McIntosh Transformer causing noise in speakers while disconnected.

[ClassG33]

Sorry you're correct this uses toroidal transformers

All g, my illustrious brotha

 

[mhardy6647]

well, I am having fun -- and learning in real time.

The sound definitely has a buzzy quality to me.

 

[mhardy6647]

Why do I get a different one?


View attachment 357727

(that's after the 25 sec where the buzz is audible)

Just the frequency range, I think. Yours goes down to DC mine to 20 Hz.

 

[Doodski]

Using @mhardy6647's image and frequency analysis I am going to call this one at 60 Hz fundamental.

 

[Sokel]

Just the frequency range, I think. Yours goes down to DC mine to 20 Hz.

We have very different levels,mine is some 30db lower.
Is it the whole video you captured or the buzzing only as I did?

 

[mhardy6647]

Using @mhardy6647's image and frequency analysis I am going to call this one at 60 Hz fundamental.
View attachment 357730

Good call. The bold horizontal line is 100 Hz, so, yup, peak at 60 Hz.

 

[mhardy6647]

We have very different levels,mine is some 30db lower.
Is it the whole video you captured or the buzzing only as I did?

Just a bit, then I cropped out (so to speak) the non-noisy part and focused on a few seconds where it is relatively loud. The y-axis is arbitrary to my "record level" and not normalized to anything meaningful. I was simply looking for Gestalt.

 

[Doodski]

Good call. The bold horizontal line is 100 Hz, so, yup, peak at 60 Hz.

That's the thing with not having a proper oscilloscope to view and maybe even a frequency counter the waveform in the noise sample makes it difficult to acquire the waveform duration details.

 

[Golf]

I tried Adobe Audition on a portion between seconds 31 and 34 of the sample. Looks like this:

 

[Doodski]

I tried Adobe Audition on a portion between seconds 31 and 34 of the sample. Looks like this:

View attachment 357738

Do you have a option available for logarithmic scaling? That is better I think.

 

[Golf]

Do you have a option available for logarithmic scaling?

Sure:

 

[Doodski]

Sure:

View attachment 357741

The freq bandwidth analysis output is better although something is preventing the peak value from being indicated as a larger dB value at the fundamental frequency of 60 Hz. To the degree that the curve is distorted and not accurate.

 

[ClassG33]

The freq bandwidth analysis output is better although something is preventing the peak value from being indicated as a larger dB value at the fundamental frequency of 60 Hz. To the degree that the curve is distorted and not accurate.

I'll be back to do a more accurate sample soon... make 20 minutes. stay tuned, this is very interesting....

 

[Golf]

I have to admit that I’m not familiar with this at all ...

Playing around with the »FFT Size«, i. e. enlarging it, leads to results like this:

 

[mhardy6647]

That's the thing with not having a proper oscilloscope to view and maybe even a frequency counter the waveform in the noise sample makes it difficult to acquire the waveform duration details.

Well, I've got a scope (several ) and I do know how to use 'em -- but it was easier to point and click, you know? Even though I don't really know the more subtle ins and outs of Audacity. I am a brute force user of the s/w.

 

[Golf]

I'll be back to do a more accurate sample soon... make 20 minutes. stay tuned, this is very interesting....

Btw ... As long as this mysterious phenomenon isn’t fully understood, better don’t allow any folks with cardiac pacer anywhere near that »beaming« device

 

[ClassG33]

Btw ... As long as this mysterious phenomenon isn’t fully understood, better don’t allow any folks with cardiac pacer anywhere near that »beaming« device

I have had this happen in my Audio Design Career

 

[Doodski]

I have to admit that I’m not familiar with this at all ...

Playing around with the »FFT Size«, i. e. enlarging it, leads to results like this:

View attachment 357744

I think what is required is to get a fresh sample with the present settings and make the sample that is derived from the peak of a waveform. That should result in success.

 

[ClassG33]

Can you zoom in for more resolution of the waveform timing details. What we need is the time interval for a complete 360 degrees of the distorted sinusoidal waveform and that time duration will give us the frequency.... 1/T = F

Here is a better sample video, tome starting about 1:09, and I go through it twice on each driver.
-you'll notice it's almost the same on each Driver, but different amplitude (probably source harmonic matching crossover)
- the amp gives a slight surge a second after powering on, when the Transformer activates, and produces a brief surge of this same tone, but ceases. Is this normal upon powering on? I know the initial "click" is..
- speaker is 100% disconnected and on a foam pad to show you its not picking it up from the cabinet (I can't lift it and do the video simultaneously to prove).
- phone can't really hear the tweeter, but it's there. Don't mistake the high-F background noise with the tweeter; My 1945 home has a lot of background noise, including adjacent laundry room, adjacent mech room with furnace running, new baby boy having story time upstairs, etc.

Noise Sample ~1:09

 

[ClassG33]

I think what is required is to get a fresh sample with the present settings and make the sample that is derived from the peak of a waveform. That should result in success.

Posted