I don't know how Chen determines input overload - whether he's seeing digital overrun in the data or measuring clipping in the audio input. I suspect you are seeing the inevitable narrow range between hum/noise on the audio output or the high noise floor of the particular sound card and the limited signal handling capability (ca. 4.5V P-P) of a single rail, 5V device.
However, unless a 24 bit soundcard is using +/- 12V supplies for the analog input circuitry, there simply isn't 130 dB of usable range between the noise floor and clipping levels. A properly designed 16 bit sound card will produce in excess of 90 dB dynamic range if the designer takes care with the noise floor issues. That is sufficient to handle signals from the receiver's noise floor to S9+20 dB with- out AGC in most cases if the receiver itself linear. For signals above that level it will take some kind of AGC simply to keep levels within the linear range of the receiver's front end/mixers. 73, ... Joe, W4TV On 8/20/2012 8:58 AM, Bill Frantz wrote: > Using a SignaLink USB interface, I often see cocoaModem > reporting input overload from strong signals on 20 meters with > the noise barely visible on the waterfall. This occurs with both > the K3 and the Small Wonder Labs PSK-20 driving the system. (The > Icom 706 MK2G generates a lot more noise.) I reduce the gain on > the SignaLink, or with the K3 on the RF gain, but that pushes > weak signals too low to decode. > > I am willing to accept that the SignaLink isn't a good sound > card, but need evidence/alternatives. > > Cheers - Bill, AE6JV > > On 8/20/12 at 22:08, li...@subich.com (Joe Subich, W4TV) wrote: > >> With real receivers you probably will never notice the difference in >> dynamic range between a 16 and 24 bit sound card. The receiver AGC >> - even only modest AGC - will keep the dynamic range presented to the >> sound card well within anything the 16 bit sound card can handle. >> That is *unless* the 16 bit card is poorly designed with internal >> noise that wastes a significant part of the 16 bit range. >> >> Even without AGC, the range between "background noise" (the noise >> floor of the demodulation process including sky noise, thermal >> noise in the IF, etc.) and the clipping point of the audio output >> can be considerably less than 90 dB in "real" receivers. Audio >> output levels tend to range from around 10 mV with "no signal" to >> just under 5V P-P (2V RMS) at best ... that's less than 50 dB. >> Even if one assumes the software can decode a 1 mV signal in the >> 10 mV nose floor, the resulting dynamic range is still much less >> than 90 dB provided by a properly designed 16 bit sound card and >> unless the 24 bit soundcard uses other than the typical 5V power >> supply, its real performance will be limited by the same 5V P-P >> audio levels! >> >> Taken a step further - if the receiver produces a 1 mV noise floor >> with 24V P-P output (+/- 12V supplies), that's *still* less than >> 90 dB of range. > ----------------------------------------------------------------------- > Bill Frantz | If the site is supported by | Periwinkle > (408)356-8506 | ads, you are the product. | 16345 > Englewood Ave > www.pwpconsult.com | | Los Gatos, > CA 95032 > > ______________________________________________________________ > Elecraft mailing list > Home: http://mailman.qth.net/mailman/listinfo/elecraft > Help: http://mailman.qth.net/mmfaq.htm > Post: mailto:Elecraft@mailman.qth.net > > This list hosted by: http://www.qsl.net > Please help support this email list: http://www.qsl.net/donate.html > ______________________________________________________________ Elecraft mailing list Home: http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/mmfaq.htm Post: mailto:Elecraft@mailman.qth.net This list hosted by: http://www.qsl.net Please help support this email list: http://www.qsl.net/donate.html