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: SBX TDD (Ian Buckley)
2. Re: Stepped Frequency Radar (Marcus M?ller)
3. Re: SBX TDD (Vladica Sark)
4. Re: SBX TDD (John Malsbury)
5. Re: SBX TDD (Vladica Sark)
6. Re: USRP B210 quadrature imbalance issue. (Simon Pack)
7. Phase comparator using a N200 (Juha Vierinen)
8. Re: uhd_rx_cfile: getting constant O-overflows using the USRP
B200 with debian (Marcus M?ller)
9. Reminder: Call for Proposals and Registration for GRCon14
(Michael Dickens)
----------------------------------------------------------------------
Message: 1
Date: Fri, 21 Mar 2014 09:27:14 -0700
From: Ian Buckley <[email protected]>
To: Vladica Sark <[email protected]>
Cc: [email protected]
Subject: Re: [USRP-users] SBX TDD
Message-ID: <[email protected]>
Content-Type: text/plain; charset=iso-8859-1
Absolutely switch EN_TX off, as well as the TX/RX switch. The TX/RX switch only
attenuates approximately 20dB the "not selected" path.
On Mar 21, 2014, at 6:50 AM, Vladica Sark <[email protected]> wrote:
> The output power of IQ modulator (ADL5375-05) using the DSOP pin
> (EN_TX pin).
>
> On 03/21/2014 02:47 PM, Marcus M?ller wrote:
>> I don't understand. What do you want to turn off?
>>
>>
>> On 21.03.2014 14:39, Vladica Sark wrote:
>>> Thanks Marcus,
>>
>>> The question is, if I do not turn off the RF stage of the IQ
>>> modulator, would it burn the antenna switch for TX/RX, or
>>> eventually some PA on the output signal path?
>>
>>> BR, Vladica
>>
>>
>>> On 03/21/2014 02:15 PM, Marcus M?ller wrote:
>>>> Hi Vladica,
>>>>
>>>> this depends on your requirements. USRPs have independent TX/RX
>>>> chains, so you can just transmit samples whenever you want
>>>> (using burst flags, timed commands, zero padded sample streams,
>>>> or modified FPGA, which of course is way more complex), and RX
>>>> all the time.
>>>>
>>>> Depending on what you change in the FPGA, you might or might
>>>> not be able to use the aforementioned features, depending on
>>>> how "high level" or "host interface specific" they are and how
>>>> far up the processing chain you can/want to go.
>>>>
>>>> Greetings, Marcus
>>>>
>>>> On 21.03.2014 13:53, Vladica Sark wrote:
>>>>> Hi there,
>>>>
>>>>> Anybody knows how to use the TDD with SBX board?
>>>>
>>>>> Is it enough to activate/deactivate U6 switch (TX/RX), or I
>>>>> should also turn off the transmitter?
>>>>
>>>>> I am changing a bit the verilog code, so I want to do this in
>>>>> hardware.
>>>>
>>>>> BR, Vladica
>>>>
>>>>> _______________________________________________ USRP-users
>>>>> mailing list [email protected]
>>>>> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>>>>
>>>>>
>>>>
>>>>
>>>>>
>>
>>>>>
> _______________________________________________ USRP-users mailing
>>>> list [email protected]
>>>> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>>>>
>>
>>>>
>>>>
>>> _______________________________________________ USRP-users
>>> mailing list [email protected]
>>> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>>
>>>
>>
>> _______________________________________________ USRP-users mailing
>> list [email protected]
>> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>>
>
> _______________________________________________
> USRP-users mailing list
> [email protected]
> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
------------------------------
Message: 2
Date: Fri, 21 Mar 2014 17:36:27 +0100
From: Marcus M?ller <[email protected]>
To: [email protected]
Subject: Re: [USRP-users] Stepped Frequency Radar
Message-ID: <[email protected]>
Content-Type: text/plain; charset=ISO-8859-1
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi Haroon Muhammad,
TX and RX are generally derived of the same clock, and thus phase
coherent, if that answers your question.
Greetings,
Marcus
On 21.03.2014 14:38, Haroon Muhammad wrote:
> Hi!
>
> Please advice if the WBX along with N-210 can be used to implement
> a coherent Stepped Frequency Radar which detects the phase
> difference at each transmitted frequency? Can any other
> daughterboard serve the purpose?
