Hi, I'm really hoping someone can help me with this. I have four USRP2s and I'm making use of the MATLAB/SIMULINK Communications Blockset. I've been trying to run the transmission demo using each of the USRP2s. Unfortunately I'm consistently picking up intermodulation distortion on all four radios. At the bottom of the e-mail, I've placed links to the spectrum analyzer graphs I've captured from each of the radios. In addition I thought it prudent to provide all other potentially relevant information: - USRP2 version 4 - XCVR2450 Daughterboard - Firmware is as specified on the Mathworks website for the XCVR2450 ( http://www.mathworks.com/support/solutions/en/data/1-CUN7JZ/index.html?product=CB&solution=1-CUN7JZ) - Antennae used are the VERT2450 from Ettus (I've placed antennae on RF1 and RF2 - not sure if this is correct?) - The waveform transmitted should be a 30kHz Real Sinusoid at 2.45GHz - Simulation Gain was set at 10 - Interpolation was set at 512 While I do realize that I could use a bandpass filter to isolate the 30kHz waveform, I believe this would then prevent me from using frequency-based channel access schemes such as OFDMA. I am also aware that I seem to have a frequency shift from my desired central frequency of 2.45GHz - which is something that I am also unsure of how to fix. Lastly, I need to be able to detect when a signal is received by one of the USRP2s. I have read about various synchronization methods making use of auto-correlation with known signals - does anyone know of a simple, reliable method that can be easily implemented? Any thought would be appreciated!! http://i1215.photobucket.com/albums/cc518/RYANVANDENBERGH/USRP2%20Debugging/USRP2_40Subnet_10Gain.jpg http://i1215.photobucket.com/albums/cc518/RYANVANDENBERGH/USRP2%20Debugging/USRP2_30Subnet_10Gain.jpg http://i1215.photobucket.com/albums/cc518/RYANVANDENBERGH/USRP2%20Debugging/USRP2_20Subnet_10Gain.jpe http://i1215.photobucket.com/albums/cc518/RYANVANDENBERGH/USRP2%20Debugging/USRP2_10Subnet_10Gain.jpg Kind regards, Ryan van den Bergh

I haven't done exactely what you do but I have made some related observations. When I am doing OFDM (IEEE802.11a but with only 38 sub-carriers), I can get around 10-12dBm output without impairing the EVM (i.e. the modulation accuracy) to much. A good digital TX base-band signal level is RMS=6000. Conclusion: Test different base-band levels and gain settings and watch the spectrum. BR/ Per On Thu, 2011-01-13 at 18:21 +0200, Ryan van den Bergh wrote: > Hi, > > I'm really hoping someone can help me with this. > > I have four USRP2s and I'm making use of the MATLAB/SIMULINK > Communications Blockset. I've been trying to run the transmission demo > using each of the USRP2s. Unfortunately I'm consistently picking up > intermodulation distortion on all four radios. At the bottom of the > e-mail, I've placed links to the spectrum analyzer graphs I've > captured from each of the radios. > > In addition I thought it prudent to provide all other potentially > relevant information: > - USRP2 version 4 > - XCVR2450 Daughterboard > - Firmware is as specified on the Mathworks website for the XCVR2450 > (http://www.mathworks.com/support/solutions/en/data/1-CUN7JZ/index.html?product=CB&solution=1-CUN7JZ) > - Antennae used are the VERT2450 from Ettus (I've placed antennae on > RF1 and RF2 - not sure if this is correct?) > - The waveform transmitted should be a 30kHz Real Sinusoid at 2.45GHz > - Simulation Gain was set at 10 > - Interpolation was set at 512 > > While I do realize that I could use a bandpass filter to isolate the > 30kHz waveform, I believe this would then prevent me from using > frequency-based channel access schemes such as OFDMA. I am also aware > that I seem to have a frequency shift from my desired central > frequency of 2.45GHz - which is something that I am also unsure of how > to fix. > > Lastly, I need to be able to detect when a signal is received by one > of the USRP2s. I have read about various synchronization methods > making use of auto-correlation with known signals - does anyone know > of a simple, reliable method that can be easily implemented? > > Any thought would be appreciated!! > > http://i1215.photobucket.com/albums/cc518/RYANVANDENBERGH/USRP2% > 20Debugging/USRP2_40Subnet_10Gain.jpg > http://i1215.photobucket.com/albums/cc518/RYANVANDENBERGH/USRP2% > 20Debugging/USRP2_30Subnet_10Gain.jpg > http://i1215.photobucket.com/albums/cc518/RYANVANDENBERGH/USRP2% > 20Debugging/USRP2_20Subnet_10Gain.jpe > http://i1215.photobucket.com/albums/cc518/RYANVANDENBERGH/USRP2% > 20Debugging/USRP2_10Subnet_10Gain.jpg > > Kind regards, > > Ryan van den Bergh > > _______________________________________________ > USRP-users mailing list > USRP-users@lists.ettus.com > http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com

On 01/13/2011 08:21 AM, Ryan van den Bergh wrote: > Hi, > > I'm really hoping someone can help me with this. > > I have four USRP2s and I'm making use of the MATLAB/SIMULINK > Communications Blockset. I've been trying to run the transmission demo > using each of the USRP2s. Unfortunately I'm consistently picking up > intermodulation distortion on all four radios. At the bottom of the > e-mail, I've placed links to the spectrum analyzer graphs I've captured > from each of the radios. > > In addition I thought it prudent to provide all other potentially > relevant information: > - USRP2 version 4 > - XCVR2450 Daughterboard > - Firmware is as specified on the Mathworks website for the XCVR2450 > (http://www.mathworks.com/support/solutions/en/data/1-CUN7JZ/index.html?product=CB&solution=1-CUN7JZ) > > - Antennae used are the VERT2450 from Ettus (I've placed antennae on RF1 > and RF2 - not sure if this is correct?) > - The waveform transmitted should be a 30kHz Real Sinusoid at 2.45GHz > - Simulation Gain was set at 10 > - Interpolation was set at 512 > It is likely that you are overdriving the signal. Use a lower amplitude. Matt

