Hi Jeon,

short answer: Don't worry. Everything will be ok.
long answer:
I do have a bit of a tummy ache about this sentence.
> I wonder that such input range (+- 50 mV) is quite acceptable for GNU
> Radio to manipulate.
There's two things that aren't so great about that sentence:
1. What you're visualizing is not +-50mV. It's the numbers that come out
of your USRP sink, which happen to have a magnitude < ca 0.05. But
that's not really a voltage; it's what the signal processing on the USRP
makes of the the ADC measurement before it reaches your PC and gets
converted to floating point numbers. Now, with your LFRX, which is
basically a voltage follower, and your USRP, which most likely actually
has a 2Vpp measurement range, that might even actually be quite close to
the voltage in Volt, but it's not really necessarily so. The USRP is
*not* a calibrated measurement device, and unless you calibrate
yourself, the only thing you can say is "these samples' values are
proportional to the voltage seen by the ADC" (hint: in your case,
calibration is trivial. Take a small known voltage and apply it between
inner and outer conductor, compare to your visualization, and do it
again, in reverse polarity).
2. The assumption that GNU Radio deals with "mV". It doesn't. It really
just sees sequences of numbers -- so, 0.05 in your case. It doesn't care
(and it doesn't even have any mechanism to know) what these numbers
represent.

So, GNU Radio internally works with floating point numbers, most of the
time (the orange connections are simple 32 bit floats, the blue ones are
complex numbers composed of 32 bit float real and 32 bit float imaginary
part). These numbers are accurate down to about 10^-16 ; you're far away
from that, so unless you do something mathematically extremely unstable,
you should be just fine.

Best regards,
Marcus



On 30.07.2015 07:37, Jeon wrote:
> I am building a communication system which uses light. For the system,
> I've buiult a custom analog circuit and connected it to LFRX with
> SMA-BNC-Alligator clip.
>
> A simple dry run gives the following:
>
>
>
> As you can see, data is transmitted with on off keying.
> (Please ignore some ripples and fluctuation, it's due to 60 Hz
> fluorscent light interference. I'll fix it later.)
>
> I wonder that such input range (+- 50 mV) is quite acceptable for GNU
> Radio to manipulate.
>
> Since it's my first time to physically implement a communication
> system, I only have mathematical and theoretical knowledge, but have
> little experience and sense about dBm, rx sensitivity and so.
>
> But as I can see the waveform not so bad, I think I can manipulate it
> by equalizing or something...
>
> PS: In addition, that 60 Hz interference, would it be better if I
> filter out that at the analog circuit with high pass filter? Or is it
> just ok to use high pass filter block in GNU Radio. I think the former
> is better to reduce the computational cost of GNU Radio. And it is
> more proper to filter out before it passes through the ADC...
>
> Regards,
> Jeon.
>
>
> _______________________________________________
> Discuss-gnuradio mailing list
> Discuss-gnuradio@gnu.org
> https://lists.gnu.org/mailman/listinfo/discuss-gnuradio

_______________________________________________
Discuss-gnuradio mailing list
Discuss-gnuradio@gnu.org
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio

Reply via email to