that should also have been factored in when working out what is left for the poor designer to work with.
There was a time in UK audio where 'hair shirt hifi' was in. Back then putting it in a plain metal box was a sign that you had spent the money in the right place. Naim cleverly chose an extrusion that looked quite nice with silver accents. Exposure just looked boring.
There was a time in UK audio where 'hair shirt hifi' was in. Back then putting it in a plain metal box was a sign that you had spent the money in the right place. Naim cleverly chose an extrusion that looked quite nice with silver accents. Exposure just looked boring.
With Rasperry pi and Arduino is at least educating people that PWM is a way to get a digital signal converted to an analog one.
But i guess it does't get easier by formats like PDM. Of course it is analog and can be treated as an analog signal. But it can also be treated as a digital signal and with digital interfaces.
And of cource you don't have to convert a PWM stream to analog. You could also treat PWM as an digital signal. (Not a loudspeaker element of course, but a MCU) So i'm not that suprised people get it mixed up.
^ Thank you!!!!!!!
This is a serious question. I'm not trolling. I swear. If it's OT, once again... apologies. Nomenclature / terms etc., with this topic and others cause a lot of confusion in my wee brain. It seems an appropriate place to ask.
Is it appropriate, in your opinion, to not refer to a signal AT ALL as analog or digital? It seems to me that what we do with that signal is the differentiating factor. One can use a continuous sine wave (often considered analog and perhaps it may be by definition) to actuate an ON OFF / 1 or 0 receiver as long as the receiver is set for such a purpose. Thus... in my wee brain... I have to try and think about what we're doing with the signal vs. the signal itself.
Simplified - Who cares?
This is a serious question. I'm not trolling. I swear. If it's OT, once again... apologies. Nomenclature / terms etc., with this topic and others cause a lot of confusion in my wee brain. It seems an appropriate place to ask.
Is it appropriate, in your opinion, to not refer to a signal AT ALL as analog or digital? It seems to me that what we do with that signal is the differentiating factor. One can use a continuous sine wave (often considered analog and perhaps it may be by definition) to actuate an ON OFF / 1 or 0 receiver as long as the receiver is set for such a purpose. Thus... in my wee brain... I have to try and think about what we're doing with the signal vs. the signal itself.
Simplified - Who cares?
PWM and digital are two different things.
Digital is a stream of numbers, where the actual analog properties of the stream (pulse levels, clock timing) do not matter (within the system limits of course).
PWM is a purely analog system, a stream of pulses where the levels and timing define the signal.
The last drawing in #122 nicely illustrates how you can convert an analog signal into a bunch of numbers, a digital signal.
Is this used in consumer equipment? (The pic says for measurment purposes).
Jan
Digital is a stream of numbers, where the actual analog properties of the stream (pulse levels, clock timing) do not matter (within the system limits of course).
PWM is a purely analog system, a stream of pulses where the levels and timing define the signal.
The last drawing in #122 nicely illustrates how you can convert an analog signal into a bunch of numbers, a digital signal.
Is this used in consumer equipment? (The pic says for measurment purposes).
Jan
The analog part in PWM, is the time variable.
In a micro controller you just sample this in discrete steps. (aka sample rate)
That doesn't make a signal all of a sudden digital.
@jan.didden was just literally 1 second before me 😀
Yes, those ones make me happy, I use them as well. 🙂
You want a super flat inductance vs current curve, and make sure the max current is linear until the PEAK current of the signal.
Extremely fair point, in fact that it's literally what it's all about.
Well, there is some marketing psychology involved as well.
Since the original UCD patent has been expired, I am still very curious to see what big major company is gonna pick this up.
Because from a PCB design perspective, a TPA32xx/TPA31xx is so much nicer vs a UCD discrete amplifier.
The only major difference is the Post Filter Feedback loop (there are a few other things here and there, but they are pretty minor).
In high quantities, it's easy to produce a extremely well performing Class-D amplifier for about 10-15 bucks no problem at all.
This is also the beauty for switching amplifiers, the need for matching components is a lot less!
You won't find any that also will pass you any EMC testing.
I don't know any professional amplifier, incl the Hypex, Purifi etc ones, that don't use a ferrite core.
As long as the curve is still linear and the inductor doesn't saturate yet, it's fine.
Hi Jan
Short answer: YES !
Longer answer: The TacT amps and also the TI ones derived from the same technology work like that. You need very fast fast counters combined with upsampling and noise-shaping in order to reach reasonable audio quality. If my memory serves me right they use a switching frequency of 352.8 kHz and 8 bit resolution for the timing of the PWM. The other famous Jan from this forum would know the exact details.......
On the TacT Millenium (and its successors) the loudness is set with the supply voltage of the output bridge in order to keep resolution high at low loudness settings.
