usrp-users@lists.ettus.com
Discussion and technical support related to USRP, UHD, RFNoC
View all threadsProblem with replicas.
Hello,
This may be an easy question for the experts here. I'm trying to reuse some
code a colleague has done and I encountered a problem he might not noticed
nor corrected.
I want to transmit samples uninterruptedly over the 474 MHz channel. The
code he gave me comes with something like:
usrp->set_rx_freq(uhd::tune_request_t(474e6, 8e6));
A RF analyzer capture shows this
https://dl.dropboxusercontent.com/u/2696878/replicasLO_local.png, where
there are replicas everywhere (see figure
https://dl.dropboxusercontent.com/u/2696878/replicasLO.png with more freq
span). I checked the documentation and arrived at the conclusion that we
shouldn't mess with the LO_offset value in this case so I used instead:
usrp->set_rx_freq(uhd::tune_request_t(474e6));
Now it got better as the local replicas disappeared but not the ones
further away (figure
https://dl.dropboxusercontent.com/u/2696878/replicas_noLO.png). We would
be interested in having only the signal in the center freq and no signal on
its multiples. I'm not sure if there is any way to correct this or it's
just the way it is. Any advice?
(we are also using usrp->set_tx_bandwidth(10e6))
Thank you in advance.
Greetings,
Brais.
--
A little mistake. I meant set_tx_freq in both lines, i.e.:
usrp->set_tx_freq(uhd::tune_request_t(474e6));
2015-01-12 17:26 GMT+01:00 Brais Ares bares@gradiant.org:
Hello,
This may be an easy question for the experts here. I'm trying to reuse
some code a colleague has done and I encountered a problem he might not
noticed nor corrected.
I want to transmit samples uninterruptedly over the 474 MHz channel. The
code he gave me comes with something like:
usrp->set_rx_freq(uhd::tune_request_t(474e6, 8e6));
A RF analyzer capture shows this
https://dl.dropboxusercontent.com/u/2696878/replicasLO_local.png, where
there are replicas everywhere (see figure
https://dl.dropboxusercontent.com/u/2696878/replicasLO.png with more
freq span). I checked the documentation and arrived at the conclusion that
we shouldn't mess with the LO_offset value in this case so I used
instead:
usrp->set_rx_freq(uhd::tune_request_t(474e6));
Now it got better as the local replicas disappeared but not the ones
further away (figure
https://dl.dropboxusercontent.com/u/2696878/replicas_noLO.png). We
would be interested in having only the signal in the center freq and no
signal on its multiples. I'm not sure if there is any way to correct this
or it's just the way it is. Any advice?
(we are also using usrp->set_tx_bandwidth(10e6))
Thank you in advance.
Greetings,
Brais.
--
--
Hello Brais,
what is your TX sampling rate, which USRP are you using, and if
applicable, which daughterboard, at which TX gain?
Generally, if your board actually supports setting an analog bandwidth
(not all daughterboards do), then setting it to 10MHz and tuning 8MHz
from the center would have the effect that your signal might be
attenuated by the analog filters, since the analog bandwidth is
happening symmetrically to the physical LO.
Usually, I would not expect anything of only -30dB relative to your main
lobe to appear in the spectrum; so this is really worth investigating.
Best regards,
Marcus
On 01/13/2015 09:24 AM, Brais Ares via USRP-users wrote:
A little mistake. I meant set_tx_freq in both lines, i.e.:
usrp->set_tx_freq(uhd::tune_request_t(474e6));
2015-01-12 17:26 GMT+01:00 Brais Ares bares@gradiant.org:
Hello,
This may be an easy question for the experts here. I'm trying to reuse
some code a colleague has done and I encountered a problem he might not
noticed nor corrected.
