The attenuator in the AD9361 transceiver chip used on the B210 is
pretty accurate. For a certain frequency, all you need is one
measurement. Then the output power will track pretty well with the
gain setting. BTW, the TX gain range on the B210 is 0 to 89.8 in
0.2 dB steps. Above around 80 or so, wide band signals start to
get distorted with 3rd order IMD.
You have to connect your B210 directly to the spectrum analyzer,
preferably with a 20 dB or more attenuator just to be safe.
However for OFDM (or any noise like signal), you have to compensate for the spectrum analyzer resolution bandwidth. Here's an example. The signal is 5.83 MHz wide OFDM and the resolution bandwidth is 30 kHz. The compensation would then be 10 log 30000/5830000 = -22.9 dB. So the approximate -61 dBm reading on the spectrum analyzer is really -61 dBm + 22.9 dBm = -38.1 dBm of average power.
If you have a spectrum analyzer that can automatically measure channel power, it's easier and more accurate.
The spike to the right of the main signal is the DC offset shifted out of band with UHD offset tuning.
Ron
On 04/24/2018 11:45 PM, Prabhat Kumar Rai wrote:
Dear Marcus & Vitt, Thanks for your suggestion. As I am new in this field I don't know how to calibrate spectrum analyser with usrp. Is there any video tutorial or any website link you can provide will be helpful? I have searched on internet but didn't get a thing for calibration. Suggestion from other members is also welcome. Regards Prabhat On Tue, Apr 24, 2018 at 3:03 PM, Müller, Marcus (CEL) <address@hidden> wrote:You'll have to calibrate that yourself. There's no fixed formula for all possible individual devices, and all possible signals (with all possible PAPRs, which is especially relevant for OFDM). Best regards, Marcus On Tue, 2018-04-24 at 13:05 +0530, Prabhat Kumar Rai wrote:Hello, I am using tx_ofdm.grc file to transmit a signal via USRP B210. I am fixing my system as center_freq 1.8GHz, BW 20MHz, Sample rate 1.92M, Gain 20 dB, when I am checking the gain in the spectrum analyser it comes out to be -77dBm, I know that channel is giving me around 30 dBm attenuation still I am unable to compute the actual transmit power of usrp. I know that only way to know transmit power is using the combination of frequencies, sampling rates, master clock rates, analog bandwidths and gain but don't know how and where to use them, all of them are already in a usrp_sink block. Is there any other way to compute actual transmit power of Usrp or any other formulae??? I have to use that power for RSSI calculation. Regards, Prabhat _______________________________________________