>
> Regards
>
>
>
> _______________________________________________ USRP-users mailing
> list [email protected]
> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
iQEcBAEBAgAGBQJTLGqLAAoJEBQ6EdjyzlHtBXwH/15jm+E4zSa3nPojSSLGWSeb
mK8K3JSbR53XdHOvBKql1BxzwHCnubpLMFJERQsoAByx5iP9BSgfFScTTJtihlZw
E9vR4+yWQvkOfGQILsWhWph4lMC38YbqoAD/LYTDWLidudqnCkKh+CISFKS3Up4+
b8EO9u7GhMmvoG4qWYcb0LBhrPH1WB24qtGK02aWfFIxcNhV7PrSrDIjqKJkvM8B
rj4g9bjb6JQ/JC6zZcA/4/2xDc7OmXtwP8tNR56c5UJX+Q81QWwYzMWhGQe2UjdP
I5FeJbLSca9JRKN0P6PWCAaYYZ5aHQbw4sW4a88e1Ppy4sz+pQpraB/PgakycXE=
=FzC+
-----END PGP SIGNATURE-----
------------------------------
Message: 3
Date: Fri, 21 Mar 2014 18:11:55 +0100
From: Vladica Sark <[email protected]>
To: "[email protected]" <[email protected]>
Subject: Re: [USRP-users] SBX TDD
Message-ID: <[email protected]>
Content-Type: text/plain; charset=ISO-8859-1
Thanks Ian,
As I look now at the datasheet it says 20 dB.
If I transmit zeros, probably it would be ok, but
I noticed that the carrier is being propagated in the
TX path, when zeros are transmitted, so this would probably saturate the
ADC and destroy the RX signal.
Of course, I needed somebody to confirm this, and save me from
experimenting all the possible combinations :)
BR,
Vladica
On 03/21/2014 05:27 PM, Ian Buckley wrote:
> Absolutely switch EN_TX off, as well as the TX/RX switch. The TX/RX switch
> only attenuates approximately 20dB the "not selected" path.
>
> On Mar 21, 2014, at 6:50 AM, Vladica Sark <[email protected]> wrote:
>
>> The output power of IQ modulator (ADL5375-05) using the DSOP pin
>> (EN_TX pin).
>>
>> On 03/21/2014 02:47 PM, Marcus M?ller wrote:
>>> I don't understand. What do you want to turn off?
>>>
>>>
>>> On 21.03.2014 14:39, Vladica Sark wrote:
>>>> Thanks Marcus,
>>>
>>>> The question is, if I do not turn off the RF stage of the IQ
>>>> modulator, would it burn the antenna switch for TX/RX, or
>>>> eventually some PA on the output signal path?
>>>
>>>> BR, Vladica
>>>
>>>
>>>> On 03/21/2014 02:15 PM, Marcus M?ller wrote:
>>>>> Hi Vladica,
>>>>>
>>>>> this depends on your requirements. USRPs have independent TX/RX
>>>>> chains, so you can just transmit samples whenever you want
>>>>> (using burst flags, timed commands, zero padded sample streams,
>>>>> or modified FPGA, which of course is way more complex), and RX
>>>>> all the time.
>>>>>
>>>>> Depending on what you change in the FPGA, you might or might
>>>>> not be able to use the aforementioned features, depending on
>>>>> how "high level" or "host interface specific" they are and how
>>>>> far up the processing chain you can/want to go.
>>>>>
>>>>> Greetings, Marcus
>>>>>
>>>>> On 21.03.2014 13:53, Vladica Sark wrote:
>>>>>> Hi there,
>>>>>
>>>>>> Anybody knows how to use the TDD with SBX board?
>>>>>
>>>>>> Is it enough to activate/deactivate U6 switch (TX/RX), or I
>>>>>> should also turn off the transmitter?
>>>>>
>>>>>> I am changing a bit the verilog code, so I want to do this in
>>>>>> hardware.