I want to transmit samples uninterruptedly over the 474 MHz channel. The
code he gave me comes with something like:
usrp->set_rx_freq(uhd::tune_request_t(474e6, 8e6));
A RF analyzer capture shows this
https://dl.dropboxusercontent.com/u/2696878/replicasLO_local.png, where
there are replicas everywhere (see figure
https://dl.dropboxusercontent.com/u/2696878/replicasLO.png with more
freq span). I checked the documentation and arrived at the conclusion that
we shouldn't mess with the LO_offset value in this case so I used
instead:
usrp->set_rx_freq(uhd::tune_request_t(474e6));
Now it got better as the local replicas disappeared but not the ones
further away (figure
https://dl.dropboxusercontent.com/u/2696878/replicas_noLO.png). We
would be interested in having only the signal in the center freq and no
signal on its multiples. I'm not sure if there is any way to correct this
or it's just the way it is. Any advice?
(we are also using usrp->set_tx_bandwidth(10e6))
Thank you in advance.
Greetings,
Brais.
--
--
USRP-users mailing list
USRP-users@lists.ettus.com
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
The analyzer seems overdriven. I assume the transmitter is directly
connected to the input, no attenuator, and only an internal attenuation of
10dB. My guess is, attenuate it for 20 or 30 dB, and everything looks fine.
Ralph.
From: USRP-users [mailto:usrp-users-bounces@lists.ettus.com] On Behalf Of
Marcus Müller via USRP-users
Sent: Tuesday, January 13, 2015 11:14 AM
To: usrp-users@lists.ettus.com
Subject: Re: [USRP-users] Problem with replicas.
Hello Brais,
what is your TX sampling rate, which USRP are you using, and if applicable,
which daughterboard, at which TX gain?
Generally, if your board actually supports setting an analog bandwidth (not
all daughterboards do), then setting it to 10MHz and tuning 8MHz from the
center would have the effect that your signal might be attenuated by the
analog filters, since the analog bandwidth is happening symmetrically to the
physical LO.
Usually, I would not expect anything of only -30dB relative to your main
lobe to appear in the spectrum; so this is really worth investigating.
Best regards,
Marcus
On 01/13/2015 09:24 AM, Brais Ares via USRP-users wrote:
A little mistake. I meant set_tx_freq in both lines, i.e.:
usrp->set_tx_freq(uhd::tune_request_t(474e6));
2015-01-12 17:26 GMT+01:00 Brais Ares mailto:bares@gradiant.org
bares@gradiant.org:
Hello,
This may be an easy question for the experts here. I'm trying to reuse
some code a colleague has done and I encountered a problem he might not
noticed nor corrected.
I want to transmit samples uninterruptedly over the 474 MHz channel. The
code he gave me comes with something like:
usrp->set_rx_freq(uhd::tune_request_t(474e6, 8e6));
A RF analyzer capture shows this
https://dl.dropboxusercontent.com/u/2696878/replicasLO_local.png
https://dl.dropboxusercontent.com/u/2696878/replicasLO_local.png, where
there are replicas everywhere (see figure
https://dl.dropboxusercontent.com/u/2696878/replicasLO.png
https://dl.dropboxusercontent.com/u/2696878/replicasLO.png with more
freq span). I checked the documentation and arrived at the conclusion that
we shouldn't mess with the LO_offset value in this case so I used
instead:
usrp->set_rx_freq(uhd::tune_request_t(474e6));
Now it got better as the local replicas disappeared but not the ones
further away (figure
https://dl.dropboxusercontent.com/u/2696878/replicas_noLO.png
https://dl.dropboxusercontent.com/u/2696878/replicas_noLO.png). We
would be interested in having only the signal in the center freq and no
signal on its multiples. I'm not sure if there is any way to correct this
or it's just the way it is. Any advice?
(we are also using usrp->set_tx_bandwidth(10e6))
Thank you in advance.
Greetings,
Brais.
--
--
USRP-users mailing list
USRP-users@lists.ettus.com mailto:USRP-users@lists.ettus.com
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
Hello,
The same effect happened when using the next uhd example (see capture
https://dl.dropboxusercontent.com/u/2696878/uhdexample_replicas.png):
./tx_bursts --rate=8000000 --freq=474000000 --gain=0 --nsamps=80000000
The device is a USRP E110 with a SBX Board Rev 3 daughterboard.
My first though was also the analyzer, but when it gets overdriven it
usually shows a warning message on the screen and that's not happening in
this case. Anyway you may be right so I'm going to give it a try with
another analyzer tomorrow. Attenuating 20-30 dB would solve the replicas
but, can't I generate a clean signal without, let's say, 'cheating'? what
if I need a 0 or more dBm signal?