Regards
Charles
P.S.: I quite like the idea of using GaN output stages because it makes the design of the feedback loop easier.
Why would that be?
The construction of a GaN switching amplifier or MOSFET amplifier are basically the same.
It's not the only subject in audio I think the same on a daily basis.
I suppose one of the major issues is newcomers (like me) that in spite of best efforts find conflicting and/or just plain incorrect information all over the internet. 🤷♂️ I'm sure it's funny and likely frustrating, but the effort you guys take to explain things sometimes over and over and over and over is greatly appreciated.
As an example... below comes from a site I was taken to that is the first response to my Google search for "what defines an analog signal vs. a digital signal?" I've read a number of articles prior while trying to demystify myself and remember some early physics classes.
Then... here is the first graphic that appears with a similar search for "visual representation of a pulse width modulated signal"
It would be very easy for someone to think that PWM signals are digital. Now, I suppose, that uneducated people like myself should stay the heck out of it, and let the big boys discuss the matters at hand... but I definitely try to make some efforts to learn before opening my big yap (hitting the keys)....
Thanks again! Seriously.
As an example... below comes from a site I was taken to that is the first response to my Google search for "what defines an analog signal vs. a digital signal?" I've read a number of articles prior while trying to demystify myself and remember some early physics classes.
Then... here is the first graphic that appears with a similar search for "visual representation of a pulse width modulated signal"
It would be very easy for someone to think that PWM signals are digital. Now, I suppose, that uneducated people like myself should stay the heck out of it, and let the big boys discuss the matters at hand... but I definitely try to make some efforts to learn before opening my big yap (hitting the keys)....
Thanks again! Seriously.
Like these 4 in the mega $ ML53 monoblocks, that use linear supplies also ?
Cheers George
@ItsAllInMyHead
I don't know where you got that little table from, but that analog signals are "represented by sine waves" is just fundamentally incorrect.
The key factor here is the discrete part, aka a set amount of chunks with (sometimes) abrupt steps between them.
But even analog signal can have those.
It's more about 1's and 0's, or true and false statements for digital signals.
I don't know where you got that little table from, but that analog signals are "represented by sine waves" is just fundamentally incorrect.
The key factor here is the discrete part, aka a set amount of chunks with (sometimes) abrupt steps between them.
But even analog signal can have those.
It's more about 1's and 0's, or true and false statements for digital signals.
But isn’t DSD (and pulse density modulation), represented by a stream of 1’s and 0’s, yet digital to analog conversion can be done with just low pass filtering?
https://www.diyaudio.com/community/threads/the-best-dac-is-no-dac.273474/page-117
https://www.diyaudio.com/community/threads/the-best-dac-is-no-dac.273474/page-117
Last edited:
You're not alone, as evidenced by this thread! But it's been discussed and put to bed a long time ago on this forum. Perhaps we need a Class-D FAQ.
That's exactly my point... it's wrong. 🙂 I know it's wrong. However, it's literally the first site presented to me when I did the search described previously. It was just (bad) luck that when I was trying to delve a little deeper into the topic that the very first search result was ...
https://byjus.com/physics/difference-between-analog-and-digital/
I know better than to trust most information 'from the internet' (even from this forum). Even among 'experts',
or self purported ones, there is disagreement in terminology. For people learning, it's a challenge to wade through. Personally, I find text books mostly expensive and dry, but perhaps I need to continue just buying old texts, but even those disagree to some extents. I typically try to find sources on the internet. I know... I know...
Lastly, I recognize that it's semantics, but to me... important semantics. Your final sentence is tricky. I struggle a little because again... you're describing what the signal represents rather than the properties of the signal itself. To me... as a learner... that's another one of the fundamental confusions, and why I asked my question earlier. Why do we need to call a signal digital or analog? The signal just has certain properties which can be fairly easily described using electrical terms. Why do we feel the need to call the signal itself digital or analog? An "analog" signal can be used for digital purposes, and a "digital" signal can even be put through a speaker to create a facsimile of an "analog" sound. A synthesizer can create signals associated with music that could conceivably be "digital" until they hit a speaker. Is it a digital signal until it hits an analog device? Then, because the speaker is a speaker, what we hear is now classified as analog b/c the speaker just can't be digital? Yes, some of that is circular logic and theoretical, but ... the fundamental question still sticks in my mind. Why do we need to characterize the signal itself as digital or analog? People seem to be in 100% agreement over the properties of the PWM signal. They argue (or politely discuss) whether it's digital or analog. My question truly is... why does it matter if it's digital or analog? A rose by any other name...
Again... I can't emphasize enough how much I appreciate the knowledge shared on this particular site from people that clearly have established themselves as experts in the field. If I dig enough, I can find a university level education for free on this site. It's separating the information that I find challenging. I also try to understand things I should probably just leave to others... 🙂
- Status
- Not open for further replies.