>>>>>
>>>>>> BR, Vladica
>>>>>
>>>>>> _______________________________________________ USRP-users
>>>>>> mailing list [email protected]
>>>>>> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>>>>>
>>>>>>
>>>>>
>>>>>
>>>>>>
>>>
>>>>>>
>> _______________________________________________ USRP-users mailing
>>>>> list [email protected]
>>>>> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>>>>>
>>>
>>>>>
>>>>>
>>>> _______________________________________________ USRP-users
>>>> mailing list [email protected]
>>>> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>>>
>>>>
>>>
>>> _______________________________________________ USRP-users mailing
>>> list [email protected]
>>> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>>>
>>
>> _______________________________________________
>> USRP-users mailing list
>> [email protected]
>> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>
------------------------------
Message: 4
Date: Fri, 21 Mar 2014 10:39:43 -0700
From: John Malsbury <[email protected]>
To: Vladica Sark <[email protected]>
Cc: "[email protected]" <[email protected]>
Subject: Re: [USRP-users] SBX TDD
Message-ID:
<CAN5WegQ4rikM=nrvmtzeguboiazfo4xhu9w05gnumdyw-m9...@mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"
Vladica,
I might be misunderstanding your question, but can you not just use
timed-bursts? See slides starting with slide 18 of this presentation for
more info:
http://gnuradio.squarespace.com/storage/grcon13_presentations/grcon13_malsbury_phy_mac_primer.pdf
Is this with GNU Radio or UHD only?
-John
On Fri, Mar 21, 2014 at 10:11 AM, Vladica Sark <[email protected]>wrote:
> Thanks Ian,
>
> As I look now at the datasheet it says 20 dB.
> If I transmit zeros, probably it would be ok, but
> I noticed that the carrier is being propagated in the
> TX path, when zeros are transmitted, so this would probably saturate the
> ADC and destroy the RX signal.
>
> Of course, I needed somebody to confirm this, and save me from
> experimenting all the possible combinations :)
>
> BR,
> Vladica
>
> On 03/21/2014 05:27 PM, Ian Buckley wrote:
> > Absolutely switch EN_TX off, as well as the TX/RX switch. The TX/RX
> switch only attenuates approximately 20dB the "not selected" path.
> >
> > On Mar 21, 2014, at 6:50 AM, Vladica Sark <[email protected]> wrote:
> >
> >> The output power of IQ modulator (ADL5375-05) using the DSOP pin
> >> (EN_TX pin).
> >>
> >> On 03/21/2014 02:47 PM, Marcus M?ller wrote:
> >>> I don't understand. What do you want to turn off?
> >>>
> >>>
> >>> On 21.03.2014 14:39, Vladica Sark wrote:
> >>>> Thanks Marcus,
> >>>
> >>>> The question is, if I do not turn off the RF stage of the IQ
> >>>> modulator, would it burn the antenna switch for TX/RX, or
> >>>> eventually some PA on the output signal path?
> >>>
> >>>> BR, Vladica
> >>>
> >>>
> >>>> On 03/21/2014 02:15 PM, Marcus M?ller wrote:
> >>>>> Hi Vladica,
> >>>>>
> >>>>> this depends on your requirements. USRPs have independent TX/RX
> >>>>> chains, so you can just transmit samples whenever you want
> >>>>> (using burst flags, timed commands, zero padded sample streams,
> >>>>> or modified FPGA, which of course is way more complex), and RX
> >>>>> all the time.
> >>>>>
> >>>>> Depending on what you change in the FPGA, you might or might
> >>>>> not be able to use the aforementioned features, depending on
> >>>>> how "high level" or "host interface specific" they are and how
> >>>>> far up the processing chain you can/want to go.
> >>>>>
> >>>>> Greetings, Marcus
> >>>>>
> >>>>> On 21.03.2014 13:53, Vladica Sark wrote:
> >>>>>> Hi there,
> >>>>>
> >>>>>> Anybody knows how to use the TDD with SBX board?
> >>>>>
> >>>>>> Is it enough to activate/deactivate U6 switch (TX/RX), or I
> >>>>>> should also turn off the transmitter?
> >>>>>
> >>>>>> I am changing a bit the verilog code, so I want to do this in
> >>>>>> hardware.