Thank you once again,
Greetings,
Brais.
2015-01-13 15:45 GMT+01:00 Ralph A. Schmid, dk5ras via USRP-users <
usrp-users@lists.ettus.com>:
The analyzer seems overdriven. I assume the transmitter is directly
connected to the input, no attenuator, and only an internal attenuation of
10dB. My guess is, attenuate it for 20 or 30 dB, and everything looks fine.
Ralph.
From: USRP-users [mailto:usrp-users-bounces@lists.ettus.com] *On Behalf
Of *Marcus Müller via USRP-users
Sent: Tuesday, January 13, 2015 11:14 AM
To: usrp-users@lists.ettus.com
Subject: Re: [USRP-users] Problem with replicas.
Hello Brais,
what is your TX sampling rate, which USRP are you using, and if
applicable, which daughterboard, at which TX gain?
Generally, if your board actually supports setting an analog bandwidth
(not all daughterboards do), then setting it to 10MHz and tuning 8MHz from
the center would have the effect that your signal might be attenuated by
the analog filters, since the analog bandwidth is happening symmetrically
to the physical LO.
Usually, I would not expect anything of only -30dB relative to your main
lobe to appear in the spectrum; so this is really worth investigating.
Best regards,
Marcus
On 01/13/2015 09:24 AM, Brais Ares via USRP-users wrote:
A little mistake. I meant set_tx_freq in both lines, i.e.:
usrp->set_tx_freq(uhd::tune_request_t(474e6));
2015-01-12 17:26 GMT+01:00 Brais Ares bares@gradiant.org bares@gradiant.org:
Hello,
This may be an easy question for the experts here. I'm trying to reuse
some code a colleague has done and I encountered a problem he might not
noticed nor corrected.
I want to transmit samples uninterruptedly over the 474 MHz channel. The
code he gave me comes with something like:
usrp->set_rx_freq(uhd::tune_request_t(474e6, 8e6));
A RF analyzer capture shows this
https://dl.dropboxusercontent.com/u/2696878/replicasLO_local.png https://dl.dropboxusercontent.com/u/2696878/replicasLO_local.png, where
there are replicas everywhere (see figure
https://dl.dropboxusercontent.com/u/2696878/replicasLO.png https://dl.dropboxusercontent.com/u/2696878/replicasLO.png with more
freq span). I checked the documentation and arrived at the conclusion that
we shouldn't mess with the LO_offset value in this case so I used
instead:
usrp->set_rx_freq(uhd::tune_request_t(474e6));
Now it got better as the local replicas disappeared but not the ones
further away (figure
https://dl.dropboxusercontent.com/u/2696878/replicas_noLO.png https://dl.dropboxusercontent.com/u/2696878/replicas_noLO.png). We
would be interested in having only the signal in the center freq and no
signal on its multiples. I'm not sure if there is any way to correct this
or it's just the way it is. Any advice?
(we are also using usrp->set_tx_bandwidth(10e6))
Thank you in advance.
Greetings,
Brais.
--
--
USRP-users mailing list
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
--
Usually you can produce a clean signal with 0dBm without problems; I assume that the analyzer shows something that is not there, and the warning only is shown when the analyzer fears for its live from overload. There is no warning at a non-lethal overload that just produces stuff that is not there J Set the analyzers attenuation to a higher value and see what is happening, or add an external 30dB attenuator, this leaves your signal untouched. Possibly your signal is just fine, and the images are produced in the input stage of the analyzer.
Ralph.
From: Brais Ares [mailto:bares@gradiant.org]
Sent: Tuesday, January 13, 2015 16:52
To: Ralph A. Schmid, dk5ras
Cc: Marcus Müller; usrp-users
Subject: Re: [USRP-users] Problem with replicas.
Hello,
The same effect happened when using the next uhd example (see capture https://dl.dropboxusercontent.com/u/2696878/uhdexample_replicas.png ):
./tx_bursts --rate=8000000 --freq=474000000 --gain=0 --nsamps=80000000
The device is a USRP E110 with a SBX Board Rev 3 daughterboard.
