1
\$\begingroup\$

I wired a Fostex FT17H to a YAMAHA R-S202D receiver and put a 1µF capacitor on the + line (like the manual suggests). If I play now a "normal" audio file (<20kHz) everything is great but when I play an ultrasonic audio file (>20kHz) the tweeter speaker emits a crack/pop sound when starting the audio and when it is finished.

The speaker and receiver are brand new.

Why is this happening?

\$\endgroup\$
5
  • 1
    \$\begingroup\$ "when I play an ultrasonic audio file (>20kHz)" - what exactly is in this 'ultrasonic' audio file, and why are you trying to put it through an audio amplifier? \$\endgroup\$ Commented Jan 16, 2018 at 14:07
  • 1
    \$\begingroup\$ It sounds like your audio file has a DC offset in it. Have you looked at it in an audio editor? \$\endgroup\$
    – Dave Tweed
    Commented Jan 16, 2018 at 14:41
  • \$\begingroup\$ @DaveTweed I generated only a sine wave (without offset) and the crack persists. \$\endgroup\$
    – pythonimus
    Commented Jan 29, 2018 at 13:26
  • 3
    \$\begingroup\$ How did you shape the beginning and end of your tone? If you just start and stop it abruptly, this creates wide sidebands that can be audible. If you ramp the signal level more gradually over a few milliseconds (e.g., linear ramp, raised cosine, etc.), this effect will disappear. \$\endgroup\$
    – Dave Tweed
    Commented Jan 29, 2018 at 13:32
  • \$\begingroup\$ @DaveTweed I m shaping linear out. That helps, but to get the crack sound unnoticable I have to fade out a 22kHz tone over 1 second! This cannot be normal? \$\endgroup\$
    – pythonimus
    Commented Jan 29, 2018 at 15:09

2 Answers 2

1
\$\begingroup\$

If the signal source using a virtual ground and the output cap connecting to the speaker is removing a large DC offset, it takes a few time constants to pass before the DC offset is cancelled (or reformed) on power up/power down.

Also, the manual says you must use a non-polar cap in your connection. A simple electrolytic cap will leak in the half-wave that is alloying reverse bias and creating a DC offset and a pop on power up/power down. You can make a non-polar cap by connecting two electrolytes neg-to-neg and using the two positive sides as your "non-polar" cap at half the capacitance of each single cap.

\$\endgroup\$
-1
\$\begingroup\$

Because at ultrasonic frequencies the capacitor looks more and more like a short circuit - don't do it! Putting ultrasonic frequencies into speakers designed for the audio range can result in destruction of the tweeter

\$\endgroup\$
14
  • 1
    \$\begingroup\$ This has nothing to do with why there might be a "pop" at the beginning/end of the file. \$\endgroup\$
    – Dave Tweed
    Commented Jan 16, 2018 at 14:44
  • \$\begingroup\$ Possibly not, but that will be the least of his problems if he ups the power. Tweeters rated at "80 wats RMS!" seldom take even a fraction of that power without smoking. \$\endgroup\$ Commented Jan 16, 2018 at 16:14
  • \$\begingroup\$ The tweeter speaker is designed to play up to 50kHz, and the manual also states that you have to use a capacitor: madisoundspeakerstore.com/pdf/FT17H_OM_E.pdf Should I connect it directly to the amplifier? Because the manual states to not to. \$\endgroup\$
    – pythonimus
    Commented Jan 17, 2018 at 9:43
  • \$\begingroup\$ @pythonimus No. The capacitor is there to block low fequencies. However, when tweeters are rated for "80W RMS" as I have seen in the past, they do not mean that the tweeter can take 80W of ultrasonic frequencies. Typically it is something like "typical distribution of frequencies in the audio music range" at 80W. Quite often this is approximated by Pink Noise \$\endgroup\$ Commented Jan 17, 2018 at 13:19
  • \$\begingroup\$ @pythonimus Try and find out what the real power rating is for signals above (say) 20kHz \$\endgroup\$ Commented Jan 17, 2018 at 13:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.