> >>>>>
> >>>>>> BR, Vladica
> >>>>>
> >>>>>> _______________________________________________ USRP-users
> >>>>>> mailing list [email protected]
> >>>>>> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
> >>>>>
> >>>>>>
> >>>>>
> >>>>>
> >>>>>>
> >>>
> >>>>>>
> >> _______________________________________________ USRP-users mailing
> >>>>> list [email protected]
> >>>>> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
> >>>>>
> >>>
> >>>>>
> >>>>>
> >>>> _______________________________________________ USRP-users
> >>>> mailing list [email protected]
> >>>> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
> >>>
> >>>>
> >>>
> >>> _______________________________________________ USRP-users mailing
> >>> list [email protected]
> >>> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
> >>>
> >>
> >> _______________________________________________
> >> USRP-users mailing list
> >> [email protected]
> >> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
> >
>
> _______________________________________________
> 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/20140321/0d9037f1/attachment-0001.html>
------------------------------
Message: 5
Date: Fri, 21 Mar 2014 18:48:27 +0100
From: Vladica Sark <[email protected]>
To: "[email protected]" <[email protected]>
Subject: Re: [USRP-users] SBX TDD
Message-ID: <[email protected]>
Content-Type: text/plain; charset=ISO-8859-1
Hi John,
This looks nice, but I have to retransmit what I have received,
and the time from receiving to retransmitting the same thing should be
in us (microseconds). If I transfer it on a PC and schedule it for
retransmission, than it would take some ms. Also, some preamble
detection is needed, which I am already doing in FPGA.
You are right it can be done with timed-bursts, but I didn't explained
my complete scenario. So in this case because the delay from receiving
the frame, to retransmission is quite small, it must be done in hardware.
BR,
Vladica
On 03/21/2014 06:39 PM, John Malsbury wrote:
> Vladica,
>
> I might be misunderstanding your question, but can you not just use
> timed-bursts? See slides starting with slide 18 of this presentation
> for more info:
>
> http://gnuradio.squarespace.com/storage/grcon13_presentations/grcon13_malsbury_phy_mac_primer.pdf
>
> Is this with GNU Radio or UHD only?
>
> -John
>
>
>
> On Fri, Mar 21, 2014 at 10:11 AM, Vladica Sark <[email protected]
> <mailto:[email protected]>> wrote:
>
> Thanks Ian,
>
> As I look now at the datasheet it says 20 dB.
> If I transmit zeros, probably it would be ok, but
> I noticed that the carrier is being propagated in the
> TX path, when zeros are transmitted, so this would probably saturate the
> ADC and destroy the RX signal.
>
> Of course, I needed somebody to confirm this, and save me from
> experimenting all the possible combinations :)
>
> BR,
> Vladica
>
> On 03/21/2014 05:27 PM, Ian Buckley wrote:
> > Absolutely switch EN_TX off, as well as the TX/RX switch. The
> TX/RX switch only attenuates approximately 20dB the "not selected" path.
> >
> > On Mar 21, 2014, at 6:50 AM, Vladica Sark <[email protected]
> <mailto:[email protected]>> wrote:
> >
> >> The output power of IQ modulator (ADL5375-05) using the DSOP pin
> >> (EN_TX pin).
> >>
> >> On 03/21/2014 02:47 PM, Marcus M?ller wrote:
> >>> I don't understand. What do you want to turn off?
> >>>
> >>>
> >>> On 21.03.2014 14 <tel:21.03.2014%2014>:39, Vladica Sark wrote:
> >>>> Thanks Marcus,
> >>>
> >>>> The question is, if I do not turn off the RF stage of the IQ
> >>>> modulator, would it burn the antenna switch for TX/RX, or
> >>>> eventually some PA on the output signal path?
> >>>
> >>>> BR, Vladica
> >>>
> >>>
> >>>> On 03/21/2014 02:15 PM, Marcus M?ller wrote:
> >>>>> Hi Vladica,
> >>>>>
> >>>>> this depends on your requirements. USRPs have independent TX/RX
> >>>>> chains, so you can just transmit samples whenever you want
> >>>>> (using burst flags, timed commands, zero padded sample streams,
> >>>>> or modified FPGA, which of course is way more complex), and RX
> >>>>> all the time.
> >>>>>
> >>>>> Depending on what you change in the FPGA, you might or might
> >>>>> not be able to use the aforementioned features, depending on
> >>>>> how "high level" or "host interface specific" they are and how
> >>>>> far up the processing chain you can/want to go.
> >>>>>
> >>>>> Greetings, Marcus
> >>>>>
> >>>>> On 21.03.2014 13 <tel:21.03.2014%2013>:53, Vladica Sark wrote:
> >>>>>> Hi there,
> >>>>>
> >>>>>> Anybody knows how to use the TDD with SBX board?
