With respect to the discussion about clocked designs vs self oscillating designs, I put together some fundamental thoughts and examinations.
The notes include multiple graphics, so I put all in one pdf.
🙄🙄
Well, after being rejected by the diyaudio upload system I had to split it in multiple documents... sorry for that.
🙄🙄
The notes are intended as a new starting point for discussion.
Comments, experiences, comparisons and examinations are highly welcome.
The notes include multiple graphics, so I put all in one pdf.
🙄🙄
Well, after being rejected by the diyaudio upload system I had to split it in multiple documents... sorry for that.
🙄🙄
The notes are intended as a new starting point for discussion.
Comments, experiences, comparisons and examinations are highly welcome.
Hi Choco
Very interesting.
I think our ears and brain do not sum like the two resistors. That is why you dont hear it.
I think that beating must be shown in both signals, i.e. like a 2 kHz signal in either of the outputs, for the ear to hear.
I therefore think that the beating signal must be injected into the other chanel to be heard, either through the PSU, air or other.
Could you post the LTSpice file for further experiments?
Best regards Baldin 🙂
Very interesting.
I think our ears and brain do not sum like the two resistors. That is why you dont hear it.
I think that beating must be shown in both signals, i.e. like a 2 kHz signal in either of the outputs, for the ear to hear.
I therefore think that the beating signal must be injected into the other chanel to be heard, either through the PSU, air or other.
Could you post the LTSpice file for further experiments?
Best regards Baldin 🙂
Hi Markus
Interesting. The beep cannot be anything else than effect of AM demodulation which took place somewhere (I'd suspect generators if I were to guess).
A sum of two sines of different frequency should simply add up according to superposition prinicple, and so do the spectra of them creating something identical to AM signal of a given carrier and 1kHz modulation. It's just two supersonic fourier components .
Remember it only takes one diode to demodulate AM signal and 2kHz is in the best hearable frequency range.
The only intriguing issue is your "uncertain perception that there might be something."
That's what you get close to upper hearing range. I did myself a generator testing of what I really can hear and around 19kHz I was sure I was hearing the tone. But at 20.5kHz I got something like what you've described, that was not a hearable beep of any sort, but a dizzy feeling. I ask my friend to switch on and off the generator and although I didn't experience pure kind of hearing, I seemed to have known perfectly if a generator is on or off.
Possibly also (little speculating here) the so-called 'binaural beating' is doing some tricks. It is a very puzzling brainf\/cking experience well in audible range, because it triggers the same parts of brain, which are responsible for sensing direction of the incoming noise (phase comparison, sensing the Doppler effect).
I wouldn't be very much shocked if it worked a bit over the real hearing threshold.
Regards,
Adam
Interesting. The beep cannot be anything else than effect of AM demodulation which took place somewhere (I'd suspect generators if I were to guess).
A sum of two sines of different frequency should simply add up according to superposition prinicple, and so do the spectra of them creating something identical to AM signal of a given carrier and 1kHz modulation. It's just two supersonic fourier components .
Remember it only takes one diode to demodulate AM signal and 2kHz is in the best hearable frequency range.
The only intriguing issue is your "uncertain perception that there might be something."
That's what you get close to upper hearing range. I did myself a generator testing of what I really can hear and around 19kHz I was sure I was hearing the tone. But at 20.5kHz I got something like what you've described, that was not a hearable beep of any sort, but a dizzy feeling. I ask my friend to switch on and off the generator and although I didn't experience pure kind of hearing, I seemed to have known perfectly if a generator is on or off.
Possibly also (little speculating here) the so-called 'binaural beating' is doing some tricks. It is a very puzzling brainf\/cking experience well in audible range, because it triggers the same parts of brain, which are responsible for sensing direction of the incoming noise (phase comparison, sensing the Doppler effect).
I wouldn't be very much shocked if it worked a bit over the real hearing threshold.
Regards,
Adam
Last edited:
Hi,
Interesting topic. can help us reflect on some things.
I would like to add something to the last questions at the end of the pdf
In the case of two ultrasonic frequencies (2KHz with a difference), sent to two speakers.
We can close one ear and hear perfectly the beep. This shows that:
1) The difference is not produced (as a result) of our brain.
2) Our two ears are actually two independent channels, and our brain does not use for differences in frequencies but in amplitude and delay.
Regards
Interesting topic. can help us reflect on some things.
I would like to add something to the last questions at the end of the pdf
In the case of two ultrasonic frequencies (2KHz with a difference), sent to two speakers.
