Hello, Does the Ettus N320 employ any type of AGC? If so is there a detailed description of it anywhere? Or can it be disabled? I notice that if I send in a strong signal the digital output from the SDR is approx +/-1. If I send in a weak signal it's also +/-1 just noisier. I've been experimenting with sending in amplitude modulated pulses. Seems like over 'long' timescales there is some sort of automatic adjustment going on that is corrupting my measurements. What am I missing here? Conversely it appears I am able to control the TX signal by just sending out smaller numbers, and I don't see this (or am I just missing it)? Thanks in advance for any help or insight you can provide! Thanks, Steve

On 13/11/2023 05:08, Steve Hamn wrote: > Hello, > > Does the Ettus N320 employ any type of AGC? If so is there a detailed > description of it anywhere? Or can it be disabled? I don't believe the N320 implements any type of AGC. > > I notice that if I send in a strong signal the digital output from the > SDR is approx +/-1. If I send in a weak signal it's also +/-1 just > noisier. I've been experimenting with sending in amplitude modulated > pulses. Seems like over 'long' timescales there is some sort of > automatic adjustment going on that is corrupting my measurements. What > am I missing here? A narrowband signal right at the tuning center can get modified by the built-in DC offset compensation in the DSP logic   in the radio.  Look into "offset tuning": https://files.ettus.com/manual/page_general.html And the API bits here: https://files.ettus.com/manual/structuhd_1_1tune__request__t.html The idea is that you ask the radio to offset the LO away from the DC anomaly, and then the DSP in the radio shifts the bandwidth  so that you get exactly the desired center frequency that you asked for. What do you mean by "weak" and "strong" signal.  The maximum recommended input to the radio is -15dBM, and its narrowband   MDS at max gain is probably somewhere south of -120dBm.    So both your "weak" and "strong" signals may just be clipping, or   you're mis-interpreting the data that you are collecting.  The radio really does have a quite high dynamic range, but it   does not implement AGC as far as I know.  That would be something you'd have to implement in your signal processing flow. > > Conversely it appears I am able to control the TX signal by just > sending out smaller numbers, and I don't see this (or am I just > missing it)? > > Thanks in advance for any help or insight you can provide! > > Thanks, > > Steve > > _______________________________________________ > USRP-users mailing list -- usrp-users@lists.ettus.com > To unsubscribe send an email to usrp-users-leave@lists.ettus.com

