[time-nuts] DMTD for different input frequencies

[Oz, in DFW]

On 11/14/2010 12:08 PM, Magnus Danielson wrote:
> In the first stage, the input signals is mixed by the average
> frequency (37 MHz in this case) causing the beat frequencies to become
> roughly the same (27 MHz in this case). The second stage would then
> act as as the normal offset local oscillator and beat-frequency mix-down. 
Magnus,

I'm writing you directly because I want to avoid a nitpicking flame war
or I agree with your math, but your nomenclature seems a bit confusing. 
The arithmetic mean or average is 27 Mhz.  It seems that for  the
general case is the local oscillator is the arithmetic mean plus the
lower frequency.  You want to have high side injection for the low
frequency and low side injection for the high frequency.

I'd call the average frequency the "intermediate frequency" and define
it as the arithmetic mean of the two frequencies of interest, because it
literally is and because it conforms to common use for superheterodyne
receivers.  Likewise I'd call the injection the "first local oscillator"
or "first injection" and define it as the intermediate frequency plus or
minus the frequencies of interest for comparison.  The "normal" DMTD
oscillator could then be called the "second local oscillator" or "second
injection"

One side effect of this approach is that common mode variations on the
oscillators being compared will either be enhanced or reduced because
those effects are inverted in frequency for the lower frequency.  I
haven't run the math to prove this, but it looks like it'll be
proportional to the ratio of their frequencies. 

I need to think more about the impact of variations between the first
and second LOs. 

-- 
mailto:oz at ozindfw.net    
Oz
POB 93167 
Southlake, TX 76092 (Near DFW Airport)