are you generally ignorant of the assumptions, theory that most everyone else is using
or are you just unwilling to understand, communicate with the rest of us, want to stir things up with "contradictions" that rely on your own built-in assumptions
for one "bandwidth limit" doesn't exclusively mean 1st order analog filters - making your statement sound absurd
16/44 digital audio is pretty ubiquitous - the leading % of music releases, reissues - the audio source most used today
at the dawn of CD digital audio high order analog bandlimiting filters were used
today most digital audio DAC systems include flat to 0.1 dB to 20 kHz digital filters in addition to using up sampling to relax requirements on analog anti-imaging/reconstruction filters, allowing them to be lower order, higher corner frequency to avoid degrading flatness to 20 kHz
or are you just unwilling to understand, communicate with the rest of us, want to stir things up with "contradictions" that rely on your own built-in assumptions
for one "bandwidth limit" doesn't exclusively mean 1st order analog filters - making your statement sound absurd
16/44 digital audio is pretty ubiquitous - the leading % of music releases, reissues - the audio source most used today
at the dawn of CD digital audio high order analog bandlimiting filters were used
today most digital audio DAC systems include flat to 0.1 dB to 20 kHz digital filters in addition to using up sampling to relax requirements on analog anti-imaging/reconstruction filters, allowing them to be lower order, higher corner frequency to avoid degrading flatness to 20 kHz
Last edited:
It's heart warming to see so much agreement on the simple questions of audio. 😛
I liked the picture of the square and sine waves. Easy for simple minds like mine to grasp. Thanks for that one.
I liked the picture of the square and sine waves. Easy for simple minds like mine to grasp. Thanks for that one.
Fourier Series Simulators/Calculators (First link, inch up the harmonic number to build a square wave, and remember that a square wave is made up of entirely odd harmonics)
Does this help anyone? I'm happy to (try to) explain basic signal theory so people can understand what's up.
So as far as a 1 kHz square wave input, band limiting it (assuming a 20 kHz brickwall filter), we'd have the summation of 1, 3, 5, 7, 9, 11, 13, 15, 17, 19 kHz sinusoids. Which, if you try in the link above, gives you a pretty decent square wave! 100 Hz square wave has up to the 199th (remember, odd harmonics) harmonic, so it looks even closer to a "perfect" square wave.
Perhaps I burned down a strawman in talking about 20 kHz square waves, but there seemed to be a confusion about what constituted a "audio signal" and I, perhaps incorrectly, assumed people knew that a square wave is the summation of a (infinite) series of odd harmonics above the fundamental. Mea culpa.
Here's the more literal math: Fourier Series--Square Wave -- from Wolfram MathWorld
Does this help anyone? I'm happy to (try to) explain basic signal theory so people can understand what's up.
So as far as a 1 kHz square wave input, band limiting it (assuming a 20 kHz brickwall filter), we'd have the summation of 1, 3, 5, 7, 9, 11, 13, 15, 17, 19 kHz sinusoids. Which, if you try in the link above, gives you a pretty decent square wave! 100 Hz square wave has up to the 199th (remember, odd harmonics) harmonic, so it looks even closer to a "perfect" square wave.
Perhaps I burned down a strawman in talking about 20 kHz square waves, but there seemed to be a confusion about what constituted a "audio signal" and I, perhaps incorrectly, assumed people knew that a square wave is the summation of a (infinite) series of odd harmonics above the fundamental. Mea culpa.
Here's the more literal math: Fourier Series--Square Wave -- from Wolfram MathWorld
z is zepto (1E-21). NIST has a good summary of the prefixes used in the metric system: Prefixes
I wholeheartedly agree. As is uS in this context. I'm willing to let u pass for µ, but S is siemens, the unit for electrical conductance. Slew rate is measured in volt per second (V/s). Not volt per siemens (V/S).
I know. Feel the love... 🙂
+1.
