I did an experiment and the values match pretty much exactly what theory
would predict.
I don't have a splitter, so I used a 1m cable on the channel A of my B210
as my reference, and compared a 1m and a 2m RF cable on channel B.
The 1m cables have a velocity factor of 0.83 (enforcer SI-195), and the 2m
cable (RG223) has a velocity factor of 0.66. The difference in electrical
lengths is 220 degrees ((((2 meter) * (100 MHz) / (c * 0.66)) - ((1 meter)
* (100 MHz) / (c * 0.83)))*360 = 219.2)
I ran the UHD example txrx_loopback_to_file with the following command:
"./txrx_loopback_to_file --tx-rate=4e6 --rx-rate=4e6 --tx-freq=100e6
--rx-freq=99.99e6 --tx-channels="0,1" --rx-channels="0,1" --tx-gain=45
--rx-gain=45 --type=float"
I then analyzed the results in octave using the following commands:
data1 = read_complex_binary('usrp_samples.00.dat' );
data2 = read_complex_binary('usrp_samples.01.dat' );
data2_ang = angle(data2);
data1_ang = angle(data1);
ang_delta = mod(data1_ang - data2_ang + pi, 2*pi)-pi;
mean_delta = mean(ang_delta)*180/pi
With both channels using the 1m cable, the mean phase delta was ~1 degree
as expected, and with the second channel using the 2m cable, the phase
difference returns as -142 degrees, which is pretty much exactly what was
expected based on the difference in the electrical length of the two
cables. The attached plot also clearly shows the second signal lagging by
> 180 degrees. The code returns angles between -180 and +180 to make it
easier to average the reference where the phase is close to zero.
I'm not sure why you are seeing a much smaller result than you expect (the
velocity factor will make the change in electrical length greater instead
of smaller), but you might want to try simulating your phase extraction
using signals with a known phase offset (generated in gnuradio perhaps) to
debug what's going on.
Cheers,
William
On Tue, May 8, 2018 at 2:34 PM, Reinhold Frederick William Hollender <
whollen...@gmail.com> wrote:
>
>
> On Tue, May 8, 2018 at 1:51 PM, Marcus D. Leech via USRP-users <
> usrp-users@lists.ettus.com> wrote:
>
>> On 05/08/2018 02:17 PM, Reinhold Frederick William Hollender via
>> USRP-users wrote:
>>
>>
>>
>> The additional cable length acts as a time delay, so I would calculate
>> the expected phase difference based on the frequency of the baseband
>> signal, not the RF signal. I have a couple different cable lengths, so I
>> can try running an experiment if I get some time later today.
>>
>>>
>>> Thanks again,
>>> Steve
>>>
>>> P.S. How do you respond within the mailing list? Each time I send a new
>>> mail with similar subject and copying the previous replies manually...
>>>
>>
>> The list is copied on each email, so simply reply to all instead of
>> replying to just the sender.
>>
>> Cheers,
>> William
>>
>>
>>>
>>> The phase offset at baseband should be the same as at RF. If it wasn't,
>> then things like PSK modulators that do their modulation at baseband would
>> not
>> work correctly at RF. That clearly isn't the case.
>>
>> I do phase-sensitive interferometry for radio astronomy. My fringe rates
>> are exactly what theory predicts from the *RF* frequency and antenna
>> spacing. If there was "phase scaling", that would not be the case.
>>
>>
> Doh! You are correct of course. That's what I get for relying on my
> intuition instead of actually looking at the math!
>
>>
>>
>>
>>>
>>> Steve , I got the feeling from your original message, though you didn’t say
>>> it outright, that you might be using the same USRP its self as the signal
>>> generator with loopback cables.
>>>
>>> If so, beware “self receive” via paths other than your calibrated length
>>> cables confusing the results, a sensitive full duplex radio like this hears
>>> its self via various leakage paths especially with RX and TX tuned to the
>>> same freq.
>>>
>>> (I’m assuming in the original email the cable you added was 50.8cm or
>>> 0.508m, right?)
>>>
>>> -Ian
>>>
>>> >* On May 8, 2018, at 9:34 AM, ROBIN TORTORA via USRP-users <usrp-users at
>>> >lists.ettus.com
>>> ><http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com>>
>>> >wrote:
>>> *> >* some small things:
>>> *> >* I am pretty sure you cables are not phase matched (costs about 5K for
>>> a 26” matched pair), so you will get some difference there. Not sure how to
>>> quantify.
