[Hackrf-dev] transmitting problem

[Michael Ossmann]

On Tue, Jun 30, 2015 at 08:11:08AM -0700, Peter Shipley wrote:
>
> Was it the harmonics that prevented the hackrf from  transmiting then
> ?

I doubt it.  The harmonics themselves probably do not have any harmful
effects other than making you a worse neighbor on the spectrum.
However, the harmonics are symptoms of abrupt transitions between
frequencies in your generated waveform.  These transitions may cause
problems with the receiver, but it depends on how the receiver is
implemented.  The best way to avoid them is to use a Numerically
Controlled Oscillator as John suggested and to use a pulse shaping
filter.  The typical goal is to produce a waveform continuously varying
phase.

On the other hand, the frequency transitions in the captured waveform
are also fairly abrupt, so I'm guessing that your receiver doesn't care
about phase discontinuities.

> Oddly when eyeballing the data and gnuplot the data looks almost
> identical except for my data being perfectly centered, the higher &
> low frequency are exactly the same but for the phase shift.

There are several things different between your generated waveform and
your captured waveform.  Here are the ones I've noticed (with a guess at
the order of importance):

1. inversion of bits
2. verification that every bit is correct
3. lack of repetition
4. total duration
5. frequency offset (may be correctable with hackrf_transfer -f)
6. abrupt transition between frequencies

Mike