We can close one ear and hear perfectly the beep. This shows that:
1) The difference is not produced (as a result) of our brain.
2) Our two ears are actually two independent channels, and our brain does not use for differences in frequencies but in amplitude and delay.
Regards
Last edited:
Hi Adam,
yes I also think, that it must be sort of a demodulation effect.
Because the beating signal itself does not look like being audibile.
Besides your fourier description, we could also describe it as one ultrasonic signal which grow and disappear, just the envelope would be in the audible frequency range (==> that' what we see on the scope.)
Hi Baldin,
I am attaching the LT Spice file.
It should run with the standard libraries.
The buffered transmission line is not mandatory to get the beating, but reflecting real world a little better.
yes I also think, that it must be sort of a demodulation effect.
Because the beating signal itself does not look like being audibile.
Besides your fourier description, we could also describe it as one ultrasonic signal which grow and disappear, just the envelope would be in the audible frequency range (==> that' what we see on the scope.)
Hi Baldin,
I am attaching the LT Spice file.
It should run with the standard libraries.
The buffered transmission line is not mandatory to get the beating, but reflecting real world a little better.
One more examination:
Examination 5:
Two signal generators, two ultrasonic frequencies, difference around 2kHz.
Both signal generators have the ground connected (like all previous examinations). Now the tweeter is connected between both hot ends, like in a bridged amp.
==> Strong 'beeeeeepp' again...
This might be an indicator that bridged configurations of selfresonant topologies are more critical than unbridged.
Examination 5:
Two signal generators, two ultrasonic frequencies, difference around 2kHz.
Both signal generators have the ground connected (like all previous examinations). Now the tweeter is connected between both hot ends, like in a bridged amp.
==> Strong 'beeeeeepp' again...
This might be an indicator that bridged configurations of selfresonant topologies are more critical than unbridged.
One more examination:
Examination 5:
Two signal generators, two ultrasonic frequencies, difference around 2kHz.
Both signal generators have the ground connected (like all previous examinations). Now the tweeter is connected between both hot ends, like in a bridged amp.
==> Strong 'beeeeeepp' again...
This might be an indicator that bridged configurations of selfresonant topologies are more critical than unbridged.
Examination 5:
Two signal generators, two ultrasonic frequencies, difference around 2kHz.
Both signal generators have the ground connected (like all previous examinations). Now the tweeter is connected between both hot ends, like in a bridged amp.
==> Strong 'beeeeeepp' again...
This might be an indicator that bridged configurations of selfresonant topologies are more critical than unbridged.
Yes, definitely. Generator is much more probable than speaker
Yes. It is done on frequency domain. And strongly band-limited before transformation.
No, but air is much more linear than the output of the generator.
You mean I was around 3 mHz ??
But obviously I am suffering from a heavy 'posting arrhythmia'. 😛
In order to check if the generators act as demodulators, I repeated examination 3 and examination 5 by using a class AB audio amplifier for buffering the generators.
==> Silence.
All together this appears to me that such beating becomes an issue if you have nonlinear elements in the power stage area (I am able to get back the beep even by doing a purposedly poor contact at the tweeter's connector 😀 ).
Under the condition of high quality power stages, low distortion tweeters and good contacts it seems to be not much of headache.
==> Silence.
All together this appears to me that such beating becomes an issue if you have nonlinear elements in the power stage area (I am able to get back the beep even by doing a purposedly poor contact at the tweeter's connector 😀 ).
Under the condition of high quality power stages, low distortion tweeters and good contacts it seems to be not much of headache.
I used the simulation file from my posting #6 for a FFT.
Changed the input voltage down to 1pV and extended the simulation time up to 11ms. Time frame for FFT was 1ms to 11ms.
Also the FFT is of the opinion that there is no audible content.
In the zoom we can see both oscillation frequencies, but checking over the entire range down to 100Hz there is no content of f1-f2.
f1-f2 is just the frequency of a resulting amplitude modulation.
It will become audible as soon as we introduce unpleasant nonlinear elements that act as simple AM demodulator.
Changed the input voltage down to 1pV and extended the simulation time up to 11ms. Time frame for FFT was 1ms to 11ms.
Also the FFT is of the opinion that there is no audible content.
In the zoom we can see both oscillation frequencies, but checking over the entire range down to 100Hz there is no content of f1-f2.
f1-f2 is just the frequency of a resulting amplitude modulation.
It will become audible as soon as we introduce unpleasant nonlinear elements that act as simple AM demodulator.
Attachments
- Status
- Not open for further replies.