Thanks for the response. Reading that page about the tune_request_t it says "*The daughterboards that support this functionality are: WBX (all revisions), WBX-120, SBX (all revisions), SBX-120, CBX, CBX-120, UBX, UBX-160*" so will that work for the N320? I actually am trying to avoid the DC signal and have manually tuned 2MHz above the LO (fc=220MHz, BW=2MHz-4MHz) but still see this behavior. I haven't exactly measured the dBm of my signals. I'm transmitting over the air from one N320 and receiving from a second N320, but by strong/weak I mean my "weak" signal is -30dB down from my "strong" signal. I don't see any clipping in my "strong" signal. I guess what should I expect from the digital samples out from the SDR and how does that work? Should I see +/- 1.0 for -15dBm, and then +/-0.001 for a -45dBm signal? (my understanding is that's NOT how it works). Is there some kind of adaptive behavior that autoscales the data, and are there any details available how that works or how to turn it off? It appears to me from my sporadic testing that in short time scales (e.g. microseconds) a XdB change in signal level does result in an XdB change in digital output, but from longer timescales (e.g. 100's of milliseconds or just receive command to receive command) they do not appear correlated. Thank you, Steve On Mon, Nov 13, 2023 at 8:04 AM Marcus D. Leech <patchvonbraun@gmail.com> wrote: > On 13/11/2023 05:08, Steve Hamn wrote: > > Hello, > > > > Does the Ettus N320 employ any type of AGC? If so is there a detailed > > description of it anywhere? Or can it be disabled? > I don't believe the N320 implements any type of AGC. > > > > I notice that if I send in a strong signal the digital output from the > > SDR is approx +/-1. If I send in a weak signal it's also +/-1 just > > noisier. I've been experimenting with sending in amplitude modulated > > pulses. Seems like over 'long' timescales there is some sort of > > automatic adjustment going on that is corrupting my measurements. What > > am I missing here? > A narrowband signal right at the tuning center can get modified by the > built-in DC offset compensation in the DSP logic > in the radio. Look into "offset tuning": > > https://files.ettus.com/manual/page_general.html > > And the API bits here: > > https://files.ettus.com/manual/structuhd_1_1tune__request__t.html > > The idea is that you ask the radio to offset the LO away from the DC > anomaly, and then the DSP in the radio shifts the bandwidth > so that you get exactly the desired center frequency that you asked for. > > > > What do you mean by "weak" and "strong" signal. The maximum recommended > input to the radio is -15dBM, and its narrowband > MDS at max gain is probably somewhere south of -120dBm. So both > your "weak" and "strong" signals may just be clipping, or > you're mis-interpreting the data that you are collecting. The radio > really does have a quite high dynamic range, but it > does not implement AGC as far as I know. That would be something > you'd have to implement in your signal processing flow. > > > > > > Conversely it appears I am able to control the TX signal by just > > sending out smaller numbers, and I don't see this (or am I just > > missing it)? > > > > Thanks in advance for any help or insight you can provide! > > > > Thanks, > > > > Steve > > > > _______________________________________________ > > USRP-users mailing list -- usrp-users@lists.ettus.com > > To unsubscribe send an email to usrp-users-leave@lists.ettus.com > _______________________________________________ > USRP-users mailing list -- usrp-users@lists.ettus.com > To unsubscribe send an email to usrp-users-leave@lists.ettus.com >

 On 13/11/2023 05:08, Steve Hamn wrote:

 detailed

 I don't believe the N320 implements any type of AGC.

 from the

 modulated

 measurements. What

 A narrowband signal right at the tuning center can get modified by
 the
 built-in DC offset compensation in the DSP logic
    in the radio.  Look into "offset tuning":

 https://files.ettus.com/manual/page_general.html

 And the API bits here:

 https://files.ettus.com/manual/structuhd_1_1tune__request__t.html

 The idea is that you ask the radio to offset the LO away from the DC
 anomaly, and then the DSP in the radio shifts the bandwidth
   so that you get exactly the desired center frequency that you
 asked for.



 What do you mean by "weak" and "strong" signal.  The maximum
 recommended
 input to the radio is -15dBM, and its narrowband
    MDS at max gain is probably somewhere south of -120dBm.    So both
 your "weak" and "strong" signals may just be clipping, or
    you're mis-interpreting the data that you are collecting.  The
 radio
 really does have a quite high dynamic range, but it
    does not implement AGC as far as I know.  That would be something
 you'd have to implement in your signal processing flow.

 _______________________________________________
 USRP-users mailing list -- usrp-users@lists.ettus.com
 To unsubscribe send an email to usrp-users-leave@lists.ettus.com

