It's all about the Software

Hello Simon,

Further to my previous post about the Pluto uncommanded transmit bug when starting or changing bandwidth in Console.

As I noted, I first noticed a similar problem had newly occurred using Satsagen, the spectrum analyzer program for the Pluto. After I brought it to the developer's attention he came up with a fix. I got his permission to share the info here.

His reply follows-

"In SATSAGEN, when the Generator is not used, and only the SA is running, is issued this command before setting the sampling frequency:

iio_device_attr_write(phy,"calib_mode","manual_tx_quad"); //(phy is the device descriptor of "ad9361-phy")

The above function disables the automatic TX quadrature feature that generates spikes.
Otherwise, when the generator or TSA is used, the same function is called before every sampling frequency setting but with different modes, depending on the user setting choice. If the user sets the Auto TX calibration check on the Pluto device, then the function is called with "auto" parameter:

iio_device_attr_write(phy,"calib_mode","auto");

Else the function is called with "manual_tx_quad" like before.

Warning: Using the TX with the calibration mode to manual_tx_quad doesn't generate spikes, but because the TX quadrature feature is off, the output could be out of specification! For this, in SATSAGEN, if the TX quadrature calibration is disabled, this is demanded to the user via the TX IQ balance and the Phase Gain settings panels! "

Once again, this bug causes full rf power uncommanded transmit of the Pluto whenever Console is first started or when the bandwidth setting is changed. There is a fix.

regards, Roy

Hm, at least for Plutos the output usually must be amplified significantly before it can be used in a useful transmitter. I wonder if it is possible to send the TX signal to the unused TX output. If so then it could receive a 50 ohm termination. The brief calibration transmit could be directed there and then the output could be switched to the real output for normal operation. I wonder if that would work right or if internal leakage problems exist that would preclude it from performing as it should.

{^_^}