You can download a real time spectrum analyzer and signal generator for your PC, baudline if you have a linux machine. Generate a 1k sine at half VP-P out (use resistor, lower output voltage is alright) and listen to it with the sound card and spectrum analyzer. With a filter like that you will probably see bumps above and below 1k even with a cheap old card. I'd be surprised if you didn't see one or two well under 80dB down. If the filter is not in a feedback loop you are free to play with values and circuits until you like what you see on the analyzer and hear, without any possible stability issues.
Last edited:
The waveforms that I posted are from my own amplifier. I will post 1khz capture when I find some time to take it.
Carrier residual is part of the class D concept, it does not introduce any distortion. A few volts peak to peak of clean carrier residual at the output are found in many high power class D amplifiers.
In fact, the PWM signal could be just fed to the loudspeaker and the result in THD would be the same, only audio frequencies matter and carrier residual components are well above 20khz (the first component happens at switching frequency, the second at 2f, 3f, ..., around 500khz, 1Mhz, 2Mhz... for this tripath chip).
Carrier residual is part of the class D concept, it does not introduce any distortion. A few volts peak to peak of clean carrier residual at the output are found in many high power class D amplifiers.
In fact, the PWM signal could be just fed to the loudspeaker and the result in THD would be the same, only audio frequencies matter and carrier residual components are well above 20khz (the first component happens at switching frequency, the second at 2f, 3f, ..., around 500khz, 1Mhz, 2Mhz... for this tripath chip).
I think you will notice lots of distortion right at the amp terminals even with a resistance load. Once you add wire, crossover components, and speaker motors, the chance for harmonics and intermodulation by driving the whole system with a large portion of even rounded off carrier goes way up .
But I forgot, the carrier residual probably Will freak out the input of a cheap sound card, making the scene look even worse than it really is when sampled at a lower frequency. You could try picking off a low level 1k test tone at the speaker and then after an attenuator and high slope filter well into the audio band look for components below 2 or 3Khz. You wouldn't need an expensive analyser for that, but you might have a to cob up a little circuitry. Anything other than the test tone below the second harmonic and between integer harmonics is the stuff that you Really don't want to hear.
But I forgot, the carrier residual probably Will freak out the input of a cheap sound card, making the scene look even worse than it really is when sampled at a lower frequency. You could try picking off a low level 1k test tone at the speaker and then after an attenuator and high slope filter well into the audio band look for components below 2 or 3Khz. You wouldn't need an expensive analyser for that, but you might have a to cob up a little circuitry. Anything other than the test tone below the second harmonic and between integer harmonics is the stuff that you Really don't want to hear.
Last edited:
I will try to take a picture of the phenomen but I'm not sure to be successful. I can try to make a little video sequence.
I've just taken a video file and I'm waiting ten minutes, as soon as the uploading will be done on You Tube I will show the link.
I do not understand anything, now the amplifier doesn't go to protection !!!! and I tried on L & R channels !!!
I've tested with 3 waveforms, tell me what you're thinking about that, thanks in advance
Ok, I'm back, this is the link : http://www.youtube.com/watch?v=3xjF4UVDuys
Last edited:
LOL! It'd probably sound better if you stuck a nail in the center.
Last edited:
There was an error in uploading the file, I've reuploaded it, this is the new link :
YouTube - AR output signals forms
YouTube - AR output signals forms
Everyone tells me it's normal, but it's not my mind. The power resistors (2 black 390 ohms 5 w) of the RC damper used after the LC filter are getting very very hot when powers goes to 10 watts and more (with an input sinus of 30 KHz). A full power, in five seconds, they begin to smoke and I must lower the input level to avoid them burning. They do not burn at 1000 Hz full power. Why did'nt they put a low pass filter in the input to protect the output filter ?
I tried the amplifier with a sacrified loudspeaker and it burned when the power was about 10watts.
I definitively hate this amplifier.
I tried the amplifier with a sacrified loudspeaker and it burned when the power was about 10watts.
I definitively hate this amplifier.
Last edited:
Feding class D amplifiers with high amplitude stuff above 5Khz is not a good idea. Only professional amplifiers with limiter/protection circuits and UcD style circuits can handle that. Everything else will smoke output filter resistors (UcD style circuits don't need these resistors).
This is not exclusive of class D, many class AB amplifiers will fail due to high stress (smoked output zobels), overheating of components or just cross-conduction in the output stage if they are fed with high amplitude stuff above 5khz (no matter whether a load is connected or not).
The output filter capacitors also have a limit on the amount of AC voltage they can handle above approx. 5khz. They can melt too, film caps exhibit enough losses for that to happen, they are not as good as many people think. The ones with wider terminal spacing can handle less HF AC, so they are worse.
For example, I had to design a limiter for my amplifiers matching the maximum voltage vs frequency plot shown on output capacitors datasheet. It allows full output for 1 second before acting, so it prevents damage without disturbing audio.
Anyway, the pictures tell the whole story. This amplifier comes from class AB designers making their very first steps into class D and getting a lot of undeserved money in the process.
They think that throwing exotic components on a poorly laid out PCB is enough to get a good amplifier. Everything is done wrong.
It's a pure scam.
Nearly all small high-end companies are like that.
For example, a couple of inexpensive UcD modules will outperform this amplifier in every measurable field.
This is not exclusive of class D, many class AB amplifiers will fail due to high stress (smoked output zobels), overheating of components or just cross-conduction in the output stage if they are fed with high amplitude stuff above 5khz (no matter whether a load is connected or not).
The output filter capacitors also have a limit on the amount of AC voltage they can handle above approx. 5khz. They can melt too, film caps exhibit enough losses for that to happen, they are not as good as many people think. The ones with wider terminal spacing can handle less HF AC, so they are worse.
For example, I had to design a limiter for my amplifiers matching the maximum voltage vs frequency plot shown on output capacitors datasheet. It allows full output for 1 second before acting, so it prevents damage without disturbing audio.
Anyway, the pictures tell the whole story. This amplifier comes from class AB designers making their very first steps into class D and getting a lot of undeserved money in the process.
They think that throwing exotic components on a poorly laid out PCB is enough to get a good amplifier. Everything is done wrong.
It's a pure scam.
Nearly all small high-end companies are like that.
For example, a couple of inexpensive UcD modules will outperform this amplifier in every measurable field.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.