Hi Per, Thanks for the advice. Glad to know someone got it right and I'm not running around in the dark :-). Kind regards, Ryan On 2011/01/13 06:55 PM, Per Zetterberg wrote: > I haven't done exactely what you do but I have made some related > observations. > > When I am doing OFDM (IEEE802.11a but with only 38 sub-carriers), I can > get around 10-12dBm output without impairing the EVM (i.e. the > modulation accuracy) to much. > > A good digital TX base-band signal level is RMS=6000. > > Conclusion: > Test different base-band levels and gain settings and watch the > spectrum. > > BR/ > Per > > > > On Thu, 2011-01-13 at 18:21 +0200, Ryan van den Bergh wrote: >> Hi, >> >> I'm really hoping someone can help me with this. >> >> I have four USRP2s and I'm making use of the MATLAB/SIMULINK >> Communications Blockset. I've been trying to run the transmission demo >> using each of the USRP2s. Unfortunately I'm consistently picking up >> intermodulation distortion on all four radios. At the bottom of the >> e-mail, I've placed links to the spectrum analyzer graphs I've >> captured from each of the radios. >> >> In addition I thought it prudent to provide all other potentially >> relevant information: >> - USRP2 version 4 >> - XCVR2450 Daughterboard >> - Firmware is as specified on the Mathworks website for the XCVR2450 >> (http://www.mathworks.com/support/solutions/en/data/1-CUN7JZ/index.html?product=CB&solution=1-CUN7JZ) >> - Antennae used are the VERT2450 from Ettus (I've placed antennae on >> RF1 and RF2 - not sure if this is correct?) >> - The waveform transmitted should be a 30kHz Real Sinusoid at 2.45GHz >> - Simulation Gain was set at 10 >> - Interpolation was set at 512 >> >> While I do realize that I could use a bandpass filter to isolate the >> 30kHz waveform, I believe this would then prevent me from using >> frequency-based channel access schemes such as OFDMA. I am also aware >> that I seem to have a frequency shift from my desired central >> frequency of 2.45GHz - which is something that I am also unsure of how >> to fix. >> >> Lastly, I need to be able to detect when a signal is received by one >> of the USRP2s. I have read about various synchronization methods >> making use of auto-correlation with known signals - does anyone know >> of a simple, reliable method that can be easily implemented? >> >> Any thought would be appreciated!! >> >> http://i1215.photobucket.com/albums/cc518/RYANVANDENBERGH/USRP2% >> 20Debugging/USRP2_40Subnet_10Gain.jpg >> http://i1215.photobucket.com/albums/cc518/RYANVANDENBERGH/USRP2% >> 20Debugging/USRP2_30Subnet_10Gain.jpg >> http://i1215.photobucket.com/albums/cc518/RYANVANDENBERGH/USRP2% >> 20Debugging/USRP2_20Subnet_10Gain.jpe >> http://i1215.photobucket.com/albums/cc518/RYANVANDENBERGH/USRP2% >> 20Debugging/USRP2_10Subnet_10Gain.jpg >> >> Kind regards, >> >> Ryan van den Bergh >> >> _______________________________________________ >> USRP-users mailing list >> USRP-users@lists.ettus.com >> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com > >

On 2011/01/13 09:13 PM, Matt Ettus wrote: > On 01/13/2011 08:21 AM, Ryan van den Bergh wrote: >> Hi, >> >> I'm really hoping someone can help me with this. >> >> I have four USRP2s and I'm making use of the MATLAB/SIMULINK >> Communications Blockset. I've been trying to run the transmission demo >> using each of the USRP2s. Unfortunately I'm consistently picking up >> intermodulation distortion on all four radios. At the bottom of the >> e-mail, I've placed links to the spectrum analyzer graphs I've captured >> from each of the radios. >> >> In addition I thought it prudent to provide all other potentially >> relevant information: >> - USRP2 version 4 >> - XCVR2450 Daughterboard >> - Firmware is as specified on the Mathworks website for the XCVR2450 >> (http://www.mathworks.com/support/solutions/en/data/1-CUN7JZ/index.html?product=CB&solution=1-CUN7JZ) >> >> >> - Antennae used are the VERT2450 from Ettus (I've placed antennae on RF1 >> and RF2 - not sure if this is correct?) >> - The waveform transmitted should be a 30kHz Real Sinusoid at 2.45GHz >> - Simulation Gain was set at 10 >> - Interpolation was set at 512 >> > > > It is likely that you are overdriving the signal. Use a lower amplitude. > > Matt > > Hi Matt, Thanks for the response and sorry for the double post. I've decreased the amplitude of the Sine Wave (tried 1e-3, 1e-6, 1e-9, and 1e-12) and simultaneously decreased the Gain to 0. I'm still getting the same thing. I've also shifted the Frequency between 2.4GHz and 2.5GHz in fairly small increments of 1MHz and nothing changes. Is it possible that it's a software problem with the Firmware or MATLAB? Kind regards and thanks for the help, Ryan