Send USRP-users mailing list submissions to
[email protected]
To subscribe or unsubscribe via the World Wide Web, visit
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
or, via email, send a message with subject or body 'help' to
[email protected]
You can reach the person managing the list at
[email protected]
When replying, please edit your Subject line so it is more specific
than "Re: Contents of USRP-users digest..."
Today's Topics:
1. Re: Severe DC offset (Matt Ettus)
2. Re: How to build USRP2 binary for BASIC RX/TX from scratch
(Matt Ettus)
3. Re: Severe DC offset ([email protected])
----------------------------------------------------------------------
Message: 1
Date: Mon, 14 Mar 2011 23:33:53 -0700
From: Matt Ettus <[email protected]>
To: [email protected]
Cc: [email protected], [email protected]
Subject: Re: [USRP-users] Severe DC offset
Message-ID: <[email protected]>
Content-Type: text/plain; charset=ISO-8859-1
On 03/13/2011 02:21 PM, [email protected] wrote:
> I'm trying to decode 80211.g packets using the USRP2 and the XUPV5.
>
> When I use the old USRP2 code (not the UHD build, but code obtained off
> of gnuradio head), I don't receive packets. This appears to be caused
> by enormous power in the DC band. When I collect similar traces using
> UHD code, I see no such DC offset and I can decode packets without issue.
>
> I'm wondering what could have happened to cause this discrepancy. The
> DSP code appears to be the same between the two FPGA builds. My current
> belief is that something changed in the software configuration during
> the move to UHD.
The host code, settings, and just about everything else is different.
Are you sure you are getting the right data type? If you interpret a
floating point number as an int, or vice versa, you will see huge DC
offsets.
> Is there some kind of software post-processing going on in the new
> build? Do I need to manually remove the DC offset using the the
> dcoffset registers in the rx dsp block?
>
> I have attached power spectra showing the difference between UHD and the
> old code load.
It is hard to tell since you are using a very non-standard setup. In
any case, the UHD version works for you, so I would suggest using that.
All development is happening on the UHD version, and the old version
has not been touched in nearly a year.
Matt
------------------------------
Message: 2
Date: Mon, 14 Mar 2011 23:36:57 -0700
From: Matt Ettus <[email protected]>
To: Uzair Rauf <[email protected]>
Cc: [email protected], Farooq Shah <[email protected]>,
Farooq Shah <[email protected]>
Subject: Re: [USRP-users] How to build USRP2 binary for BASIC RX/TX
from scratch
Message-ID: <[email protected]>
Content-Type: text/plain; charset=ISO-8859-1
On 03/09/2011 11:25 PM, Uzair Rauf wrote:
>
> Hi
> I am a student who is working on the radar signal processing project
> using custom firmware, retaining as much of the originalcode as we can.
> To wet my feet, I must build the binaries forUSRP2 for BASIC RX/TX from
> scratch. I got the git but amsomewhat lost in the maze. I need to know
> which is the top levelmakefile which builds the entire project OR
> perhaps a bunch of makefilesthat build parts of it. Also the process of
> going from the coed to the files to be burned on the SD. I do have
> xilinx ISE12.4.
The usrp2/top/u2_rev3 directory has the top level makefile you need.
http://www.ettus.com/uhd_docs/manual/html/usrp2.html#load-the-images-onto-the-sd-card
Matt
------------------------------
Message: 3
Date: Tue, 15 Mar 2011 08:20:48 -0400
From: [email protected]
To: "Matt Ettus" <[email protected]>
Cc: [email protected], [email protected]
Subject: Re: [USRP-users] Severe DC offset
Message-ID: <[email protected]>
Content-Type: text/plain; charset=ISO-8859-1; DelSp="Yes";
format="flowed"
Thanks for the response.
We're pulling data directly out of rx_control, which I assume is not
floating point, and routing it to our own serdes core. But I
emphasize that the DC offset issue we are observing happens even if we
collect traces using rx_stremaing_samples and decode them in
simulation. Is rx_streaming_samples generating a floating point
format? Moreover, if we look at the samples as routed to the XUPV5,
and attempt to decode them offline, the power spectra look similar.
With respect to the UHD code, is there a document describing the
various control surfaces. It's a little difficult to trace the code.
I do notice that a lot of the SPI configuration has conceptually moved
into the host side code.
Once question is whether the dc offset compensation capabilities of
the dsp_rx_core are being used by either line of code. It seems that
this could have an impact, but it appears that this capability is not
used. We also wonder whether there might be an issue in the old code
of using USRP Rev3 vs. USRP Rev4. UHD seems to carry around this
information in a much more visible way, although the two board builds
are treated the same.
I guess in UHD, we'll target vita_rx_control. It doesn't seem like a
difficult port.
-Elliott
Quoting "Matt Ettus" <[email protected]>:
> On 03/13/2011 02:21 PM, [email protected] wrote:
>> I'm trying to decode 80211.g packets using the USRP2 and the XUPV5.
>>
>> When I use the old USRP2 code (not the UHD build, but code obtained off
>> of gnuradio head), I don't receive packets. This appears to be caused
>> by enormous power in the DC band. When I collect similar traces using
>> UHD code, I see no such DC offset and I can decode packets without issue.
>>
>> I'm wondering what could have happened to cause this discrepancy. The
>> DSP code appears to be the same between the two FPGA builds. My current
>> belief is that something changed in the software configuration during
>> the move to UHD.
>
> The host code, settings, and just about everything else is different.
> Are you sure you are getting the right data type? If you interpret a
> floating point number as an int, or vice versa, you will see huge DC
> offsets.
>
>> Is there some kind of software post-processing going on in the new
>> build? Do I need to manually remove the DC offset using the the
>> dcoffset registers in the rx dsp block?
>>
>> I have attached power spectra showing the difference between UHD and the
>> old code load.
>
> It is hard to tell since you are using a very non-standard setup. In
> any case, the UHD version works for you, so I would suggest using that.
> All development is happening on the UHD version, and the old version
> has not been touched in nearly a year.
>
> Matt
>
------------------------------
_______________________________________________
USRP-users mailing list
[email protected]
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
End of USRP-users Digest, Vol 7, Issue 26
*****************************************
