Output signal quality

[cochenob]

It doesn't.  This was my original point....using the shipped example...as is...with no changes...results in this behavior.  This is why I initially set the tone frequency to 1Hz rather than cutting the wire and wiring put DC values for I and Q.  I wanted to try and convince myself that it wasn't something silly I was doing with my code.

I can try some of the other shipped examples just to verify this behavior.

[cochenob]

Some more info...

If I increase the carrier frequency above 300MHz or so, then the carrier shape does look sinusoidal.  As you decrease the carrier, the shape gets worse and worse. Is this distortion at lower frequencies expected?  I should have some results from a USRP with a LF front end later today.

The amplitude modulation still remains of course.  Again, this might be the uneven magnitude response of the analog front end, but the envelope doesn't vary according to the FM tone frequency, so I'm still not sure what's going on in this case.

[ErikL]

Today I learned something new. The waveforms you see is expected. When you look at waveforms generated by this kind of direct conversion architecture on an oscilloscope, you see all of the harmonics. The mixer does not put out sine waves, and that is why you get all the harmonics. What you are seeing is expected, and normal.

[cochenob]

Interesting!  Is this due to the particular component choice (i.e. - the oscillator or mixer) or is it more of an architecture thing (that I should expect to see with any radio that does mod/demod in a similar fashion)? 

Also...I'm assuming that the mixers on the Rx end exhibit the same behavior.  Is there any implication there?  i.e. - what does this mean for the demodulated IQ signals?  Anything I need to be cautious of?

For purposes of comparing performance to other SDR's...I'm just trying to wrap my head around whether or not I should take this behavior is a big deal or not.  I'm leaning towards, "no, it's not an issue", but it would help to understand a little more as to why.

Thanks for the help!

---

Brandon

[lkeithr]

Hi Erik, hope it's okay to jump in as we've seen the same thing in our BasicTX 1-250MHz daughterboard transmitter. At some frequencies, the output looks sinusoidal (<30, 100-150); at others, the frequency is correct but who knows what to call the waveform.

What's the basic transmit architecture? Are we dealing with a fully digital signal synthesis, with mixing done in the FPGA before the DAC?

Is the full solution as simple as, add a bandpass filter in the FPGA? (Simple, not easy?) Or at the output, if we can sacrifice the power?

Thanks,

Luke