[srslte-users] Calibrating time_adv_nsamples

[Lorem Ipsum]

Hello all,

I'm trying to set up a link over a fairly lame channel emulator. I
managed to establish the communication by playing with
time_adv_nsamples, eventually going as high as ~300 samples by
(very-very) exhaustive search, and the link seemed to be stable. After
having done some tweaks to eliminate IQ imbalance, I'm unable to get
the system working properly. The UE is able to find the cell,
connects, but drops the connection immediately after. I believe the IQ
tweaks should not affect in-band performance in an adverse manner.

I'd like to reconfirm that time_adv_nsamples is correctly calibrated.
Is there a reliable and systematic method to do so? Also, what else
could have gone wrong?
I'd appreciate any suggestions.
Best regards,
L.