Why class-D amplifiers are crazy efficient
Vložit
- čas přidán 5. 07. 2024
- In this video, we take a look at how class-D audio amplifiers work. These amplifiers achieve efficiencies over 90%, making them the most popular choice for audio power amplification in modern products.
(AKIO TV) MMXXIV
/ actual_akiotv
akiotv.nl - Věda a technologie
Happy new year everyone! (bit late I know but still)
Happy new year to you too!
i think one reason class D have a bad rap is, as you said, the carrier frequency. early D amps used lower frequencies as well as not having fully fleshed out board designs for EMI rejection. this lead to a coloring of the audio that was coming through. this was especially noticeable when bi-amping using tweets/mids. car audio really showed the worst of the worst in that reguard. now days switching ferquencies are MUCH higher, parts are FAR cheaper to make and more easily obtained. loads of board designs have come out that allow intigration of power/output sections with out sacraficing quality. back in the day you had to fully isolate power boards from the audio switching board to really have a chance at ultra low THD numbers. now - its just not that hard to do anymore. things are smaller, solid state, pre-engineered.
one thing that still blows my mind is how much of a delta you get with percieved output going from A to D. a 12w class A tube amp sounds every bit as lout as a 90W class D. thats before we even get into class H or BD or what ever other combination that has come up with over the years.
At these high frequencies are we not really talking about DC-DC conversion. So it is bucket or/boost converter. So here why would there be EMI? The coils prevent any sudden change of current in or out the box. Just make the whole box small so that all waveguide / cavity frequencies are even higher, and eliminate parasitic capacity?
Turning an analog waveform into a pulse train is not without its consequences. It creates a bunch of high frequency energy. While old class D worked much like this video's diagrams (triangle wave modulator into a comparator). Modern day class D shapes the pulse train in a way to push the frequency of the noise as high as possible. This makes it easier for the output filter, which needs to use big and expensive components, to function.
On a cheap design, the output filter is designed to barely pass EMI regulations. They really don't care if the product annoys your hamster. They figure, hamsters do not have enough money to buy an amplifier.
@@moddaudio switching frequencies in the MHz range are possible. Just apply this switched voltage onto a coil, which does not exhibit too much parasitic capacity which makes it a band stop filter.
Funfact: The power grid where HVDC is used runs these same digital amps!
These run at 99,5% near lossless.
A very clear explanation for a neophyte. Thanks a lot for your videos. And happy new year.
man i love your videos great learning experience as a to be electrical engineer
Thanks for the video. I finally understand the difference.
glad I could help :)
Interesting video as usual, love your work bro
Thanks, i didn't know that!
Thanks👍
thx
Often class D amplifiers have fully differential output, eliminating the coupling capacitor.
Also, sigma delta modulation pushes the switching noise further out of band.
That was really clearly explained. I was thinking you would make a good teacher, like teaching the kids technology.
By the way, whether Class D or Class A is better is a hot topic in the hifi community. Most tend to swear by the traditional Class A, but according to what I hear, the difference is all in how you design the Class D. There is a lot more to it in terms of subtle distortions creeping in. You get a kind of delay plus a ringing effect on bipolar transistors, and your A to D converter may be non-linear. The ultimate way is to keep it entirely digital and use GaN FETs in the output. If you want a sharp edge to 16 bit accuracy at 20kHz you are talking about switching at microwave frequencies. GaN FETs can do it and the results are suppose to sound superb, but not cheap. Feedback can iron out harmonic distortion, but increases transient distortion, so for ultimate quality you want totally linearity and zero feedback.
yep, I have the NAD c298 (similar to NAD M23) to upgrade from my already great Parasound 2125v2.
i’ve had other great class d: Parasound zonemaster 2.
