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: Custom RFNoC block connection through UHD API and
accessing front panel GPIO (Jonathon Pendlum)
2. Re: Is there a MULTIPLY block in RFNoC? (Jonathon Pendlum)
3. Re: [EXTERNAL] Re: low frequency ADC flicker noise (Ettus
x310) (Hofer, Russell Dwayne)
4. Re: Is there a MULTIPLY block in RFNoC? ([email protected])
5. Re: Is there a way to store modulated samples generated by
USRP (Qurat-Ul-Ann Akbar)
6. Retrieving samples up-sampled to 2.4GHz from the USRPN210
(Qurat-Ul-Ann Akbar)
7. Re: Retrieving samples up-sampled to 2.4GHz from the USRPN210
(Kyeong Su Shin)
8. Difference between Clock Recovery and Audio rate
synchronization. (Benny Alexandar)
----------------------------------------------------------------------
Message: 1
Date: Fri, 9 Jun 2017 13:22:03 -0400
From: Jonathon Pendlum <[email protected]>
To: "Jean Michel Cioranesco (jm)" <[email protected]>
Cc: "[email protected]" <[email protected]>
Subject: Re: [USRP-users] Custom RFNoC block connection through UHD
API and accessing front panel GPIO
Message-ID:
<cagdo0utnlsmyayhwvyo+v1zw__8dr_bsqjtp83ctk8+qdex...@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"
Hi,
Now I am encountering some difficulties in the next steps and I hope you
> guys can help me.
>
> I am unsure how I could connect the usrp block on the receiver side. I was
> thinking to use the following script (but I could not successfully verify
> it is connected on the datapath):
>
> ./rfnoc_rx_to_file --radio-id 0 --block-id customblock
>
> Is this how you guys are connecting your custom blocks? Do you have a
> tutorial or other information that would help me connect the block
> successfully?
>
> I would like to use the UHD API directly to make the connection in RFNoC
> if possible.
>
That will work and is a good example to follow on how to build your own
RFNoC UHD apps. We do not have a pure UHD API tutorial on how to connect
blocks yet, although the example code is very straightforward.
> My second question relates to the RTL design of the custom block. I would
> like to connect an output of my custom core directly to one of the
> frontpanel GPIO pin.
>
> Is there any example out there on how to do that or other useful
> information?
>
The GPIO is connected to the radio core, so you have two options: 1) Modify
the RTL to connect one of the GPIO pins to your block instead of the radio
core or 2) Send a command packet to the radio core to set the GPIO.
> My third and final question is about the UHD version compability of the
> API, I am now working under UHD_4.0.0.rfnoc-0-unknown to develop the custom
> FPGA image. Will the software I will develop to connect the RFNoC blocks be
> portable to UHD version 3.10.1.1?
>
While behind the scenes UHD is using RFNoC in 3.10+, we will not enable
full RFNoC support in UHD until the 4.0 release.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20170609/f5c42c65/attachment-0001.html>
------------------------------
Message: 2
Date: Fri, 9 Jun 2017 13:24:19 -0400
From: Jonathon Pendlum <[email protected]>
To: "Swanson, Craig" <[email protected]>
Cc: "[email protected]" <[email protected]>
Subject: Re: [USRP-users] Is there a MULTIPLY block in RFNoC?
Message-ID:
<cagdo0urvfr8wd0onxvyij7mt24zl+fq-d7w1kptsesw-ko_...@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"
Hi Craig,
Depends on what you want to do. If you want to multiply by a constant, we
have the digital gain RFNoC block. If you want to multiply two streams,
you'll need to take mult.v, mult_rc.v, or complex_multiplier.v and wrap
that in a RFNoC block.
On Fri, Jun 9, 2017 at 9:50 AM, Swanson, Craig <
[email protected]> wrote:
> Jonathon,
>
> In gnuradio there is a multiply block and I need something equivalent in
> RFNoC.
