[Hammarlund] Re: Sweep Method for Alignment - let's get DB correct.

[Kenneth G. Gordon]

On 10 Apr 2006 at 8:38, John Marshall wrote:

> Ken, I think you've been doing fine until the corrections tripped you
> up.  It's a slippery and confusing slope when we start talking about
> "dB power" and dB voltage".

Yup. I was winging it there. :-)

> 
> Decibels ALWAYS refer to a power ratio,

Of course! Thanks for the reminder. 

> specifically dB = 
> 10*LOG(P1/P2), where LOG represents a base 10 logarithm.  3dB is
> approximately a two to one power ratio, 6dB is four to one.
> 
> IF impedance is constant, the relationship can be written as dB =
> 20*LOG(V1/V2) so that a 1.414 to one voltage ratio equals 3 dB and a 2
> to one voltage ratio equals 6 dB.  Oscilloscopes measure voltage and
> it's a lot easier to measure a 2X ratio on a scope than 1.414X, so the
> 6dB bandwidth is kind of a "natural" for swept measurements.
> 
> However, 3dB power does not correspond to "6dB voltage", it
> corresponds to a 1.414:1 voltage ratio which is a 2:1 power ratio.  If
> you measure a 2:1 voltage ratio, it's 6dB (not "6dB voltage") because
> dB is always a power ratio.
> 
> Since bandpass filter rejection increases gradually as you move
> farther to either side of the center frequency, the bandwidth can be
> expressed at any rejection value you like.  For communications
> receiver IF bandwidths, the "6dB bandwidth" seems to be the one most
> often used, and this ties in with the simplicity of measuring a 2:1
> voltage ratio.
> 
> John, KU4AF
> Pittsboro, NC

Thank you, John. I am going to forward this on to both lists as a 
correction.

Ken W7EKB