactions: mterbekke and Mooly
I would also add that a square wave with zero rise time would be "dirty" (i.e. rich in harmonics) to the point that there could be EMC issues with nearby electronics such as receivers and other sensitive components.

Some systems implement slew rate limiting for that very purpose.
 
As for those who complain about the opposite, too much dynamic range, and especially with classical music that starts out quiet but then builds to a thunderous crashing, I tell them to get more powerful amplifiers and better speakers.

Perfect Circle. My 40 watts continuous/channel can't play them very loud.

A partial solution is to use a steep slope high pass filter @20 Hz. Their mixes contain a whole lot of subsonic trash. I imagine DSP could address this issue.
 
I created "perfect" square wave signal on CD artifically, by editing 0s and 1s into an 1 kHz pattern .wav and converting ot to audio CD format. By playing it back, you get a square wave limited by the analog part of the CD-player. But I understand that even the DAC has its own rise/fall time limits. Anyway, it is good enough for test purposes, because it does not contain the limitations of the recording chain.
 
I would also add that a square wave with zero rise time would be "dirty" (i.e. rich in harmonics) to the point that there could be EMC issues with nearby electronics such as receivers and other sensitive components.

Some systems implement slew rate limiting for that very purpose.
That's also the reason why digital wireless communication systems don't use square waves. They have to occupy no more bandwidth than necessary, so they use modulated sine waves with filtered modulating signals.
 
Many people will accept without question the lies and distortions of marketing people, while summarily rejecting the claims of experts like engineers that actually design the products.

Muricans are stupid and getting stupider. That's what makes Murica great.

View attachment 1118094
FYI, you should not be sharing photoshopped images without at least a warning of substantial editing.

This is the original https://i.kym-cdn.com/entries/icons/original/000/001/296/morans.jpg
 
  • Like
Reactions: ejp
lcsaszar - That results in aliasing. To avoid aliasing, you have to sum the harmonics only up to fs/2. The resulting waveform is intentionally bandwidth-limited.

If you need real square waves, get a hardware generator.
Ed
Yes. The resulting squarewave in the analog domain contains harmonics way beyond the sampling frequency. That is intentional in my case.
 
A friend and I got into a debate about what a 'digital signal, sent over-the-air, as in when receiving OTA HDTV from broadcast towers' appears like electrically. Pardon me if this isn't the best place to ask but it does apply to music too and I thought solid state amplifiers would be suitable.

The friend said that the 'digital signal', where his understanding was that the digital signal was that of a square wave, was 'perfect' and meaning in its 'squareness and exactly SHARP corners'. The signal was never, ever 'in-between' and only always exactly electrically representing a '0' or a '1'.

I argued that there is no way possible to create an absolutely perfect square wave signal, which would represent 0's and 1's, because there is always a bit of overshoot on the, say, on the 'way up', and then some over compensation resulting in a bit of undershoot, and that repeats but that signal does quickly 'level off' at the 'flat top' which is supposed to be there.

My argument was based on the fact that NOTHING, except only in the realm of mathematics perhaps, when in the real world, can go from say '0 volts to 1 volt' instantaneously and without any over and/or undershoot before it settles to its 'correct/wanted' amount/voltage/etc. So, therefore, my argument was that even broadcast 'square waves' cannot possibly be 'perfectly digital' (in his vision of the signal) which he thought was so.

And, furthermore, there isn't anything such as a 'digital antenna' which are named that and sold with those words. Maybe that was his problem with me arguing the topic because he had bought a whole new $100+ 'digital antenna' when then I showed him a few minutes later on that a simple piece of wire could receive the signal just fine as it did with the former NTSC broadcast signal (which everyone called 'analog television' or something like that).

Which of us is correct? I am curious.
Thanks!

I'd say that neither of you were discussing the same thing and talking at cross purposes!

The friend was saying that digital signals are transmitted with square waves and that's how the 1s and 0s are perfectly received.

Then you replied that you can't make perfect square waves in real life.

Whilst you may be correct in that, that wasn't what you said your friend was getting at, you just happened to pick up on one single aspect of his claim ... which really wasn't the point of it (as is being relayed to us at least).

Even if a square wave were to be the method of transmission, the threashold of conditions needed to switch states is where the accuracy of transmission would be key. The squareness of the wave, the rise time etc, would then only determine the timing of these state changes, a buffer and re-clocking will overcome that.

So - your friend's understanding of digital signal transmission is wrong, but his round-about assertion that the outcome of even a square wave transmission can be perfect, in the sense that the states will switch perfectly in relation to those changes in the signal is correct in theory .... but not becuase the square wave is perfect. At which point, arguing about the possibility of perfect square waves existing is a change of topic.

And then there's probably a lot reasons why square waves are not used to transmit digital signals anyway... so it's all moot.
 
  • Like
Reactions: mterbekke
FYI, you should not be sharing photoshopped images without at least a warning of substantial editing.

This is the original https://i.kym-cdn.com/entries/icons/original/000/001/296/morans.jpg
That's a daft thing to assert. It's not a point of information you're putting across, it's nothing more than an opinion.

If it's the photoshopping that contains the meaningful message then it's there for a reason and is the point being made.

You're kind of daftly saying that people should give an exhaustive history and origin of a photo whenever they use a meme... Do you know how memes come about and are used? They deliberately do not reference original sources of images - that is their point.

You should instead just be saying that you don't like the message that it's perveying and want it known that that message has been created by someone from another photo..

Of course, "should" or "should not" means nothing - I don't have any influence over you, nor you the previous poster - again, it's only an expression of my feelings, not an instruction.
 
FYI, you should not be sharing photoshopped images without at least a warning of substantial editing.

This is the original https://i.kym-cdn.com/entries/icons/original/000/001/296/morans.jpg

I had no idea.

The original picture is no more flattering to stupid Muricans. The stupid is quite close to home for me. As much as I try to get along with people, as an Official Smarty Pants I can't help but out blatant ignorance, and it's not always popular.

I often wonder if I had been born 20 years later whether or not I would have been one of those that slipped on their own slobber. I was inculcated with a strong work ethic and great value for education. I was taught that I could rise above my roots with education, integrity, hard work, and perseverance. And I did. Believe it or not, these used to be bona fide American values.

Sorry if you guys don't like it, but it pains me greatly to see my once great country becoming a shallow parody of its former self. I owe it to my father, who fought in WWII and would be devastated if he saw what's left of the country today.