Hi out there,
I have several questions on Frequency Hopping.
I’m trying to send a constant with the *_siggen.py files on different
frequencies in short intervalls.
Because those frequencies are spread over a band of 25MHz I need to
change
the center frequency of the board very fast(within micro seconds).
Now the question.
Is the usrp1/2 capable of this via the functions in python? How much
time
does the usrp need to change the center frequency?
I have read a few articles abut the minimal available time interval on
desktop pcs. On ubuntu (hardy) there is a file called
/usr/src/linux-Headers-2.6.24-25/kernel/Kconfig.hz where I can adjust
the
“Timer Interrupt Frequency Configuration” which allows 1000Hz which
represents 0.001 seconds. (important for python timer functions)
Now the question.
Is this 0.001 s limit really the shortest time span available for this
pc?
It seems to be too long.
Is there an other possibility to send ~8 carriers distributed over a
bandwitdh of 25MHz without preprocessed data?
Realtime/Frequency Hopping (tH=~several micro seconds)
On Tue, Dec 8, 2009 at 6:43 AM, Hanno Jung
[email protected]wrote:
Is this 0.001 s limit really the shortest time span available for this pc?
It seems to be too long.
Is there an other possibility to send ~8 carriers distributed over a
bandwitdh of 25MHz without preprocessed data?
Realtime/Frequency Hopping (tH=~several micro seconds)
I once looked in to this, and I believe that Matt once told me it takes
~200us for a daughterboard (I think it might have been RFX2400) to lock
to a
new frequency. You can always set your bandwidth to a larger range and
then
transmit at different offsets within that range.
On Tue, Dec 8, 2009 at 12:05 PM, George N. [email protected]
wrote:
I once looked in to this, and I believe that Matt once told me it takes
~200us for a daughterboard (I think it might have been RFX2400) to lock to a
new frequency. You can always set your bandwidth to a larger range and then
transmit at different offsets within that range.
No need to re-tune as long as you’re within 32MHz (maybe 64MHz due to
complex sampling?).
Since you have a good amount of bandwidth to work with at the host,
you could just “hop” at the host level with your samples and a mixer.
Or, if you need more bandwidth, you can put the hopping sequence into
the FPGA going into the DUC and go to a different phase accumulation
based on how many samples have just been transmitted.
Plenty of ways to skin this cat without having to retune as long as
you can fit within the DAC bandwidth.
Brian
Brian P. schrieb:
Or, if you need more bandwidth, you can put the hopping sequence into
Hmm I thought about this before too, but what do you mean with
So if you constrained yourself to that bandwidth, you could just drive
being able to tune the synthesizer to the exact frequency (at least I
Hi,
what about calculating on the host? At the end I will have a signal at
~25Mhz (baseband) which will be converted up by the usrp.
Is this really possible? It seems to me that this will be to high for
the pc to calculate.
Am I right?
Seems like I have to shift work to the fpga.
On Sun, Dec 13, 2009 at 4:05 PM, Hanno [email protected]
wrote:
Hi,
what about calculating on the host? At the end I will have a signal at
~25Mhz (baseband) which will be converted up by the usrp.
Is this really possible? It seems to me that this will be to high for the pc
to calculate.
Am I right?
It all depends on the bandwidth of the signal in question.
Your first question made me think you had a smaller bandwidth signal
jumping around a 25MHz bandwidth somewhat quickly. Since the USRP2
can handle 25MHz of bandwidth being sent to it, you can place your
smaller bandwidth signal anywhere within that 25MHz as calculated at
the host.
Does that make sense?
Seems like I have to shift work to the fpga.
If you have a 25MHz bandwidth signal that you need to move around the
spectrum, then yes - you will have to shift inside the FPGA or look
into how quickly you can retune the front end.
I hope this is a bit more clear.
Brian
Brian P. schrieb:
That’s exactly the case this is all about 
It’s a constant/sine which should hop within a band of 25Mhz.
So I will try to implement the hopping directly on the host without
retuning.
Thx