Tom
True. He appeared to be referring to your post 57, but may have misunderstood what you were claiming. 1kHz square waves have already been dealt with by me: they (when band limited, as they must be) contain lower slew rates than a 20kHz sine wave.Struth said:
Any band limiting means the response is not actually flat anywhere. Fortunately this does not matter, as the 20kHz limit includes the slope towards and beyond 20kHz. Why do people imagine that the 20kHz limit was some sort of brick-wall filter?
If you don't like bandlimiting then you must get all your music from live acoustic instruments only. All sound reproduction is out, as is any music relying on amplifiers for live sound. Sorry, I forgot: your ears act as bandlimiters too!
I go with 1us per peak volt out. So if the rails are +-50 V, at least 50V/us.
For small signal rise time, with band limiting in the 200-500 KHz range, you should see rise times (note this is not the same as SR) of better than 50 ns.
Simple rule of thumb, and it works well. No need to agonize over this.
😎
For small signal rise time, with band limiting in the 200-500 KHz range, you should see rise times (note this is not the same as SR) of better than 50 ns.
Simple rule of thumb, and it works well. No need to agonize over this.
😎
Your not stating what voltage should be reached after those 10us makes me weary of this blanket definition.
Ignoring the use of "S" as a secondary matter, compared to the basic concept.
Why? .................. 😕
Not dissing, just curious about where that number came from.
from SR = 2*Pi*f*Vp
If you look in the black square in the spread sheet you will see a range of SR's that cover 1Vpk/us to 2Vpk/us. The full power (undistorted) bandwidths are between 150 and 300 kHz. This pretty much represents the current range for SOTA power amplifiers. Sure, some will pick lower power bandwidths, and a few have even higher (I have 2 designs that are up at 500 kHz power bandwidth). But 1Vpk/us puts you in the right ball park.
This is a useful heuristic - Bob Cordell also talks about it in his book.
For a useful discussion on SR see James Solomon's tutorial here: http://hifisonix.com/wordpress/wp-content/uploads/2010/10/James-E-Solomon-Opamp-Tutorial.pdf
If you look in the black square in the spread sheet you will see a range of SR's that cover 1Vpk/us to 2Vpk/us. The full power (undistorted) bandwidths are between 150 and 300 kHz. This pretty much represents the current range for SOTA power amplifiers. Sure, some will pick lower power bandwidths, and a few have even higher (I have 2 designs that are up at 500 kHz power bandwidth). But 1Vpk/us puts you in the right ball park.
This is a useful heuristic - Bob Cordell also talks about it in his book.
For a useful discussion on SR see James Solomon's tutorial here: http://hifisonix.com/wordpress/wp-content/uploads/2010/10/James-E-Solomon-Opamp-Tutorial.pdf
Last edited:
"1Vpk/us" mean nothing, same as a few pages ago with the dubbious notion of volts RMS when expressing a slew rate, just write 1V/us, we know that if it s a sine 1V is its peak voltage...
Last edited:
Agonize a little over it. BW and rise time are not independent, BW[GHz]=0.35/RT[ns]
So for 500kHz BW the rise time calculates to 70ns, 175ns for 200kHz. 50ns leads to a 700KHz bandwidth. If one doesn't want the local AM stations to enter the input stage and get demodulated, 200-300KHz is a sweet spot. Then attenuation in the RF range is large enough to avoid the input stage entering nonlinearities due to the RF parasitically.
1V/us SR for each peak output volt is already over an order of magnitude over what is required to avoid SID (slewing induced distortions) or AAMF TIM. This rule of thumb is good enough. Now that you mentioned it, please explain why one would need 500V/us from a CFA topology 100W/8ohm audio amplifier.
The huge confusion between rise time and slew rate comes from missing their sources; while the rise time is natural in linear bandwidth limited amplifiers, as a small signal parameter, slew rate is essentially a nonlinear large signal effect. They have virtually nothing in common, a designer can control both independently. However, in the particular and trivial case of a single pole Miller compensated amp with a long tail pair at the input, the Miller cap controls both, this adds to the illusion the two phenomena are related.
- Status
- Not open for further replies.