My first though was also the analyzer, but when it gets overdriven it usually shows a warning message on the screen and that's not happening in this case. Anyway you may be right so I'm going to give it a try with another analyzer tomorrow. Attenuating 20-30 dB would solve the replicas but, can't I generate a clean signal without, let's say, 'cheating'? what if I need a 0 or more dBm signal?
Thank you once again,
Greetings,
Brais.
2015-01-13 15:45 GMT+01:00 Ralph A. Schmid, dk5ras via USRP-users usrp-users@lists.ettus.com:
The analyzer seems overdriven. I assume the transmitter is directly connected to the input, no attenuator, and only an internal attenuation of 10dB. My guess is, attenuate it for 20 or 30 dB, and everything looks fine.
Ralph.
From: USRP-users [mailto:usrp-users-bounces@lists.ettus.com] On Behalf Of Marcus Müller via USRP-users
Sent: Tuesday, January 13, 2015 11:14 AM
To: usrp-users@lists.ettus.com
Subject: Re: [USRP-users] Problem with replicas.
Hello Brais,
what is your TX sampling rate, which USRP are you using, and if applicable, which daughterboard, at which TX gain?
Generally, if your board actually supports setting an analog bandwidth (not all daughterboards do), then setting it to 10MHz and tuning 8MHz from the center would have the effect that your signal might be attenuated by the analog filters, since the analog bandwidth is happening symmetrically to the physical LO.
Usually, I would not expect anything of only -30dB relative to your main lobe to appear in the spectrum; so this is really worth investigating.
Best regards,
Marcus
On 01/13/2015 09:24 AM, Brais Ares via USRP-users wrote:
A little mistake. I meant set_tx_freq in both lines, i.e.:
usrp->set_tx_freq(uhd::tune_request_t(474e6));
2015-01-12 17:26 GMT+01:00 Brais Ares mailto:bares@gradiant.org bares@gradiant.org:
Hello,
This may be an easy question for the experts here. I'm trying to reuse
some code a colleague has done and I encountered a problem he might not
noticed nor corrected.
I want to transmit samples uninterruptedly over the 474 MHz channel. The
code he gave me comes with something like:
usrp->set_rx_freq(uhd::tune_request_t(474e6, 8e6));
A RF analyzer capture shows this
https://dl.dropboxusercontent.com/u/2696878/replicasLO_local.png https://dl.dropboxusercontent.com/u/2696878/replicasLO_local.png, where
there are replicas everywhere (see figure
https://dl.dropboxusercontent.com/u/2696878/replicasLO.png https://dl.dropboxusercontent.com/u/2696878/replicasLO.png with more
freq span). I checked the documentation and arrived at the conclusion that
we shouldn't mess with the LO_offset value in this case so I used
instead:
usrp->set_rx_freq(uhd::tune_request_t(474e6));
Now it got better as the local replicas disappeared but not the ones
further away (figure
https://dl.dropboxusercontent.com/u/2696878/replicas_noLO.png https://dl.dropboxusercontent.com/u/2696878/replicas_noLO.png). We
would be interested in having only the signal in the center freq and no
signal on its multiples. I'm not sure if there is any way to correct this
or it's just the way it is. Any advice?
(we are also using usrp->set_tx_bandwidth(10e6))
Thank you in advance.
Greetings,
Brais.
--
--
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
--
https://dl.dropboxusercontent.com/u/2696878/gradianfirma.jpg
Hi Brais,
Attenuating 20-30 dB would solve the replicas but, can't I generate a clean signal without, let's say, 'cheating'?
well, attenuation to protect a measurement device surely doesn't sound
like cheating :)
I think there might be a misunderstanding expressed in your sentence:
If the images (spurious peaks) go away by inserting attenuators, than
the signal emitted by the USRP is correct, but the fact that amplifiers
in the analyzer are driven outside their linear regions introduces
quadratic (and higher order terms) in their time-domain transfer
function, leading to a situation where the signal entering the device is
modulated with itself, yielding signals at multiples of the input
frequency. Thus, if attenuation is really solving the problem, you're
not really having one, and the signal is fine, but the analyzer's
analysis is wrong.