>
> Thanks,
>
> Craig
>
>
> *Craig F. Swanson*
>
> *Research Engineer II *
> *Information and Communications Laboratory*
> *Communications, Systems, and Spectrum Division*
> *Georgia Tech Research Institute*
>
>
> *Room 560 250 14th St NW *
> *Atlanta, GA 30318*
> *Cell: 770.298.9156 <(770)%20298-9156>*
> http://www.gtri.gatech.edu
> <https://mail.gtri.gatech.edu/owa/redir.aspx?C=c20925f2f0af4dd29329ddf0701ecfff&URL=http%3a%2f%2fwww.gtri.gatech.edu%2f>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20170609/288334fd/attachment-0001.html>
------------------------------
Message: 3
Date: Fri, 9 Jun 2017 17:33:42 +0000
From: "Hofer, Russell Dwayne" <[email protected]>
To: Leandro Echevarr?a <[email protected]>, "Attila Kinali"
<[email protected]>, "[email protected]"
<[email protected]>
Subject: Re: [USRP-users] [EXTERNAL] Re: low frequency ADC flicker
noise (Ettus x310)
Message-ID:
<[email protected]>
Content-Type: text/plain; charset="utf-8"
Thanks for your inputs and thoughts, they are very helpful. Leo, what did you
mean by two repeating images ? I think you may have been referring to Frf =
Fif +/- Flo; in our case, we are starting at 1 kHz to 2 MHz, so the two images
are at the same frequency.
See the attached plot (I hope this is visible in the mailing list). While the
knee of the 1/f noise is much lower than 10 MHz, you do get some 1/f noise @ 10
MHz; which is why I am interested in switching at a higher frequency.
[cid:[email protected]]
Thoughts so far ? to be honest, the thought of just building an entirely
different SDR with a low frequency friendly ADC is becoming more attractive.
That being said, we do want to stick with the x310 architecture because it
gives us some things for free (well, not really free, but less overall cost),
allowing us to focus our research efforts in other areas. I explored the
switch idea a bit more, and it does pose challenges; I am dropping it for now.
I have been looking into the ?simple? mixer solution; it appears to be the best
option so far. Noise and distortion are important considerations. I am
considering 50 MHz as the local oscillator mix frequency, anyone have any
advice/experience mixing in this arena? Unfortunately, most COTS mixers are
more geared toward high frequency applications (@ both IF and RF); the level of
community support is pretty low.
Russell
From: Leandro Echevarr?a [mailto:[email protected]]
Sent: Wednesday, May 31, 2017 10:42 AM
To: Attila Kinali; [email protected]<mailto:[email protected]>
Cc: Hofer, Russell Dwayne
Subject: [EXTERNAL] Re: [USRP-users] low frequency ADC flicker noise (Ettus
x310)
Hey Russell,
Could you not use a different frequency in the local oscillators in the
receiving path so as to convert to a slightly off-DC frequency and sample two
repeating images of the spectrum of your signal of interest? Of course, this is
assuming your signal is not already in baseband (which I assume given most
daughterboards for the x310 are for passband signals) and that you can perform
some sort of channel filtering around those 2 MHz of bandwidth you're
interested in before digitalizing (otherwise you'd get aliasing).
Regards,
Leo.
On Wed, May 31, 2017 at 1:22 PM Attila Kinali via USRP-users
<[email protected]<mailto:[email protected]>> wrote:
Moin
On Wed, 31 May 2017 11:50:51 -0400
"Marcus D. Leech via USRP-users"
<[email protected]<mailto:[email protected]>> wrote:
> On 2017-05-31 10:53, Hofer, Russell Dwayne via USRP-users wrote:
>
> > We have developed a system that utilizes Ettus X310 radios in order to
> > perform low frequency data collection (0 - 2 MHz). The system is
> > ruggedized for outdoor use, and utilizes an x310 along with some of our own
> > custom hardware. We have used these systems several times in multi-day
> > continuous collection intervals.
> >
> > There is one glaring concern that I found when doing noise comparisons
> > between the new Ettus based system and a legacy data collection system.
> > The low frequency 1/f flicker noise from the X310's ADC is significantly
> > impacting our performance, to the point where noise levels become
> > dramatically worse than our legacy system below 25 kHz (we decimate to 4
> > MSPS using the standard FPGA image, and then perform an FFT to observe the
> > ambient noise). 25 kHz is the cross-over point, and the noise level
> > continues to increase at a rate of 1/f - i.e. very strong noise levels
> > exist at lower frequencies. Not only that, there is a significant level of
> > flicker noise across the entire bandwidth; my testing shows that the ADC's
> > 1/f flicker noise doesn't level out until 10-20 MHz - so it appears that we
Yes, the flicker noise of high speed ADCs is quite high. Though a corner
frequency of >10MHz seems a bit too high. The measurements I have seen
(from Enrico Rubiola one or two years ago) suggest that the corner frequency
is usually in the order of 100kHz to 1MHz. Maybe there is power supply noise
or reference noise coupling into to the ADC.
