Phase Noise at 24 GHz
Posted: Thu Dec 21, 2017 8:37 pm
I have spent the last few days optimising my 24 GHz system, built in the main from DMC modules, for DATV. The challenge has been to reduce the phase noise of the ADF5355 local oscillator source to acceptable levels. The ADF5355 runs at 11.805 GHz, is amplified in an old Sat LNB preamp to +10 dBm and then doubled in a "Franco" doubler.
Initial tests gave good results on narrow band (with a fairly pure tone from the Bell Hill 24 GHz beacon) but MiniTioune would not achieve a lock on DATV signals. I then applied the modifications detailed in this month's ScatterPoint to reduce the on-board power supply noise, and also enabled the clock doubler and increased the charge pump current as suggested by G4JNT. Both these modifications were needed to reduce the phase noise to an acceptable level. The effect of the noise can be seen on the received constellation plot below, with the elongated quadrant responses. The effect of the noise seemed relatively independent of SR; it has similar effects at 333KS and 4MS, but I suppose that depends to some extent on the noise spectrum.
I do plan to publish the ADF5355 parameters that I used, and make a PIC program available for the "ADF4351" PCB sold by the BATC shop that will work with the ADF5355, but I think that there may be scope for more optimisation, so I'll try this first.
Today's 2-way contact was 29 km, and confirmed that the ADF5355 is usable at 24 GHz. After equipment modifications, Noel and I hope to start increasing the distances.
Dave
G8GKQ
Initial tests gave good results on narrow band (with a fairly pure tone from the Bell Hill 24 GHz beacon) but MiniTioune would not achieve a lock on DATV signals. I then applied the modifications detailed in this month's ScatterPoint to reduce the on-board power supply noise, and also enabled the clock doubler and increased the charge pump current as suggested by G4JNT. Both these modifications were needed to reduce the phase noise to an acceptable level. The effect of the noise can be seen on the received constellation plot below, with the elongated quadrant responses. The effect of the noise seemed relatively independent of SR; it has similar effects at 333KS and 4MS, but I suppose that depends to some extent on the noise spectrum.
I do plan to publish the ADF5355 parameters that I used, and make a PIC program available for the "ADF4351" PCB sold by the BATC shop that will work with the ADF5355, but I think that there may be scope for more optimisation, so I'll try this first.
Today's 2-way contact was 29 km, and confirmed that the ADF5355 is usable at 24 GHz. After equipment modifications, Noel and I hope to start increasing the distances.
Dave
G8GKQ