Greetings,
Marcus
On 01/13/2015 04:51 PM, Brais Ares wrote:
Hello,
The same effect happened when using the next uhd example (see capture
https://dl.dropboxusercontent.com/u/2696878/uhdexample_replicas.png):
./tx_bursts --rate=8000000 --freq=474000000 --gain=0 --nsamps=80000000
The device is a USRP E110 with a SBX Board Rev 3 daughterboard.
My first though was also the analyzer, but when it gets overdriven it
usually shows a warning message on the screen and that's not happening in
this case. Anyway you may be right so I'm going to give it a try with
another analyzer tomorrow. Attenuating 20-30 dB would solve the replicas
but, can't I generate a clean signal without, let's say, 'cheating'? what
if I need a 0 or more dBm signal?
Thank you once again,
Greetings,
Brais.
2015-01-13 15:45 GMT+01:00 Ralph A. Schmid, dk5ras via USRP-users <
usrp-users@lists.ettus.com>:
The analyzer seems overdriven. I assume the transmitter is directly
connected to the input, no attenuator, and only an internal attenuation of
10dB. My guess is, attenuate it for 20 or 30 dB, and everything looks fine.
Ralph.
From: USRP-users [mailto:usrp-users-bounces@lists.ettus.com] *On Behalf
Of *Marcus Müller via USRP-users
Sent: Tuesday, January 13, 2015 11:14 AM
To: usrp-users@lists.ettus.com
Subject: Re: [USRP-users] Problem with replicas.
Hello Brais,
what is your TX sampling rate, which USRP are you using, and if
applicable, which daughterboard, at which TX gain?
Generally, if your board actually supports setting an analog bandwidth
(not all daughterboards do), then setting it to 10MHz and tuning 8MHz from
the center would have the effect that your signal might be attenuated by
the analog filters, since the analog bandwidth is happening symmetrically
to the physical LO.
Usually, I would not expect anything of only -30dB relative to your main
lobe to appear in the spectrum; so this is really worth investigating.
Best regards,
Marcus
On 01/13/2015 09:24 AM, Brais Ares via USRP-users wrote:
A little mistake. I meant set_tx_freq in both lines, i.e.:
usrp->set_tx_freq(uhd::tune_request_t(474e6));
2015-01-12 17:26 GMT+01:00 Brais Ares bares@gradiant.org bares@gradiant.org:
Hello,
This may be an easy question for the experts here. I'm trying to reuse
some code a colleague has done and I encountered a problem he might not
noticed nor corrected.
I want to transmit samples uninterruptedly over the 474 MHz channel. The
code he gave me comes with something like:
usrp->set_rx_freq(uhd::tune_request_t(474e6, 8e6));
A RF analyzer capture shows this
https://dl.dropboxusercontent.com/u/2696878/replicasLO_local.png https://dl.dropboxusercontent.com/u/2696878/replicasLO_local.png, where
there are replicas everywhere (see figure
https://dl.dropboxusercontent.com/u/2696878/replicasLO.png https://dl.dropboxusercontent.com/u/2696878/replicasLO.png with more
freq span). I checked the documentation and arrived at the conclusion that
we shouldn't mess with the LO_offset value in this case so I used
instead:
usrp->set_rx_freq(uhd::tune_request_t(474e6));
Now it got better as the local replicas disappeared but not the ones
further away (figure
https://dl.dropboxusercontent.com/u/2696878/replicas_noLO.png https://dl.dropboxusercontent.com/u/2696878/replicas_noLO.png). We
would be interested in having only the signal in the center freq and no
signal on its multiples. I'm not sure if there is any way to correct this
or it's just the way it is. Any advice?
(we are also using usrp->set_tx_bandwidth(10e6))
Thank you in advance.
Greetings,
Brais.
--
--
USRP-users mailing list
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
--
There will always be some harmonic content in the output--it's a
natural consequence of what mixers do for a living.
In purpose-built radios, there's nearly always a post-mixer filter to
eliminate harmonics, but in purpose-built radios, the operating
bandwidth is quite small.
The SBX operates over a 10:1 frequency range--there's no way to
(economically) provide switchable filter banks to eliminate the
inevitable mixer harmonics, so "RF hygiene" at that level must
necessarily be an and-user problem.
