[BC] processing way back when

Robert Orban rorban
Wed Dec 20 17:50:08 CST 2006


At 06:35 PM 12/19/2006, Dave Dunsmoor wrote:
> > .... IIRC, the
> > audio was modulated as RF to permit gain control in the RF domain without
> > thumps.
>
>
>Bob, I'm going to show some ignorance here now, what exactly are you telling
>us here? I don't think I'm following your description, or I don't see what
>you're
>describing.
>
>thanks for a circuit operation lesson,

Thump in a tube-type limiter is leakage of the gain-control signal into the 
audio. In conventional tube-type limiters, the audio is always in push-pull 
form and gain reduction is done by varying the bias voltage of two matched 
variable-mu tubes that also pass the audio signal. When perfectly balanced, 
this push-pull arrangement nulls out the thump, because the thump is a 
common-mode signal.

When you modulate the audio on RF in the form of double sideband suppressed 
carrier, you can now use a single variable-mu tube to change gain. The gain 
control signal will appear as low frequency thumps at the output of the 
tube. However, when the DSB suppressed carrier signal is changed back to 
audio via a product detector, the thumps get frequency-shifted outside the 
audio band (to approximately the carrier frequency) and can be readily 
filtered out.

As an interesting side-effect, this process does not add harmonic 
distortion even though the variable-mu tubes, being nonlinear devices, 
always add nonlinear distortion to the audio. If the carrier is at 1 MHz, 
for example, the second harmonic caused by nonlinearity in the gain-change 
tube will appear at 2 MHz. After product detection, this harmonic will 
appear at 1 MHz and can be readily filtered out. This also happens to be 
the operating principle of a RF clipper, which is another audio processing 
technique that produces no harmonic distortion. (Some of our processors 
have used a "computed RF clipper" using Hilbert transforms to exploit this 
fact. See US Patent 4,495,643.)

Both techniques, however, produce in-band IM distortion. So there is still 
a requirement to minimize the audio signal level on the variable-mu tube in 
order to have the audio on an approximately linear part of the tube's 
transfer characteristic. Moreover, in the RF limiter there is still benefit 
to using two tubes in push-pull because this eliminates even-order IM 
distortion.

Bob Orban





More information about the Broadcast mailing list