On 10/26/2020 09:16 AM, Jon Beniston wrote:
Hi Marcus,
I think this may be a problem in UHD's
d9361_device_t::_tx_quadrature_cal_routine()
This calculates a variable max_cal_freq, that's the maximum frequency used
by the calibration. It then checks this against bbbw (the baseband
bandwidth), but it doesn't seem to check the LPF's bandwidths to make sure
they are sufficiently wide.
Looking at what I think is the equivalent in the Analog Devices reference
API:
https://github.com/analogdevicesinc/no-OS/blob/master/ad9361/sw/ad9361.c
in the function: ad9361_tx_quad_calib(), on line 3179, they compare the
required frequency against the RX and TX LPF's bandwidth, and if they aren't
4x greater, it looks like they are temporarily increased during the
calibration.
Cheers,
Jon
Thanks for your persistence on this.
I'll put together a ticket for Ettus R&D to look at.
_______________________________________________
USRP-users mailing list
[email protected]
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