The E300, does have switchable filter-banks to reduce this problem,
but it's hard for them to be perfect for every application, because
every application is different.
Having to put in filtering is very often a requirement in commercial
systems deployments, to keep out-of-band spurs and harmonics out of the
final RF amplifier stages before the antenna. It's often necessary to do
this on the output side of the amplifier, because the amplifer itself
can introduce harmonics, depending on what mode it's operating in.
On 2015-01-13 10:51, Brais Ares via USRP-users wrote:
Hello,
The same effect happened when using the next uhd example (see capture [5]):
./tx_bursts --rate=8000000 --freq=474000000 --gain=0 --nsamps=80000000
The device is a USRP E110 with a SBX Board Rev 3 daughterboard.
My first though was also the analyzer, but when it gets overdriven it usually shows a warning message on the screen and that's not happening in this case. Anyway you may be right so I'm going to give it a try with another analyzer tomorrow. Attenuating 20-30 dB would solve the replicas but, can't I generate a clean signal without, let's say, 'cheating'? what if I need a 0 or more dBm signal?
Thank you once again,
Greetings, Brais.
2015-01-13 15:45 GMT+01:00 Ralph A. Schmid, dk5ras via USRP-users usrp-users@lists.ettus.com:
The analyzer seems overdriven. I assume the transmitter is directly connected to the input, no attenuator, and only an internal attenuation of 10dB. My guess is, attenuate it for 20 or 30 dB, and everything looks fine.
Ralph.
FROM: USRP-users [mailto:usrp-users-bounces@lists.ettus.com] ON BEHALF OF Marcus Müller via USRP-users
SENT: Tuesday, January 13, 2015 11:14 AM
TO: usrp-users@lists.ettus.com
SUBJECT: Re: [USRP-users] Problem with replicas.
Hello Brais,
what is your TX sampling rate, which USRP are you using, and if applicable, which daughterboard, at which TX gain?
Generally, if your board actually supports setting an analog bandwidth (not all daughterboards do), then setting it to 10MHz and tuning 8MHz from the center would have the effect that your signal might be attenuated by the analog filters, since the analog bandwidth is happening symmetrically to the physical LO.
Usually, I would not expect anything of only -30dB relative to your main lobe to appear in the spectrum; so this is really worth investigating.
Best regards,
Marcus
On 01/13/2015 09:24 AM, Brais Ares via USRP-users wrote:
A little mistake. I meant set_tx_freq in both lines, i.e.:
usrp->set_tx_freq(uhd::tune_request_t(474e6));
2015-01-12 17:26 GMT+01:00 Brais Ares bares@gradiant.org:
Hello,
This may be an easy question for the experts here. I'm trying to reuse
some code a colleague has done and I encountered a problem he might not
noticed nor corrected.
I want to transmit samples uninterruptedly over the 474 MHz channel. The
code he gave me comes with something like:
usrp->set_rx_freq(uhd::tune_request_t(474e6, 8e6));
A RF analyzer capture shows this
https://dl.dropboxusercontent.com/u/2696878/replicasLO_local.png [1], where
there are replicas everywhere (see figure
https://dl.dropboxusercontent.com/u/2696878/replicasLO.png [2] with more
freq span). I checked the documentation and arrived at the conclusion that
we shouldn't mess with the LO_offset value in this case so I used
instead:
usrp->set_rx_freq(uhd::tune_request_t(474e6));
Now it got better as the local replicas disappeared but not the ones
further away (figure
https://dl.dropboxusercontent.com/u/2696878/replicas_noLO.png [3]). We
would be interested in having only the signal in the center freq and no
signal on its multiples. I'm not sure if there is any way to correct this
or it's just the way it is. Any advice?
(we are also using usrp->set_tx_bandwidth(10e6))
Thank you in advance.
Greetings,
Brais.
