Re: [Discuss-gnuradio] Need some answers about FM TX

[mgray]

The same does apply to a DAC as well as an ADC.  They are both sampling 
systems.  A DAC typically has a low pass filter to cut all the alias 
images.  But you can also bandpass filter those images to produce the 
desired signal.  The system is very linear and isn't a mixer.

Here is my original post:

Another trick to get higher frequencies is to use an image that is
produced the DAC. If you program the DAC to 29.3 MHz, you'll also get RF
at the SampleRate +/- DAC Freq. So you'll get a signal at 128 - 29.3 and
128 + 29.3. So in this case 98.7 and 157.3.  The signal levels will be
lower, but with the proper bandpass filter and amplifier chain, you can
generate a complete signal.  The signal levels fall off at a sin(x)/x 
rate.

I don't recall if there is a low pass filter on the DAC channel. If there
is, you'll need to remove the filter to get the images. With a short piece
of wire stuffed in the SMA connector, you could broadcast across the room
to an FM receiver.


For an example of the band pass filters/RF chain, view:

http://www.kd7lmo.net/picobeacon_hardware.html

This is setup for operation in the UHF band at 445 MHz.



On Sat, 1 Oct 2005, Robert McGwier wrote:

> Jorge:
> 
> The paper you included and the mathematical and electrical phenomenon 
> you are talking about applies to the analog to digital converter, the 
> receiver only.  This is undersampling as you call it.  Most engineers 
> call it bandpass sampling.  It applies to the use of aliasing to bring a 
> higher frequencies down low on receive.  This cannot be done with all 
> analog to digital converters.  It can be done with those analog to 
> digital converters that can "see" the high frequency.  In order for them 
> to see the high frequency, their conversion time must be small compared 
> to the high frequency, and they must respond to the higher frequency. 
> 
> You cannot do the reverse on transmit.  It is a "one to many" problem.  
> The only way you can get a higher frequency is for some (BAD) 
> nonlinearity to be producing the higher frequency content. 
> 
> My apologies, I wish my Spanish was good.  I have included a Google 
> translation and I hope it helps rather than hurts:
> 
> "El papel usted incluido y el fenómeno matemático y eléctrico que usted 
> está hablando se aplica a convertidor a digital análogo, el receptor 
> solamente.  Esto undersampling como usted lo llama. La mayoría de la 
> llamada de los ingenieros él muestreo bandpass.  Se aplica al uso del 
> aliasing de traer frecuencias más altas abajo bajo en recibe.  Esto no 
> se puede hacer con los convertidores a digitales todo análogos.  Puede 
> ser hecha con esos los convertidores a digitales análogos que pueden 
> "ver" el de alta frecuencia.  En la orden para que consideren el de alta 
> frecuencia, su tiempo de la conversión debe ser pequeño comparado al de 
> alta frecuencia, y deben responder a la frecuencia más alta.  Usted no 
> puede hacer el revés encendido transmite.  Es "el que esta' a muchos" 
> problema.  La única manera que usted puede conseguir una frecuencia más 
> alta está para que un cierto nonlinearity (del MALO) produzca el 
> contenido de una frecuencia más alta.
> "
> 
> Bob
> 
> Jorge Chávez wrote:
> 
> > If I run the fm_tx4.py, I TX in 29.3MHz and?? Which other 
> > frecuencies??, Which one is the frecuency of the Sample Rate??
> > Please Help me, and sorry for my english
> >
> >      
> >     On Fri, Sep 30, 2005 at 10:47:47AM -0500, Jorge Chávez wrote:
> >> I'm trying to FM TX with the fm_tx4.py but it doesn work for
> >     comercial FM
> >> Band (88 to 108MHz) but I need to TX in this frecuencys.
> >> Whe I submit my needs in the gnuradio discuss I get the next tip:
> >>
> >> Another trick to get higher frequencies is to use an image that is
> >> produced the DAC. If you program the DAC to 29.3 MHz, you'll also
> >     get RF
> >> at the SampleRate +/- DAC Freq. So you'll get a signal at 128 -
> >     29.3 and
> >> 128 + 29.3. So in this case 98.7 and 157.3. The signal levels will be
> >> lower, but with the proper bandpass filter and amplifier chain,
> >     you can
> >> generate a complete signal.
> >
> >     It really works.    No sample rate change required.
> >     If you have access to a spectrum analyzer, this is easily verified.
> >     Richard Lyons' book "Understanding Digital Signal Processing" has a
> >     very clear explanation of the phenomenon.
> >
> >     There are images produced each 64 MHz (sampling rate / 2).  Every
> >     other one of them has the spectrum inverted, but that won't matter for
> >     FM.  You might try googling for "Nyquist zone" or "bandpass sampling"
> >
> >      http://www.national.com/appinfo/adc/files/Undersampling.pdf
> >
> >     Note that broadcast FM is much wider-band than narrow-band FM.
> >     Very similar, but putting together a wfm tx block would make sense.
> >
> >     If you just connect a wire to the SMA connector and the FM receiver is
> >     close by you can probably hear it.   For a real solution, you'll want
> >     a bandpass filter and amp as described above.
> >
> >      
> >
> >------------------------------------------------------------------------
> >
> >_______________________________________________
> >Discuss-gnuradio mailing list
> >address@hidden
> >http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
> >  
> >
> 
> 
>