> > 1) I know that mixing the low frequency signal to a higher frequency
> > may be an option, but I would rather not do so in order to avoid dealing
> > with the non-ideal characteristics of a mixer, specifically the
> > distortion. Or, is this the only practical way to fix it?
It is definitly one way. There are quite good mixers that can achieve
a low distortion, low noise upmixing.
> > 2) Has anyone explored the idea of chopping the input by using a
> > synchronized RF/sampling switch ? The idea is to multiply the low
> > frequency input signal by a square wave, and then after it is digitized,
> > multiply it by a square wave in the digital domain. The end result is that
> > the 1/f noise would still exist, but would be shifted higher in frequency
> > since it is only multiplied by the square wave one time in the digital
> > domain.
This is another possibiltiy, but the noise from the switches will dominate
anything else. Beside, finding switches that can handle a switching rate
of 4MHz are kind of rare. And I wouldn't trust them to be low noise either.
The mixer approach is way better and easier to do.
> > 3) Is there a different replacement ADC that may perform better (i.e.
> > have a chopper mode) ?
This is the easiest approach IMHO. The lower speed the ADC the lower
its 1/f noise corner frequency is. As you only need 4Msps you can use
one of the fasts SAR ADCs from Analog. I have not seen any 1/f corner
frequency of those, but I would be surprised if it's that high.
> High-speed, CMOS ADCs suffer from higher 1/F noise corners than there
> older bipolar counterparts.
>
> This article offers some insight:
>
> http://www.highfrequencyelectronics.com/index.php?option=com_content&view=article&id=1160:eliminate-high-speed-adc-flicker-noise-with-chopper-upgrade&catid=123&Itemid=189
>
>
> Installing a different ADC would be a *significant* amount of work,
> involving hardware surgery, significant FPGA changes, and significant
> changes in the UHD drivers. That would not be a trivial change.
Using a different ADC is by far easier than to get the right mixer
architecture to achieve low noise and low distortion.
That said: How about trying a different board?
We recently built a ADC/SDR board specifally with low noise applications
in mind. I haven't had time to really measure the 1/f corner yet, but
it is low enough (much smaller than 1MHz, probably 100kHz'ish)
that it didn't really show up in our measurements.
You can find the documentation and production files at [1].
I am currently writing a paper describing the design and how to use
it (which should be published at EFTF/IFCS next month).
Which way to go and what ADC/circuit architecture to use depends highly
on what your projects parameters are. You have told us only the bandwidth
sofar. We don't know anything about the noise level you want to achieve
or what impedance your signal has.
Attila Kinali
[1] http://www.ohwr.org/projects/r19-tdc-del-a/wiki
--
It is upon moral qualities that a society is ultimately founded. All
the prosperity and technological sophistication in the world is of no
use without that foundation.
-- Miss Matheson, The Diamond Age, Neil Stephenson
_______________________________________________
USRP-users mailing list
[email protected]<mailto:[email protected]>
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20170609/a624fa9d/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image002.jpg
Type: image/jpeg
Size: 15100 bytes
Desc: image002.jpg
URL:
<http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20170609/a624fa9d/attachment-0001.jpg>
------------------------------
Message: 4
Date: Fri, 09 Jun 2017 13:44:29 -0400
From: [email protected]
To: Jonathon Pendlum <[email protected]>
Cc: "Swanson, Craig" <[email protected]>,
[email protected]
Subject: Re: [USRP-users] Is there a MULTIPLY block in RFNoC?
Message-ID: <[email protected]>
Content-Type: text/plain; charset="us-ascii"
...and while you're at it, Craig, do a conjugate-multiply block as well
:)
[Radio astronomy interferometers...]
On 2017-06-09 13:24, Jonathon Pendlum via USRP-users wrote:
> Hi Craig,
>
> Depends on what you want to do. If you want to multiply by a constant, we
> have the digital gain RFNoC block. If you want to multiply two streams,
> you'll need to take mult.v, mult_rc.v, or complex_multiplier.v and wrap that
> in a RFNoC block.
