usrp-users@lists.ettus.com

Discussion and technical support related to USRP, UHD, RFNoC

View all threads

Phase compensate issue

HY
Hojoon Yang
Sun, Nov 12, 2017 4:44 PM

Hi guys!please let me know whether following operation is possible!we have USRP B200.All I want to do is to transmit a specific signal without distorting phase.Say, I want to transmit the cosine wave, which have center frequency fc = 3Mhz and phase is 60 degree(i.e. y = cos(2pi3e6*t + pi/3) ) through the Tx/Rx channel in USRP B200.But, I heard that there is a consistent phase offset after tuning. Thus, when we transmit the cosine signal as it is, it will have phase offset. I dont care the amplitude, but only care the exact phase.In order to compensate the phase offset, we are going to receive the signal which is transmitted from the Tx/Rx channel since we don't know the phase offset before it transmits. If we receive the signal, then we can know the phase offset 'theta'. since we know the phase offset theta, to compensate the phase offset, We multiply that signal by -theta.In summary,(a) USRP B210 transmit the signal through the Tx/Rx channel, which will have phase offset, say X.(b) the usrp receive the (a) signal to know the phase offset X through RX2 channel(i.e. A loopback using a single USRP.)(c) multiply the signal samples by -X  to compensate the phase and transmit!1. Will it be possible??2. Marcus said that The SBX has a special architecture based on a LO synthesizer that allows to have a deterministic phase after "tuning"What is the exact meaning of tuning? it means frequency lock?

Hi guys!please let me know whether following operation is possible!we have USRP B200.All I want to do is to transmit a specific signal without distorting phase.Say, I want to transmit the cosine wave, which have center frequency fc = 3Mhz and phase is 60 degree(i.e. y = cos(2*pi*3e6*t + pi/3) ) through the Tx/Rx channel in USRP B200.But, I heard that there is a consistent phase offset after tuning. Thus, when we transmit the cosine signal as it is, it will have phase offset. I dont care the amplitude, but only care the exact phase.In order to compensate the phase offset, we are going to receive the signal which is transmitted from the Tx/Rx channel since we don't know the phase offset before it transmits. If we receive the signal, then we can know the phase offset 'theta'. since we know the phase offset theta, to compensate the phase offset, We multiply that signal by -theta.In summary,(a) USRP B210 transmit the signal through the Tx/Rx channel, which will have phase offset, say X.(b) the usrp receive the (a) signal to know the phase offset X through RX2 channel(i.e. A loopback using a single USRP.)(c) multiply the signal samples by -X  to compensate the phase and transmit!1. Will it be possible??2. Marcus said that The SBX has a special architecture based on a LO synthesizer that allows to have a deterministic phase after "tuning"What is the exact meaning of tuning? it means frequency lock?
MM
Marcus Müller
Sun, Nov 12, 2017 5:19 PM

Dear Hojoon,

the B200 can't go that far down in frequency, so 3 MHz is impossible.

But let's say you want to do 300 MHz.

You say "phase is 60°", but phase is always relative to something. So,
your question only makes sense in conjuction with another wave or
theoretical source of time.

But, I heard that there is a consistent phase offset after tuning.

That isn't true for the B200. Here, the phase after tuning the LOs is
random.

Thus, when we transmit the cosine signal as it is, it will have phase offset. I dont

care the amplitude, but only care the exact phase.

What does "the same phase" even mean for two tones of different
frequencies? This makes no sense to me!

In order to compensate the phase offset, we are going to receive the
signal which is transmitted from the Tx/Rx channel since we don't know
the phase offset before it transmits. If we receive the signal, then
we can know the phase offset 'theta'. since we know the phase offset
theta, to compensate the phase offset, We multiply that signal by -theta.

You will need to tune your receiver too, since receive and send
bandwidths are the same, so without more work (you could implement
something like a staggered alternating TX and RX tuning, but that
doesn't sound like a feasible solution to any problem I can think of),
this is not a solution to a problem you could encounter.

In summary,

(a) USRP B210 transmit the signal through the Tx/Rx channel, which
will have phase offset, say X.

(b) the usrp receive the (a) signal to know the phase offset X through
RX2 channel(i.e. Aloopback using a single USRP.)

If you add another USRP to the mix, you add, again, phase uncertainty!

(c) multiply the signal samples by -X  to compensate the phase and
transmit!

You mean exp(-jX), right?

  1. Will it be possible??

Yes, I think so, but as explained above, I'm not sure what "compensated
phase" even is if you have two different frequencies.
And, your receiver would need to be able to phase-coherently sample both
your old and your new frequency – but since you had to tune TX, you
would also have to tune RX, so, no, I don't think this approach works.

  1. Marcus said that The SBX has a special architecture based on a LO
    synthesizer that allows to have a deterministic phase after "tuning"

What is the exact meaning of tuning? it means frequency lock?

Oh, ok. Here, tuning is the act of changing the mixing frequency of a
receiver or transmitter.

Best regards,
Marcus

On 12.11.2017 17:44, Hojoon Yang via USRP-users wrote:

Hi guys!

please let me know whether following operation is possible!

we have USRP B200.