On 13/11/2023 16:48, Steve Hamn wrote: > Thanks for the response. Reading that page about the tune_request_t it > says "/The daughterboards that support this functionality are: WBX > (all revisions), WBX-120, SBX (all revisions), SBX-120, CBX, CBX-120, > UBX, UBX-160/" so will that work for the N320? I actually am trying to > avoid the DC signal and have manually tuned 2MHz above the LO > (fc=220MHz, BW=2MHz-4MHz) but still see this behavior. The N320 has a different parts line-up from any of those boards, which are intended for "motherboard" type radios--including   the USRP N2xx and USRP X3xx. > > I haven't exactly measured the dBm of my signals. I'm transmitting > over the air from one N320 and receiving from a second N320, but by > strong/weak I mean my "weak" signal is -30dB down from my "strong" > signal. I don't see any clipping in my "strong" signal. > > I guess what should I expect from the digital samples out from the SDR > and how does that work? Should I see +/- 1.0 for -15dBm, and > then +/-0.001 for a -45dBm signal? (my understanding is that's NOT how > it works). Is there some kind of adaptive behavior that autoscales the > data, and are there any details available how that works or how to > turn it off? It appears to me from my sporadic testing that in short > time scales (e.g. microseconds) a XdB change in signal level does > result in an XdB change in digital output, but from longer timescales > (e.g. 100's of milliseconds or just receive command to receive > command) they do not appear correlated. > There is no autoscaling of the data, unless you're doing it yourself.    Remember that in a DSP radios the samples are   linear-proxies for antenna *VOLTAGE*,   so a change in voltage of a factor of 100 is a change in *power* of 100**2. The power in a complex-baseband signal is: PWR = AVG(I*I + Q*Q) So with this, if you change the input power by 10dB, the PWR calculated should change by approximately 10dB (there's edge   roll-off in the passband). A -15dBm signal will likely be clipped, depending your RF gain setting. > Thank you, > > Steve > > On Mon, Nov 13, 2023 at 8:04 AM Marcus D. Leech > <patchvonbraun@gmail.com> wrote: > > On 13/11/2023 05:08, Steve Hamn wrote: > > Hello, > > > > Does the Ettus N320 employ any type of AGC? If so is there a > detailed > > description of it anywhere? Or can it be disabled? > I don't believe the N320 implements any type of AGC. > > > > I notice that if I send in a strong signal the digital output > from the > > SDR is approx +/-1. If I send in a weak signal it's also +/-1 just > > noisier. I've been experimenting with sending in amplitude > modulated > > pulses. Seems like over 'long' timescales there is some sort of > > automatic adjustment going on that is corrupting my > measurements. What > > am I missing here? > A narrowband signal right at the tuning center can get modified by > the > built-in DC offset compensation in the DSP logic >    in the radio.  Look into "offset tuning": > > https://files.ettus.com/manual/page_general.html > > And the API bits here: > > https://files.ettus.com/manual/structuhd_1_1tune__request__t.html > > The idea is that you ask the radio to offset the LO away from the DC > anomaly, and then the DSP in the radio shifts the bandwidth >   so that you get exactly the desired center frequency that you > asked for. > > > > What do you mean by "weak" and "strong" signal.  The maximum > recommended > input to the radio is -15dBM, and its narrowband >    MDS at max gain is probably somewhere south of -120dBm.    So both > your "weak" and "strong" signals may just be clipping, or >    you're mis-interpreting the data that you are collecting.  The > radio > really does have a quite high dynamic range, but it >    does not implement AGC as far as I know.  That would be something > you'd have to implement in your signal processing flow. > > > > > > Conversely it appears I am able to control the TX signal by just > > sending out smaller numbers, and I don't see this (or am I just > > missing it)? > > > > Thanks in advance for any help or insight you can provide! > > > > Thanks, > > > > Steve > > > > _______________________________________________ > > USRP-users mailing list -- usrp-users@lists.ettus.com > > To unsubscribe send an email to usrp-users-leave@lists.ettus.com > _______________________________________________ > USRP-users mailing list -- usrp-users@lists.ettus.com > To unsubscribe send an email to usrp-users-leave@lists.ettus.com >

On 13/11/2023 16:48, Steve Hamn wrote: > Thanks for the response. Reading that page about the tune_request_t it > says "/The daughterboards that support this functionality are: WBX > (all revisions), WBX-120, SBX (all revisions), SBX-120, CBX, CBX-120, > UBX, UBX-160/" so will that work for the N320? I actually am trying to > avoid the DC signal and have manually tuned 2MHz above the LO > (fc=220MHz, BW=2MHz-4MHz) but still see this behavior. The N320 uses an architecture that pretty-much *requires* that offset-tuning work, in my previous response, I had mis-read   your message--sorry. Your strategy would be to offset-tune so that the DC anomaly is just outside your passband, as defined by your sample rate. Another thing is that when debugging issues like this, it can be VERY VERY useful to perform some kind of spectral analysis   on the data.  You may well have an interfering signal that is dominating your samples, and without doing an FFT, you'd   never know.