[USRP-users] Transmit DC Offset Calibration

Josh Blum josh at ettus.com
Thu Jun 21 13:23:11 EDT 2012

On 06/19/2012 09:34 PM, Steve Peters wrote:
> Hello,
> I'm hoping there is a crucial step to the DC calibration that I'm
> overlooking.  The DC offset calibration is only making my DC offset worse.
> For my setup, I'm using tx_waveforms --freq=301e6 --rate=1e6
> --wave-type=CONST --ampl=0.0. Without the calibration, the received sine
> wave has an amplitude of about 0.004, and with the calibration the
> amplitude is about 0.01 (transmission over a wire with 30dB attenuation to
> attempt fairness between tests).  This is well above the noise floor.
> I've tried setting --nsamps=1000000 and focusing on the narrow band that
> I'm operating in, but it never seems to actually reduce the DC offset.
> When running the calibration, I disconnect the internal SMB cable from the
> GDB.  Am I missing some other crucial step to the calibration?

Thats correct.

> I'm running UHD 3.4.1 on USRP N210s, rev4.
> When the TX and RX are connected to a common 10 MHz source, the sine wave
> goes away (as expected).  Unfortunately this isn't a long term solution to
> the problem for me.

I'm not sure as to why this might be happening. I'm looking into it. You
might consider moving the LO out of the passband with a tuning offset as
an alternative solution. Some notes here:


More information about the USRP-users mailing list