[Elecraft] Noisy DSP radios theory

[Brian Machesney]

Keith,

I wonder why you believe that "all DSPs" would "amplify harmonics." When
properly designed, a DSP behaves in a linear fashion - i.e., it does not
produce the harmonic distortion you describe. The criteria for "proper DSP
design" include limiting the bandwidth of the input signal to prevent
aliasing (the Nyquist criterion), limiting the amplitude of the signal to
avoid saturating the A/D converter and scaling the DSP operations to avoid
numerical overflows.

I agree with your assertion that harmonic distortion could produce a "noisy
sounding" receiver, but I can't think of any reason why well-designed
"radios... before DSP" necessarily had any more or less tendency to behave
in a linear fashion; being square-law devices, tube-type and MOSFET-based
amplifiers produce more even-order harmonic distortion which some judge as
less objectionable (purportedly because "mechanical" sound generators, like
musical instruments, produce even-order harmonics) than bipolar-junction
transistors.

-- 
73 -- Brian -- K1LI

On Thu, Jan 7, 2010 at 12:24 PM, Keith Hamilton <tuner at zoominternet.net>wrote:

> I believe it is a fault of all DSP radios that they amplify the HARMONICS
> of a sound too much. In our earlier days before DSP, radio amplifiers tended
> to amplify only the fundamental of a sound and not the partials (harmonics
> 2,3,4,5,6 etc.) ... But if the sound and ALL of its nearby harmonics are all
> amplified equally, or the harmonics are not down in amplitude far enough
> from the fundamental, you might get a busy, or noisy receiver. This could
> equate to amplifying the harmonics too much in a final PA and having a
> 'noisy" or as we call it, a dirty transmitter.
>
>