# ad9361-setup -l off
phy:            /sys/bus/iio/devices/iio:device0
ensm_mode:      fdd
sampling_freq:  4000000
rx LO:          2400000000
tx LO:          2400000000
rx bandwidth:   4000000
tx bandwidth:   4000000
rx gain mode:   manual
rx gain (dB):   30.000000 dB
tx atten (dB):  -10.000000 dB
rx rssi:        59.50 dB
loopback:       0
# rf-sweep -b  
Leakage baseline. DISCONNECT the cable from RX1A before reading this.

gain 0    peak 0          decoded 0 of 500
gain 10   peak 0          decoded 0 of 500
gain 20   peak 0          decoded 0 of 500
gain 30   peak 0          decoded 0 of 500
gain 40   peak 252        decoded 52 of 500
gain 50   peak 1167       decoded 500 of 500
gain 60   peak 3702       decoded 500 of 500
gain 70   peak 10539      decoded 500 of 500

Write these down, and compare at LOW gain -- 0 to 20 dB. Above about 50 dB
the peak saturates near 57000 regardless of where the signal came from, so a
high-gain comparison proves nothing. At low gain a cable with 30 dB of
attenuation should sit far above the numbers above; if it does not, the
signal is not going through the cable.
# rf-sweep -g "0 10 20" -t 10 
TX1A gives up to +7 dBm; RX1A takes +2.5 dBm absolute maximum.
There must be at least 30 dB of attenuation between them.
Is the attenuator in the cable? [y/N] y
 atten   gain        rssi      peak     ok payl.bad  hdr.bad  verdict
   -10      0     53.75dB       321    500        0        1  ok
   -10     10     66.75dB      1041    500        0        1  ok
   -10     20     76.75dB      3019    500        0        1  ok
#