>
> On Fri, Jun 9, 2017 at 9:50 AM, Swanson, Craig
> <[email protected]> wrote:
>
>> Jonathon,
>>
>> In gnuradio there is a multiply block and I need something equivalent in
>> RFNoC.
>>
>> Thanks,
>>
>> Craig
>>
>> CRAIG F. SWANSON
>> Research Engineer II
>>
>> _INFORMATION AND COMMUNICATIONS LABORATORY_
>> _COMMUNICATIONS, SYSTEMS, AND SPECTRUM DIVISION_
>> _Georgia Tech Research Institute_
>> Room 560
>> 250 14th St NW
>>
>> _Atlanta, GA 30318_
>> _Cell: 770.298.9156 [1]_
>> http://www.gtri.gatech.edu [2]
>
> _______________________________________________
> USRP-users mailing list
> [email protected]
> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
Links:
------
[1] tel:(770)%20298-9156
[2]
https://mail.gtri.gatech.edu/owa/redir.aspx?C=c20925f2f0af4dd29329ddf0701ecfff&URL=http%3a%2f%2fwww.gtri.gatech.edu%2f
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20170609/3ed07984/attachment-0001.html>
------------------------------
Message: 5
Date: Fri, 9 Jun 2017 14:15:22 -0500
From: Qurat-Ul-Ann Akbar <[email protected]>
To: Sumit Kumar <[email protected]>
Cc: "[email protected]" <[email protected]>
Subject: Re: [USRP-users] Is there a way to store modulated samples
generated by USRP
Message-ID:
<CAJwfNj1KgpXu-2xp4XPL=2or7kgml0efphqr0shfpjfznfv...@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"
Currently the USRP Sink block does not have any output port because it
sends the samples through the antenna. I want to save those samples right
before they reach the antenna. Can I modify the sink block such that it
sends the up-sampled samples to a file and not the antenna?
I am aware that you can store samples from USRP source but I am not sure
how to do that in software in USRP sink? Is that at all possible to get
those 2.4GHz samples instead of the baseband samples?
On Thu, Jun 8, 2017 at 11:06 PM, Sumit Kumar <[email protected]> wrote:
> If you are using GNU Radio to work with USRP, then there are examples in
> benchmark files.
>
> https://github.com/gnuradio/gnuradio/blob/master/gr-digital/examples/ofdm/
> benchmark_tx.py
>
> Its gives you option to dump binary data in a file by using --to-file
> <path to file> argument.
>
>
> On Thu, Jun 8, 2017 at 8:46 PM, Qurat-Ul-Ann Akbar via USRP-users <
> [email protected]> wrote:
>
>> Hi,
>>
>> Is there a way to store modulated samples by USRP in a file instead of
>> sending them through air ?
>>
>>
>> _______________________________________________
>> USRP-users mailing list
>> [email protected]
>> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>>
>>
>
>
> --
> --
> Sumit kumar
> Doctoral Student, UPMC
> Eurecom, BIOT
> France
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20170609/6ab3a77f/attachment-0001.html>
------------------------------
Message: 6
Date: Fri, 9 Jun 2017 15:35:18 -0500
From: Qurat-Ul-Ann Akbar <[email protected]>
To: [email protected]
Subject: [USRP-users] Retrieving samples up-sampled to 2.4GHz from the
USRPN210
Message-ID:
<cajwfnj1w31gvmjskpyjacunr1nwquauom3ge8rmfdlztbah...@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"
Hi,
I posted this question before too but I am going to re-phrase it to make it
clearer. I am using two USRPN210s with GNURadio 3.7. I want to transmit
WiFi packets from one USRP and receiver and demodulate at the other USRP
for my experiments. I have been having a lot of problems with extremely
low SNR and therefore nothing above 1MHz works. At even 5 MHz it doesn't
receive anything and the I don't see the shape of the OFDM signal at the
receiver at all (it remains at the -90 to -100 relative gain level), so 20
Mhz which is actually the WiFi BW, is a long shot. I think it is able to
decode at 1MHz because of a very narrowband. I have tried alot of things,
changing respective antenna gains, making sure no overruns underuns,
changing values of LO offsets, testing with three different USRPN210s but
nothing is working.