> >>>>>
> >>>>>> Is it enough to activate/deactivate U6 switch (TX/RX), or I
> >>>>>> should also turn off the transmitter?
> >>>>>
> >>>>>> I am changing a bit the verilog code, so I want to do this in
> >>>>>> hardware.
> >>>>>
> >>>>>> BR, Vladica
> >>>>>
> >>>>>> _______________________________________________ USRP-users
> >>>>>> mailing list [email protected]
> <mailto:[email protected]>
> >>>>>>
> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
> >>>>>
> >>>>>>
> >>>>>
> >>>>>
> >>>>>>
> >>>
> >>>>>>
> >> _______________________________________________ USRP-users mailing
> >>>>> list [email protected]
> <mailto:[email protected]>
> >>>>> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
> >>>>>
> >>>
> >>>>>
> >>>>>
> >>>> _______________________________________________ USRP-users
> >>>> mailing list [email protected]
> <mailto:[email protected]>
> >>>> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
> >>>
> >>>>
> >>>
> >>> _______________________________________________ USRP-users mailing
> >>> list [email protected] <mailto:[email protected]>
> >>> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
> >>>
> >>
> >> _______________________________________________
> >> USRP-users mailing list
> >> [email protected] <mailto:[email protected]>
> >> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
> >
>
> _______________________________________________
> USRP-users mailing list
> [email protected] <mailto:[email protected]>
> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>
>
------------------------------
Message: 6
Date: Fri, 21 Mar 2014 18:04:46 -0000
From: "Simon Pack" <[email protected]>
To: <[email protected]>
Subject: Re: [USRP-users] USRP B210 quadrature imbalance issue.
Message-ID:
<!&!AAAAAAAAAAAYAAAAAAAAAPw/1cWaKWpJvxMtGFg8nlHCgAAAEAAAAOoEexZh0rVNt6Xcj1IDS3YBAAAAAA==@commagility.com>
Content-Type: text/plain; charset="us-ascii"
Hi Martin.
The issue occurs at all frequencies. I chose 4GHz as the initial quadrature
imbalance is quite bad (quadrature image only 30dB down) and so it
illustrated the problem quite well.
Incidentally, I have found I can also get identical results if I use the
command:
uhd_siggen_gui -s 32e6 -f 4e9 -x 5e6 -g 76
If I type the frequency of 4G into the GUI, the quadrature error doesn't
change, but if I type 4.0001G it gets way better. Typing 4G back in returns
to the original frequency with the improved quadrature balance (with the
image 57dB down).
Clearly the AD9361 transmit quadrature correction algorithm is not getting
triggered at start-up.
If I stop the program and start it again, the quadrature imbalance goes back
to 30dB.
I originally noticed this issue when transmitting 20MHz LTE. The quadrature
error is pushing my EVM from 1% to 2.7%.
Cheers,
Simon
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20140321/ddd645d0/attachment-0001.html>
------------------------------
Message: 7
Date: Fri, 21 Mar 2014 20:49:32 +0000
From: Juha Vierinen <[email protected]>
To: "[email protected]" <[email protected]>
Subject: [USRP-users] Phase comparator using a N200
Message-ID:
<cagbkif2u60ksuhfyvs56vpqulsoe5gkd9qwu7f+dj35yc1a...@mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"
Hi,
For the last few weeks we've been trying to make an inexpensive phase
comparator using a USRP N200. I thought I'd share some of our experiences
here for the benefit of somebody else attempting this.
The goal is to measure relative time delay between two sinusoidal signals
on two channels by measuring the relative phase of the two channels. In our
case, we are using 5 MHz.
We first tried the most obvious approach, ie., using the two DDCs with with
the standard firmware. However, we encountering systematic relative phase
changes between two channels that were variable over time. We did remove
the residual frequency error that was there due to the finite length tuning
word, so that wasn't the cause of the problem. It was as if the two NCOs
weren't in exactly the same phase at each time step. These phase changes
are small enough that you won't notice them unless looking at the cross
channel relative phase difference very carefully (~`100 picoseconds
peak-to-peak and zero-mean). I never figured out the exact cause of this
problem, but tuning the DDC to 0 removed the issue. I ended up removing the
CIC and the half-band filter and just streamed decimated ADC counts on the
two channels, avoiding nearly all fpga processing. This is a fairly easy
modification to the UHD Verilog code (this would also be a nice to have
standard feature for the USRP N200 fpga image).
