Calibrating USRP

GNU Radio
1

Hi all guys (yet again),
I have a question about calibrating received power level of my USRP,
equipped with wbx, lftx and lfrx.
I am not sure whether it is apropriate to discuss that here.
Basically I would like to be able to determine absolute power level I am
receiving. I suppose that basic procedure for doing this would be using
super precise, expensive signal generator. Then tracing signal of know
power to daughter board and measure power level in fft sink. And maybe
also
doing this for complete band of daughterboard usage. If the previous is
correct, than I would like to ask you, how do you store calibrating
data,
inside .h file or …?

Best and thank you,

2

Discuss-gnuradio mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio

3

Thank you Marcus for super fast answer. Well, for the begining I would
like
to skip temperature calibration, only power level calibration regarding
frequency. I have following doubts:

  1. To which frequency should be tuned USRP if for example generator
    generates 400 MHz. Should it be almost the same frequency, so that
    downconverted signal falls very close to DC or something else.
  2. What is the best advice for storing calibration data.
4

I think that i am a little bit confused with calibration. Is there any
way to determine the gain of the circuitry in front of the ADC? Is it
correct enough to say that power level at the input of PGA in front of
ADC is
just ((sampled_value*Vp-p)/2^12 ^ 2 ) / Rinputpga?

The total gain in front of the ADC will vary from daughtercard to
daughtercard, and with frequency.

Calculate the power inside the flow-graph, which is proportional to
AVG(II + QQ). For RMS power, there’s an RMS block.

What you need to do is use a calibrated-in-power signal source, and
use that to calibrate your receiving setup.

You can’t reliably backtrack from the values coming out of the USRP,
because you don’t know the total gain in front of the ADC to sufficient
precision.

5

I think this should be the last question in this thread. Which input
signal
levels should I stick? OIP3 at the first LNA MGA-62563 is 23 dBm. Since
the
gain is arround 22 dB, is it ok to stay below -20dBm?

Best,
Nemanja

6

On 07/15/2013 08:55 AM, Nemanja S. wrote:

I think this should be the last question in this thread. Which input
signal levels should I stick? OIP3 at the first LNA MGA-62563 is 23
dBm. Since the gain is arround 22 dB, is it ok to stay below -20dBm?

Best,
Nemanja

Yes, signals below -20dBm should keep things reasonably linear.

The thing to remember is that these receivers are generally designed for
from-the-antenna use, where signal levels below -60dBm are quite common,
and
signals that are at -20dBm are “screamingly loud”.