my video wow! NAD C298 amplifier, little brother to the NAD M23, with Purifi Eigentakt Class D in a 7.4.4 sys
czcams.com/video/ncwCopnHJqA/video.html
Interesting, in the end it seems to be a tradeoff, more energy for better quality , or less energy for lesser quality. Makes sense thered be one, you can argue its not percievabley noticable, id argue its pointless argueing about someone elses perception, in the end we all see the world a little differently and will never be able to tell what someone else hears or seess, but we can see there is information loss, trading informaton loss for energy gain or vice versa. Id say in the end ino ones will argue against better quality, even the best cameras in the world with no efficiencey in mind cant capture the actual beauty of a moment, 2d images are so far off from the real thing is a good example, as good as they are. And why i think live instrumens will always sound bettr then recorded ones. Just more dimension like a real scence compared to a photo. Maybe one day we will be able to reproduce 3d sound like 3d holograms/pictures.
Thanks for debunking this hoax
Interesting. Class-D PWM amplifiers remind me of early push-pull valve amplifiers that relied on the output transformer to bring the two halves together. The output devices aren't the problem, it's the passive components that reconstruct the original (albeit amplified) signal.
As for "digital", I can't say that I've ever seen an amplifier with a CPU in the signal path, but while you can make bigger digits in the digital domain, you can't make higher voltages or currents by multiplying numbers. Also, pixels aren't digital. It's a mosaic; there are no numbers in it unless you assign them arbitrarily. And since when has film been analog? What signal inside film varies with the light level? Are you claiming that film grain _doesn't_ "chop up" a picture? Methinks you should learn about how these things work before pontificating about them.
Position of grains is not discrete. Position of magnetic domain boundaries on tape is not discrete. Yeah, it is all like a analog PWM. Like the old thyristors for light bulb knobs.
@@ArneChristianRosenfeldt playing a tape gives a voltage that's analogous to the signal used to record onto it. You don't play a photograph though. It's an image, a 2-dimensional image of a 3-D reality, which is one more way it's not analogous. A 1-D recording of a 1-D signal is.
@@StringerNews1 I wonder what signal you get from a tape with a modern HDD read head. I guess that the magnetic film on a HDD is much thinner than the tape with all the magnets embedded.
Photographic film is also three dimensional like magnetic tape. But here you should be able to find single grains in their xyz position using a confocal microscope. So in the time before CMOS sensors ( uh satellites have produced very hires pictures using a CCD for a long time now) you mark all films with frame numbers, cut your film. Then scan the negatives to and apply the digital effects like green screen.
@@ArneChristianRosenfeldt I'm guessing that a HDD head wouldn't work very well, because it's not likely the ideal size, and sacrifices magnetic advantages to be lightweight. It would depend on what kind of tape it is, too. Analog devices like tape heads are very specialized things.
Sure, everything in the real world is three-dimensional, but that fact doesn't give things magical properties. Photographs are still two-dimensional images. I don't know what you're getting at after that. Satellites, CMOS, etc. those are things. But I don't see how they relate to what I was saying.
@@StringerNews1 I thought that the idea with tape was to spread out tiny magnets to make them independent. 3d helps with the spread.
On a similar note the three dimensions of a photographic film allow us to have more grains per “pixel”. We don’t record on mono layer soap films or so. First photograph was on a metal surface. It was sharp, but had low contrast I think.
CCD was for me to explain why movies were still shot on film although we already had powerful computers. So there should have been a period where film was not developed for max contrast, but to see individual grains.
Class D amps are commercialized due to the cheap nature, the components are good for welders and tend to distort the signal you put in. Any 80's class AB amp is light years better than this tech however most people do not care if they got 10% distortion in their music if that is going to cost only a few dollars hence enter the death of HIFI era when mosfet were introduced in amp design.
We have way better audio now than HiFi in the 80s
You can make an absolute decent AB amplifier with welder components, but it would be able to deliver very little power because switching MOSFETs can not handle a lot of power. You can make decent AB amplifiers with IGBTs too you just need to use a level shifter.
But commercially these solutions does not make sense. A D amplifier small, cheap efficient and ticks all boxes required for high quality audio. The audiophile cult wants you to believe that class D is crap.
When blind tested, no audiophile can notice the digital vs oldschool amp.