We next noticed that single precision filtering did not provide enough
numerical accuracy. We were seeing 10 to 20 picosecond jumps, which went
away when doing the filtering in double precision. For this purpose, I
wrote a new gnuradio signal processing block that performed 25 MHz -> 10 Hz
digital downconversion with double precision floating point.
We are right now at a point where we can get relative time stability of
approximately 4e-14 in 10 seconds and 5e-15 over 100 seconds between the
two channels, which is good enough to resolve sub picosecond relative time
changes between the two channels. I've attached a plot where we tested
relative time delay between two channels, using a trombone to vary the
length of one of the channels.
juha
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20140321/6d56baeb/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: trombone_test.png
Type: image/png
Size: 33864 bytes
Desc: not available
URL:
<http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20140321/6d56baeb/attachment-0001.png>
------------------------------
Message: 8
Date: Fri, 21 Mar 2014 23:48:16 +0100
From: Marcus M?ller <[email protected]>
To: [email protected]
Subject: Re: [USRP-users] uhd_rx_cfile: getting constant O-overflows
using the USRP B200 with debian
Message-ID: <[email protected]>
Content-Type: text/plain; charset=ISO-8859-1
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi Ingmar,
all points to your bottleneck still being disk write speed.
This is still a lot:
samp-rate * bytes_per_part * parts_per_complex
=30000000/s * 4B * 2
= 240e6B/s = 240 Megabytes per second!
So your calculation with 350 MB (you said 350mb, but case is important
here... not millibit, but MegaBytes) seems a bit off.
It shows the core of your problem: SSD writes happen in Bursts and
what you can measure and display is only a temporal average of
throughput.
What most probably happens is that though your drive could possibly
sustain sufficient average write rates, at times it just can't get the
data away fast enough.
This happens for a lot of reasons - operating system scheduling,
filesystem, non-constant write speed over areas, USB possibly
interfering by causing interrupt....
Solution to this? Try increasing kernel write buffers. Use that drive
exclusively to write samples, use a filesystem that is optimized for
your case. Don't use a notebook but a In the end, everything boils
down to the fact that general purpose computers trying to do real time
processing always need to be heavily oversized to guarantee a certain
throughput.
Greetings,
Marcus
On 21.03.2014 10:54, Ingmar Splitt wrote:
> Hi developers,
>
> I have a fresh gnu radio installed with pybombs for a USRP B200. It
> runs on a Lenovo X230 with debian testing x64. When I use
> uhd_rx_cfile i get constant "O"-overruns (output given below).
>
> Taken Steps:
>
> The uhd-benchmark runs without problems:
>
> sudo ./benchmark_rate --duration 600 --rx_rate 32000000
>
> So the USB3-Port isn't the Problem. Storing in ram is also fine:
>
> sudo ./uhd_rx_cfile -f 2445000000 --samp-rate=30000000
> /tmp/test.cfile
>
> I use a Samsung Evo 840 1TB (newest firmware) and did the
> SSD-Optimizations (https://wiki.debian.org/SSDOptimization) while
> searching for the error. HDPerm-, DD- and copy-benchmarks give
> 500mb/s for the first seconds, later constant ~400mb/s writing
> speed. Debugging with "iotop" brings (output below) 90%
> io-call-workload for the benchmarks, but 0% for the uhd_rx_cfile
> which writes only with 350mb/s max. So everything should work as
> expected. Lowering the Sample-Rate only decreases the "O"s, but
> they still happen. I measured the writingspeed on disk and the
> number of the "O"s for 1 minute runtime. Running X and iotop in the
> back has nearly no impact.
>
> MHz Mb/s "O"-count 32 310 350 16 180
> 107 8 106 32 4 53 6 2
> 27 3
>
>
> Have you got any idea how to find and fix the bottleneck?
>
> ##################################################
>
> user@debian:~/tools/grc/target/bin$ sudo ./uhd_rx_cfile -f
> 2445000000 --samp-rate=30000000 /test/test.cfile -v
>
> linux; GNU C++ version 4.8.2; Boost_105400;
> UHD_003.007.000-1-ga8caec5f
>
> -- Operating over USB 3. -- Initialize CODEC control... --
> Initialize Radio control... -- Performing register loopback test...
