[USRP-users] LO phase alignment
christophe.alexandre at cnam.fr
Fri Jun 22 16:12:53 EDT 2012
i need to measure a phase difference between two photodiodes
very precisely. I think it's possible with with 2 N210, but i'm
used to use it for radio application under Matlab and not for
I've already designed a telemetry system for very precise distance
measurement (better than 10^-10), but directly with an ML605 +
FMC150 + external mixers and a single LO, the same for emission
This time, those guys asked me to realize the electronic part
of a laser telemeter, optimized for low cost and 10^-7 accuracy
(the effect of air temperature will be solved later using 2 lasers
with different wavelength, so we will use 2 telemeters in one system).
the basic principle is the following :
the TX laser is amplitude modulated with a 3.00005 GHz carrier.
on both photodiodes, RX signals are demodulated with a 3 GHz LO and the phase
difference between both 50 kHz sinus is measured.
the distance measurement is done in 2 steps :
1) put an optical reflector at distance L0. this is the 0 reference.
2) put an optical reflector at distance Lx. this is the unkown distance.
the phase difference doesn't give you directly the distance L because
at 3 GHz, the wavelength is about 10 cm in the air. So the distance is equivalent
to a difference 2.k.pi + deltaphi, and you only get deltaphi. One elegant solution
to get k is to find f1 where deltaphi = 0, increase freq and find the next deltaphi = 0
that will give you f2. then L=c/(f2-f1).
the principle is easy to understand, the optical part is not very complicated
but a 10^-7 accuracy in not so easy to get.
when the optical guys told me about what they want, i immediatly thought
about a mimo receiver (i'm also working on antenna processing algorithms)
but with phase alignment.
i tried to use 2 N210 linked with a MIMO cable, the master being external referenced
with a rubidium clock. The master generate the carrier and receive the first photodiode
signal, the slave receive the second photodiode signal.
At first, i linked directly TX with RX1 via an attenuator and put an SMA cable between
RX1 and RX2 (quick and dirty, only to get a delay). i calculate phase difference
(i use GRC because Matlab doesn't support external ref nor mimo cable) and put
a wxgui slider on carrier frequency. And of course, when i change freq from f1 to f2
then back to f1, the phase difference is not the same. So i can't use this "as is" to
measure the distance.
>From your explanations, i think it could be done if the phase difference between LOs is
the same because the calibration process is included in the measure but :
1) when you say i will get the same phase offset on each LO, i wonder about
the precision (what do you call the same offset ?).
2) how fast is the retuning ? regarding this point, i think tuning offset can help a lot.
i can find coarsely deltaphi = 0 then refine with the DSP offset.
i hope the process is more clear now.
Do you think the principle is correct ?
Conservatoire National des Arts et Métiers (CNAM)
Accès 17-1-32, Case 2D2P10
292 rue Saint Martin
75141 PARIS CEDEX 03
email : christophe.alexandre at cnam.fr
----- Original Message -----
From: John Malsbury
To: usrp-users at lists.ettus.com
Sent: Friday, June 22, 2012 8:11 PM
Subject: Re: [USRP-users] LO phase alignment
If I might chime in...
If you use the SBX retune feature, this will align all of the phase LOs. However, variations in other parts of the signal chain (mixers, filters, etc) will cause slight phase differences across the channels. So, you may still need to some level of calibration of channel-to-channel phase if you need them very closely aligned. The phase differences between the channels will change across temperature and frequency, but not with each re-tune if you use these timed commands. Calibration requirements are ultimately driven by the channel-to-channel phase and amplitude requirements of your application.
Chris, can you explain a little about how this system works? Are you doing ranging and comms with this system? I've seen similar, multi-channel optical systems with other customers.
On Fri, Jun 22, 2012 at 10:49 AM, Matt Ettus <matt at ettus.com> wrote:
To clarify here, as long as all LOs are tuned at the same time, you
will always get the same phase offset. If you retune one, you have to
retune the others simultaneously.
On Fri, Jun 22, 2012 at 1:58 AM, Christophe ALEXANDRE
<christophe.alexandre at cnam.fr> wrote:
> Hi Josh,
> i've read the link :
> Using timed commands, multiple frontends can be tuned at a specific time.
> This timed-tuning ensures that the phase offsets between VCO/PLL chains will
> remain constant after each re-tune. See notes below:
> a.. There is a random phase offset between any two frontends
> b.. This phase offset is different for different LO frequencies
> c.. This phase offset remains constant after retuning
> a.. Due to divider, WBX phase offset will be randomly +/- 180 deg after
> d.. This phase offset will drift over time due to thermal and other
> e.. Periodic calibration will be necessary for phase-coherent applications
> I still don't understand the principle you use. is it a VCO special property
> some questions still remain :
> Q1) if i change LO freq from f1 to f2 then back to f1, should i get the same
> phase offset on f1 ?
> Q2) if i turn off the N210s then on, should i get the same offset on the
> same freq ?
> Q3) if the Q1, Q2 answers are yes, what kind of phase offset precision
> should i expect ?
> Christophe ALEXANDRE
> Conservatoire National des Arts et Métiers (CNAM)
> Laboratoire CEDRIC-LAETITIA
> Département EASY
> Accès 17-1-32, Case 2D2P10
> 292 rue Saint Martin
> 75141 PARIS CEDEX 03
> email : christophe.alexandre at cnam.fr
> tel. 0140272699
> mob. 0651087311
> fax. 0140272994
> ----- Original Message ----- From: "Josh Blum" <josh at ettus.com>
> To: <usrp-users at lists.ettus.com>
> Sent: Thursday, June 21, 2012 5:53 PM
> Subject: Re: [USRP-users] LO phase alignment
>> On 06/21/2012 08:48 AM, Christophe ALEXANDRE wrote:
>>> hi John,
>>> i have already read this app note (in fact, i have started with it).
>>> I've only seen a star after phase on table 2 for SBX but no more info.
>>> what is the basic principle used to solve the phase alignment problem ?
>> Does this example help:
>>> can i use this new feature using gnuradio ?
>>> do you have any example (even C++ example) showing how to use it ?
>> Link above.
>> USRP-users mailing list
>> USRP-users at lists.ettus.com
> USRP-users mailing list
> USRP-users at lists.ettus.com
USRP-users mailing list
USRP-users at lists.ettus.com
USRP-users mailing list
USRP-users at lists.ettus.com
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the USRP-users