On 17/04/2023 22:49, Shenk, Trey E wrote:
Hi all,
Thanks for the input on this issue. I’m still trying to pin it down.
I’m using gnuradio to drive a B205mini. I’m seeing harmonics of the
carrier particularly at odd harmonics. The third harmonic, 3*carrier,
is only 10dB down from the carrier itself, and I’m seeing significant
carrier leakage.
I’ve tried changing the sampling rate, as well as modulating a QPSK
signal (both direct conversion and using some off-tuning). I still see
spectral images at the odd harmonics.
Is this known behavior for the B205mini? I’m trying to decide if I
need to change how I’m driving it, use a RF filter at the output, or
switch to a different SDR.
Thanks,
Trey
Two things.
The LO leakage can be dealt with using an LO offset.
However, the AD9361 uses a square-wave LO into the mixer, which means
that odd harmonics will make it out of the device.
This has been discussed extensively in the ADI forums, and a couple
of folks on this list have discussed it as well with
spectrum analyzer screen dumps showing it. I confirmed it with my
own modest spectrum analyzer as well. Because it's
an odd-harmonic problem, any application operating above 2GHz will
naturally not have this problem, because the 3rd
harmonic (and all other odd harmonics) will fall outside the
operating range of the AD9361 chip.
The only solution is filtering.
Some USRPs have a built-in switchable filter bank that can help with
this in some cases, but it really depends on the
frequency scheme of your application. A discrete filter bank cannot
possibly deal with all applications.
I personally tend to think of SDRs as *components* in an overall
engineered RF *system*. It is certainly the case that
ready-for-type-acceptance radios, regardless of internal
architecture, will almost always have filters at the edges to
deal with architectural unpleasantness deeper in the radio
architecture. The B2xx series were designed to be
relatively inexpensive, and as such, they don't have a switchable
filter-bank on them--and as previously pointed out,
even THEN, that may not necessarily make the radio "application
ready" in terms of formal type-acceptance, etc.
*From:* Shenk, Trey E
*Sent:* Wednesday, April 12, 2023 4:21 PM
*To:* Marcus D. Leech <patchvonbr...@gmail.com>
*Cc:* USRP-users@lists.ettus.com
*Subject:* RE: [USRP-users] Harmonic Distortion with B205mini
The original signal that I showed was an unmodulated carrier.
I tried a 10kHz complex exponential (plots included for several
harmonics). The carrier is clearly visible at all frequencies. I can
see copies of the 10kHz tone on odd multiples of the carrier, but not
on the even multiples.
*From:* Marcus D. Leech <patchvonbr...@gmail.com>
*Sent:* Monday, April 10, 2023 10:22 PM
*To:* Shenk, Trey E <trey.sh...@pnnl.gov>
*Cc:* USRP-users@lists.ettus.com
*Subject:* Re: [USRP-users] Harmonic Distortion with B205mini
On 10/04/2023 14:28, Shenk, Trey E wrote:
I first tried turning the gain down. What I saw was that the even
harmonics (2*fundamental, …) did not change power, and the odd
harmonics decreased by the same amount as the fundamental. This
means that the dBc for the odd harmonics stayed the same with
decreasing gain.
Decreasing the baseband amplitude had the exact same effect. Even
harmonics stayed at the same power level, odd harmonics decreased
while maintaining dBc.
What is the nature of the modulating signal?
If you use an example app like "tx_waveforms" with, let's say, 10kHz
SIN signal, what are the harmonic results?
*From:* Marcus D Leech <patchvonbr...@gmail.com>
<mailto:patchvonbr...@gmail.com>
*Sent:* Monday, April 10, 2023 8:39 AM
*To:* Shenk, Trey E <trey.sh...@pnnl.gov> <mailto:trey.sh...@pnnl.gov>
*Cc:* USRP-users@lists.ettus.com
*Subject:* Re: [USRP-users] Harmonic Distortion with B205mini
Check twice before you click! This email originated from outside PNNL.
Turn down the RF gain a bit as well as the baseband amplitude.
Does this make any difference?
Sent from my iPhone
On Apr 10, 2023, at 9:41 AM, Shenk, Trey E via USRP-users
<usrp-users@lists.ettus.com> wrote:
I'm using a B205mini to transmit signals. When transmitting, I
can see copies of the SOI at harmonics of the center
frequency. I ran some measurements of total harmonic
distortion, and found it to range from 39% with a 100MHz to
23% with a 1GHz carrier. The second harmonic is <-50dBc, but
the third harmonic is usually around -10dBc.
My main concern is for the lower frequency carriers, like
100MHz, because multiple harmonics will show up on a spectrum
analyzer set to a wideband. I've looked at putting an RF
filter at the output, but I need the system to be able to
switch transmit center frequencies in a range from 100MHz to 5GHz.
Is it possible to reduce the harmonics by some hardware
setting (driving with gnruadio)?
Thanks,
Trey
<carrier_freqsweep_fc0200M_gain55.png>
_______________________________________________
USRP-users mailing list -- usrp-users@lists.ettus.com
To unsubscribe send an email to usrp-users-le...@lists.ettus.com
_______________________________________________
USRP-users mailing list -- usrp-users@lists.ettus.com
To unsubscribe send an email to usrp-users-le...@lists.ettus.com