> pass -- Performing CODEC loopback test... pass -- Asking for clock
> rate 32 MHz -- Actually got clock rate 32 MHz -- Performing timer
> loopback test... pass
>
> UHD Warning: The hardware does not support the requested RX sample
> rate: Target sample rate: 30.000000 MSps Actual sample rate:
> 32.000000 MSps Using mid-point gain of 36.5 ( 0.0 - 73.0 )
> Motherboard: B200 (E6R04Z7B2) Daughterboard: B200 (RX2, A:A) Rx
> gain: 36.5 Rx baseband frequency: 2.445G Rx DDC frequency:
> -596.046m Rx Sample Rate: 32M Receiving samples until Ctrl-C
> Writing 32-bit complex floats Output filename: /test/test.cfile
> OOOOOOOOOOOO
>
>
> #########################################
>
> iotop
>
> TID PRIO USER DISK READ DISK WRITE SWAPIN IO>
> COMMAND 171 be/4 root 3.91 K/s 0.00 B/s 0.00 % 6.63 %
> [kworker/u16:3] 2656 be/4 root 0.00 B/s 0.00 B/s 0.00 %
> 0.04 % X :0 -auth /var/run/lightdm~olisten tcp vt7 -novtswitch 3502
> be/4 root 0.00 B/s 236.78 M/s 0.00 % 0.00 % python2
> ./uhd_rx_cfile -f 2~0000000 /test/test.cfile -v
>
>
> ########################################
>
> messung@debian:~/tools/grc/target/bin$ sudo hdparm -tT /dev/sda
>
> /dev/sda: Timing cached reads: 17832 MB in 2.00 seconds =
> 8921.71 MB/sec Timing buffered disk reads: 1526 MB in 3.00 seconds
> = 508.55 MB/sec
>
> _______________________________________________ USRP-users mailing
> list [email protected]
> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
iQEcBAEBAgAGBQJTLMGwAAoJEBQ6EdjyzlHtISgH+wcolkwP2Ekw71O9f+pF8lGp
08+2wkmhPP2yGxOVP/Z2L/vOqzvTR4Flj51k6y2cGZvKthO+MPBmcCkBt3v9FSLK
50jPjEVNYyMCxgaCfhOtF4n15U7VNlx5sbquF8Y9hNOxf6zdQMtjzEYu22MEtwek
Xewuy8tcD7mwQQUv2KL5KjsRofd23umftmC+aiMMXH+GCwDq0mgmZaA8Nn6g0BIf
z5UT8jxH13DBcoxeq8V1Cuz8M5ClYlo0bAHeFNiwZHPvOAU22h/dTN2qN1yseZdW
RmGV731aIpF4sTDzcrEEAPWfVD1w9sW5q7TuaVgiz2SKPMwW3SmsNzyLjELU744=
=seB2
-----END PGP SIGNATURE-----
------------------------------
Message: 9
Date: Sat, 22 Mar 2014 10:45:40 -0400
From: Michael Dickens <[email protected]>
To: GNU Radio Discussion List <[email protected]>, USRP Users
Discussion List <[email protected]>
Subject: [USRP-users] Reminder: Call for Proposals and Registration
for GRCon14
Message-ID: <[email protected]>
Content-Type: text/plain; charset=us-ascii
Greetings GNU Radio and USRP Communities,
The GNU Radio Conference 2014 (GRCon14) will take place 2014 September 15-19,
at the District Architecture Center (AIA DAC) in Washington, DC, USA [0].
Registration is open, and available via the GRCon14 Main Website [1]. If you
are interested in presenting at GRCon14, please submit your abstract via the
GRCon14 Call For Presentations Website [2]. We already have a few submissions;
please keep them coming! The submission period will close on 2014 April 04 --
just under 2 weeks away -- and we will announce the selected presentations on
2014 April 14. Tom Rondeau and I will be leading the selection process.
Best Regards,
Michael Dickens
GNU Radio Conference 2014 Co-Organizer (one of a few)
[0] GRCon14 Location : AIA DAC Website
http://aiadac.com/
[1] GRCon14 Main Website
http://gnuradio.squarespace.com/gnu-radio-conference-2014/
[2] GRCon14 Call For Presentations
http://gnuradio.squarespace.com/grc2014-call-for-presentations/
------------------------------
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 43, Issue 22
******************************************