>>> *>* Splitters have a phase mismatch, I think its called phase unbalance,
>>> proportional to cost :), but can be multiple degrees.
>>> *>* even between 2 channels on the same device, there will be some phase
>>> noise
>>> *>* Still seems far away from 18 degrees, so cant help more than above...
>>> *>
>>> >>* On May 8, 2018 at 11:17 AM "shachar J. brown via USRP-users"
>>> >><usrp-users at lists.ettus.com
>>> >><http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com>>
>>> >>wrote:
>>> *>> >>* Hi Jeff,
>>> *>>* Thanks for your response, but you understood me completely wrong.
>>> *>>* Of course I have set the RF freq in the source block. Furthermore, I
>>> have fully analized the signal from both chanels, and it appears crystal
>>> clear in all the sinks throughout the flow graph (e.g., a pure max at the
>>> correct bin after the FFT in the vector sink, and a beautiful phase gain in
>>> the time sink after the phase extraction).
>>> *>>* My problem isn't receiving the signal or analizing it. My problem is
>>> that the phase difference between the two channels does not match the
>>> theory. The wire to one channel is longer than the other by at least 1/6 of
>>> a wavelength, whilst the phase diff was only 1/10 of 2*pi.
>>> *>>* Am I understood?
>>> *>>* Does anyone have a clue what's going wrong?
>>> *>>* Thanks again,
>>> *>>* Steve
>>> *>> >> >>* On 05/07/2018 11:11 AM, shachar J. brown via USRP-users wrote:
>>> *>>* > Hi All,
>>> *>>* >
>>> *>>* > I am trying to measure the phase difference of a received signal
>>> between
>>> *>>* > the two RX ports of the B210. (In the grc I simply ran each of the
>>> two
>>> *>>* > received signals through an FFT, took the bin with highest amplitude
>>> and
>>> *>>* > extracted it's phase, and finaly - subtracted the two).
>>> *>>* >
>>> *>>* > I experimented with a single signal source generator split in two. I
>>> *>>* > first connected both RX ports with matching wires and received zero
>>> *>>* > phase difference as expected.
>>> *>>* >
>>> *>>* > Though when I added a wire of some length to one of the ports, the
>>> *>>* > received phase difference was NOT as expected by theory.
>>> *>>* >
>>> *>>* > (In short, I sent 100[Mhz] pure sine wave, thus wavelength of 3[m] or
>>> *>>* > shorter, the extra wire was 0.508[cm] long, thus I would expect a
>>> phase
>>> *>>* > diff of about 60 deg or more. Frankly I received a phase diff of
>>> about
>>> *>>* > 18 deg).
>>> *>>* >
>>> *>>* > What am I doing wrong?
>>> *>>* >
>>> *>>* > I thought maybe the phase calculation of gnuradio is done on the
>>> *>>* > baseband frequency and not on the RF, and therefore the phase diff
>>> would
>>> *>>* > be different. Is this my problem? (e.g. if the baseband is only
>>> 30[Mhz],
>>> *>>* > then expected phase diff would be 18 deg). If that is the case - how
>>> can
>>> *>>* > I know which baseband frequency the AD9361 has chosen?
>>> *>> >>* GNU Radio works with signals at baseband. It sounds like you might
>>> not
>>> *>>* have set a RF frequency in the USRP source block. The default is 0, and
>>> *>>* I'm not sure what the B210 would tune to in that case.
>>> *>> >>* I don't know whether this experiment actually works, but to do it
>>> you
>>> *>>* would tune to 99M, set the sample rate to 2M, and see what happens. The
>>> *>>* peak should be right in the middle.
>>> *>> >>* Also, make sure your signal generator is sending out a very low
>>> signal -
>>> *>>* try something like -40dBm. Max is higher, but there's no need.
>>> *>> >>* >
>>> *>>* > Thank you all for your time,
>>> *>>* > Steve
>>> *>>* >
>>> *>>* >
>>> *>> >>* _______________________________________________*
>>>
>>>
>>>
>>> _______________________________________________
>>> USRP-users mailing list
>>> USRP-users@lists.ettus.com
>>> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>>>
>>>
>>
>>
>> _______________________________________________
>> USRP-users mailing
>> listUSRP-users@lists.ettus.comhttp://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>>
>>
>>
>> _______________________________________________
>> USRP-users mailing list
>> USRP-users@lists.ettus.com
>> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>>
>>
>
_______________________________________________
USRP-users mailing list
USRP-users@lists.ettus.com
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
