On 27/01/2023 18:49, Mark-Jan Bastian wrote:
On Fri, Jan 27, 2023 at 02:37:08PM -0500, Marcus D. Leech wrote:
On 27/01/2023 00:50, Mark-Jan Bastian wrote:
When generating a complex +74 MHz and -73 MHz sinewave signal and adding them,
I see 1 MHz spur about -30 dB @ -74 MHz, irrespective of LO frequency. Thus the
spur might caused by (remaining) nonlinearity in the mixer or the subsequent
stages.
I would like to control the -30 dB spur to a lower value, or at least find out
what the source of this spur is.
Next thing could be the DAC, which supports rates up to 1600 MSPS, but is only
used at
200 MSPS. Perhaps there is a way to modify the waveform in this step, similar in
the way digital predistortion can help improve spectral output purity?
Mark-Jan
Also, have you run these utilities:
https://files.ettus.com/manual_archive/release_003_010_001_000/html/page_calibration.html
I'm quite sure the problem I see is caused by IQ imbalance. I can simulate it
with the respective
gnuradio block, with near identical results- asymetric spur, not symetric
around a
generated signal, for a whopping 0.5 factor.
I did run the calibration tools on my new system, for rx and tx chains, but
will run them
again (iterate). If that doesnt help I might modify the sourcesignal, and
follow the signal path again.
Mark-Jan
Figure 56 in the AD5375 datasheet would seem to indicate that below 1deg
phase-error, amplitude
imbalance dominates in sideband suppression.
_______________________________________________
USRP-users mailing list -- [email protected]
To unsubscribe send an email to [email protected]