Therefore, now after trying a lot I want to take the samples from the USRP
i.e. the samples which are up-sampled to 2.4GHz and sent through the
antenna. *NOT *the baseband ones generated before thye are given to the
USRP sink block. And then I want to feed these samples to the receiver and
see if it can demodualte it. Then I can manually add noise for my
experiments. Basically, I want to remove the wireless medium and see if the
code is working and then do it by hand.
Is there any way I can take these samples from inside the USRP. Is there
any API that can allow me to do that?
I would be very grateful if anyone could help me out.
Best,
Annie
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20170609/2c6cb8f6/attachment-0001.html>
------------------------------
Message: 7
Date: Fri, 9 Jun 2017 14:17:35 -0700
From: Kyeong Su Shin <[email protected]>
To: Ettus mail list <[email protected]>
Subject: Re: [USRP-users] Retrieving samples up-sampled to 2.4GHz from
the USRPN210
Message-ID:
<CAGL0V3nHV4PuhdNyW2ZvBj2esQpPOCPw=m9tkroklwmcjbt...@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"
Dear Qurat-Ul-Ann Akbar:
The samples are in baseband, and mixed up to 2.4GHz by an *analog* mixer *after
the DAC*. The receiver then receives and brings them back to the baseband
using an *analog* mixer, samples them using two ADCs, and then passes them
back to your computer. So, you really cannot get a *digital* samples of
such signals from the USRPs. Because they simply do not exist.
You can digitally mix-up the signal to 2.4GHz without using USRPs (simply
multiply a sine wave to the baseband signal and apply a filter), but I am
not sure if you want to do that. You will need to use 2.4 - 2.5GS/s at
minimum, which is a lot of data.
I don't really do WIFI experiments with my USRPs, but I don't think if this
should be happening. What daughterboards and what RF gain level do you use?
Regards,
Kyeong Su Shin
On Fri, Jun 9, 2017 at 1:35 PM, Qurat-Ul-Ann Akbar via USRP-users <
[email protected]> wrote:
> Hi,
>
> I posted this question before too but I am going to re-phrase it to make
> it clearer. I am using two USRPN210s with GNURadio 3.7. I want to transmit
> WiFi packets from one USRP and receiver and demodulate at the other USRP
> for my experiments. I have been having a lot of problems with extremely
> low SNR and therefore nothing above 1MHz works. At even 5 MHz it doesn't
> receive anything and the I don't see the shape of the OFDM signal at the
> receiver at all (it remains at the -90 to -100 relative gain level), so 20
> Mhz which is actually the WiFi BW, is a long shot. I think it is able to
> decode at 1MHz because of a very narrowband. I have tried alot of things,
> changing respective antenna gains, making sure no overruns underuns,
> changing values of LO offsets, testing with three different USRPN210s but
> nothing is working.
>
> Therefore, now after trying a lot I want to take the samples from the USRP
> i.e. the samples which are up-sampled to 2.4GHz and sent through the
> antenna. *NOT *the baseband ones generated before thye are given to the
> USRP sink block. And then I want to feed these samples to the receiver and
> see if it can demodualte it. Then I can manually add noise for my
> experiments. Basically, I want to remove the wireless medium and see if the
> code is working and then do it by hand.
>
> Is there any way I can take these samples from inside the USRP. Is there
> any API that can allow me to do that?
>
> I would be very grateful if anyone could help me out.
>
> Best,
> Annie
>
> _______________________________________________
> USRP-users mailing list
> [email protected]
> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20170609/d4872c7c/attachment-0001.html>
------------------------------
Message: 8
Date: Sat, 10 Jun 2017 14:39:00 +0000
From: Benny Alexandar <[email protected]>
To: "[email protected]" <[email protected]>
Subject: [USRP-users] Difference between Clock Recovery and Audio rate
synchronization.
Message-ID:
<hk2pr02mb13306cb28793c146ea9deba28b...@hk2pr02mb1330.apcprd02.prod.outlook.com>
Content-Type: text/plain; charset="iso-8859-1"
Hi,
If I understood correctly, clock recovery is for demodulation of digital
signals,
where as audio rate synchronization is matching the audio output frequency to
transmitter
audio clock.
-ben
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20170610/8c757451/attachment-0001.html>
------------------------------
Subject: Digest Footer
_______________________________________________
USRP-users mailing list
[email protected]
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
------------------------------
End of USRP-users Digest, Vol 82, Issue 10
******************************************