--
--
USRP-users mailing list
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com [4]
USRP-users mailing list
USRP-users@lists.ettus.com
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com [4]
--
USRP-users mailing list
USRP-users@lists.ettus.com
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com [4]
Links:
[1] https://dl.dropboxusercontent.com/u/2696878/replicasLO_local.png
[2] https://dl.dropboxusercontent.com/u/2696878/replicasLO.png
[3] https://dl.dropboxusercontent.com/u/2696878/replicas_noLO.png
[4] http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
[5] https://dl.dropboxusercontent.com/u/2696878/uhdexample_replicas.png
Hello Brais,
Look for the post "WBX daughterboard - transmit LO signal harmonics" for
example here:
http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/2013-July/007064.html
and the response from Matt Ettus. Harmonics at odd multiplies of carrier
frequency are typical for wideband transmitters sold by Ettus (those
which don't have adjustable filters at the output).
I don't know where the rest of harmonics that you observe come from.
Maybe there is saturation at some point in your transmitter or
measurement equipment.
--
Best Regards,
Piotr
W dniu 12.01.2015 o 17:26, Brais Ares via USRP-users pisze:
Hello,
This may be an easy question for the experts here. I'm trying to reuse
some code a colleague has done and I encountered a problem he might
not noticed nor corrected.
I want to transmit samples uninterruptedly over the 474 MHz channel.
The code he gave me comes with something like:
usrp->set_rx_freq(uhd::tune_request_t(474e6, 8e6));
A RF analyzer capture shows this
https://dl.dropboxusercontent.com/u/2696878/replicasLO_local.png,
where there are replicas everywhere (see figure
https://dl.dropboxusercontent.com/u/2696878/replicasLO.png with more
freq span). I checked the documentation and arrived at the conclusion
that we shouldn't mess with the /LO_offset/ value in this case so I
used instead:
usrp->set_rx_freq(uhd::tune_request_t(474e6));
Now it got better as the local replicas disappeared but not the ones
further away (figure
https://dl.dropboxusercontent.com/u/2696878/replicas_noLO.png). We
would be interested in having only the signal in the center freq and
no signal on its multiples. I'm not sure if there is any way to
correct this or it's just the way it is. Any advice?
(we are also using usrp->set_tx_bandwidth(10e6))
Thank you in advance.
Greetings,
Brais.
--
USRP-users mailing list
USRP-users@lists.ettus.com
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
That makes sense. Thank you all for the answers.
I was puzzled because the harmonics had too much power. I've learnt
something new today. Appreciate it.
I'll add some attenuators and see how it responds.
Regards,
Brais.
2015-01-13 18:03 GMT+01:00 Piotr Krysik via USRP-users <
usrp-users@lists.ettus.com>:
Hello Brais,
Look for the post "WBX daughterboard - transmit LO signal harmonics" for
example here:
http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/2013-July/007064.html
and the response from Matt Ettus. Harmonics at odd multiplies of carrier
frequency are typical for wideband transmitters sold by Ettus (those
which don't have adjustable filters at the output).
I don't know where the rest of harmonics that you observe come from.
Maybe there is saturation at some point in your transmitter or
measurement equipment.
--
Best Regards,
Piotr
W dniu 12.01.2015 o 17:26, Brais Ares via USRP-users pisze:
Hello,
This may be an easy question for the experts here. I'm trying to reuse
some code a colleague has done and I encountered a problem he might
not noticed nor corrected.
I want to transmit samples uninterruptedly over the 474 MHz channel.
The code he gave me comes with something like:
usrp->set_rx_freq(uhd::tune_request_t(474e6, 8e6));
A RF analyzer capture shows this
https://dl.dropboxusercontent.com/u/2696878/replicasLO_local.png,
where there are replicas everywhere (see figure
https://dl.dropboxusercontent.com/u/2696878/replicasLO.png with more
freq span). I checked the documentation and arrived at the conclusion
that we shouldn't mess with the /LO_offset/ value in this case so I
used instead:
usrp->set_rx_freq(uhd::tune_request_t(474e6));
Now it got better as the local replicas disappeared but not the ones
further away (figure
https://dl.dropboxusercontent.com/u/2696878/replicas_noLO.png). We
would be interested in having only the signal in the center freq and
no signal on its multiples. I'm not sure if there is any way to
correct this or it's just the way it is. Any advice?
(we are also using usrp->set_tx_bandwidth(10e6))
Thank you in advance.
Greetings,
Brais.
--
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
--