All I want to do is to transmit a specific signal without distorting
phase.

Say, I want to transmit the cosine wave, which have center frequency
fc = 3Mhz and phase is 60 degree(i.e. y = cos(2pi3e6*t + pi/3) )
through the Tx/Rx channel in USRP B200.

But, I heard that there is a consistent phase offset after tuning.
Thus, when we transmit the cosine signal as it is, it will have phase
offset. I dont care the amplitude, but only care the exact phase.

In order to compensate the phase offset, we are going to receive the
signal which is transmitted from the Tx/Rx channel since we don't know
the phase offset before it transmits. If we receive the signal, then
we can know the phase offset 'theta'. since we know the phase offset
theta, to compensate the phase offset, We multiply that signal by -theta.

In summary,

(a) USRP B210 transmit the signal through the Tx/Rx channel, which
will have phase offset, say X.

(b) the usrp receive the (a) signal to know the phase offset X through
RX2 channel(i.e. Aloopback using a single USRP.)

(c) multiply the signal samples by -X  to compensate the phase and
transmit!

  1. Will it be possible??

  2. Marcus said that The SBX has a special architecture based on a LO
    synthesizer that allows to have a deterministic phase after "tuning"

What is the exact meaning of tuning? it means frequency lock?


USRP-users mailing list
USRP-users@lists.ettus.com
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com

Dear Hojoon, the B200 can't go that far down in frequency, so 3 MHz is impossible. But let's say you want to do 300 MHz. You say "phase is 60°", but phase is always *relative* to something. So, your question only makes sense in conjuction with another wave or theoretical source of time. > But, I heard that there is a consistent phase offset after tuning. That isn't true for the B200. Here, the phase after tuning the LOs is random. > Thus, when we transmit the cosine signal as it is, it will have phase offset. I dont care the amplitude, but only care the exact phase. What does "the same phase" even mean for two tones of different frequencies? This makes no sense to me! > In order to compensate the phase offset, we are going to receive the > signal which is transmitted from the Tx/Rx channel since we don't know > the phase offset before it transmits. If we receive the signal, then > we can know the phase offset 'theta'. since we know the phase offset > theta, to compensate the phase offset, We multiply that signal by -theta. > You will need to tune your receiver too, since receive and send bandwidths are the same, so without more work (you could implement something like a staggered alternating TX and RX tuning, but that doesn't sound like a feasible solution to any problem I can think of), this is not a solution to a problem you could encounter. > > In summary, > > > (a) USRP B210 transmit the signal through the Tx/Rx channel, which > will have phase offset, say X. > > (b) the usrp receive the (a) signal to know the phase offset X through > RX2 channel(i.e. Aloopback using a single USRP.) > If you add another USRP to the mix, you add, again, phase uncertainty! > > (c) multiply the signal samples by -X to compensate the phase and > transmit! > You mean exp(-jX), right? > 1. Will it be possible?? Yes, I think so, but as explained above, I'm not sure what "compensated phase" even is if you have two different frequencies. And, your receiver would need to be able to phase-coherently sample both your old and your new frequency – but since you had to tune TX, you would also have to tune RX, so, no, I don't think this approach works. > 2. Marcus said that The SBX has a special architecture based on a LO > synthesizer that allows to have a deterministic phase after "tuning" > > What is the exact meaning of tuning? it means frequency lock? > Oh, ok. Here, tuning is the act of changing the mixing frequency of a receiver or transmitter. Best regards, Marcus On 12.11.2017 17:44, Hojoon Yang via USRP-users wrote: > > Hi guys! > > > > please let me know whether following operation is possible! > > > > we have USRP B200. > > > > All I want to do is to transmit a specific signal without distorting > phase. > > > > Say, I want to transmit the cosine wave, which have center frequency > fc = 3Mhz and phase is 60 degree(i.e. y = cos(2*pi*3e6*t + pi/3) ) > through the Tx/Rx channel in USRP B200. > > > > But, I heard that there is a consistent phase offset after tuning. > Thus, when we transmit the cosine signal as it is, it will have phase > offset. I dont care the amplitude, but only care the exact phase. > > > In order to compensate the phase offset, we are going to receive the > signal which is transmitted from the Tx/Rx channel since we don't know > the phase offset before it transmits. If we receive the signal, then > we can know the phase offset 'theta'. since we know the phase offset > theta, to compensate the phase offset, We multiply that signal by -theta. > > > In summary, > > > (a) USRP B210 transmit the signal through the Tx/Rx channel, which > will have phase offset, say X. > > (b) the usrp receive the (a) signal to know the phase offset X through > RX2 channel(i.e. Aloopback using a single USRP.) > > (c) multiply the signal samples by -X to compensate the phase and > transmit! > > > 1. Will it be possible?? > > > > 2. Marcus said that The SBX has a special architecture based on a LO > synthesizer that allows to have a deterministic phase after "tuning" > > What is the exact meaning of tuning? it means frequency lock? > > > > _______________________________________________ > USRP-users mailing list > USRP-users@lists.ettus.com > http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com