50kHz pitch shifter

[Scott Gravenhorst]

<smoke blowing>
Hmm.  I guess alot depends on the band on which the noise exists
and how the "intelligence" bacame modulated on it.

If the intelligence is simply mixed with the 50kHz sine, the 
carrier can be stripped off with a notch filter.

The sine wave, if constant, would be uninteresting, but the noise
and other perturbations present might be.  If the noise and other
signals are already in the audio range and are amplitude modulated 
on the 50 kHz sine, a high-Q notch filter tuned to 50 kHz might work.
There's more trouble, I think, if the carrier is not very stable.  
A wider notch might work.  Again, much depends on the band where 
the desired signal sits.  

If the desired signals are frequency modulating the sine, the
"intelligence" can be recovered perhaps using a phase locked loop
locking to the 50kHz sine, then listening to the VCO's control
voltage.

Once recovered, If the noise is in a band above human hearing, then 
indeed, you will need a shifter of one kind or another.  A pitch 
shifter will preserve harmonic relationships, a frequency shifter 
will not, but rather the hertz differences.  (that may not be a 
precise definition and I'm sure that it will be tweaked by others 
if it's not).  It may be interesting to "hear" this using both
approaches.  It may be possible to use digital methods to stretch
the signal out in time, like sampling it at one rate and playing back
at another.  That's a whole other can of worms and again, alot 
depends on the frequency band of the intelligence signal.

I think the ring mod approach might also work, then followed by 
a notch filter to strip the sine wave.  But this method provides
no demodulation that I'm aware of, just shifting the whole mess.

Anyway, I'd say what you're attempting sounds very interesting.
</smoke blowing>

Toby Paddock <tpaddock at seanet.com> wrote:
>Hello,
>A friend is doing a physics experiment for school and 
>one of the resulting signals is a 50kHz sine wave with 
>noise on it. The noise and ringing is one indication 
>of what is going on. We thought it might be worth 
>bonus points to be able to listen to it. 
>So we need a pitch shifter or divider.
>
>Maybe a 1/100 divider. Sample one cycle and play it 
>back slow and do it again. Sync could be direct from 
>the clean sine source. The wave is periodic and fairly 
>repetitive, cycle to cycle. But the sample rate would 
>be pretty high.  
>
>Maybe a ring mod (my favorite). Assuming the carrier 
>is a clean sine wave would the output have the same 
>waveform as the input. It would not need to be perfect 
>as long as you could hear the change. Any tips on 
>what ring mods like 50kHz inputs? 
>I have a sample AD633, but have never built a ring mod.
>
>Ideas?
>
>Thanks,
> - -- -  Toby Paddock
>
>

-- Scott Gravenhorst
-- Linux Rex, Linux Vobiscum  |  RedWebMail by RedStarWare
-- FatMan: www.teklab.com/~chordman
-- NonFatMan: members.xoom.com/_XMCM/chordman/index.html
-- The 21st century does NOT start in the year 2000!!!