I use ADF4360-9 as a heterodyne at freqs 125...140 MHz (step and PFD freq = 25 kHz).
Ref = 55700 kHz (R = 2228),shematic as recommended in ADIsimPLL, both inductors on pins 9,10 = 136 nH (100 nH + 36 nH LQW18A), Loop filter CPP_2C.
Spurs at freqs 139125...139450 kHz (N = 5565...5578) are about -55 dbc.
At 139125 spurs are on 250 kHz from carrier (both sides),
At 139450 spurs are on 425 kHz from carrier (both sides).
All frequencies outside the band 139125...139450 are OK, spurs < 70 dbc.
Can anybody test this situation on EvBoard?
When reference harmonics mixes with the o/p frequency to produce a beat frequency at the PFD frequency or multiples of the PFD, this increases the expected spur levels at the PFD/multiples of PFD, this is more obvious when low values of PFD are chosen ( 50Khz and smaller).
If extra, low pass filtering on the input Reference signal reduces the spurs then this would indicate the above issue.
DVDD supplies the reference circuit as well as the digital section, removing the decoupling caps from the DVDD pin has reduced this spur , as to why this is happening… is difficult to explain, it appears that by removing coupling from the DVDD into the ground plane, we reduce the above mechanism. Assuming you have seperate decoupling caps on AVDD and DVDD, can you try removing the decoupling caps on DVDD?
No, removing caps didn't help, the LED lock indicator began to blink.
I've tried to increase PFD frequency to 400...1200 kHz. The spurs were decreased, but the frequency error is too big.
Did you try low pass filtering the Reference?
I've tried to filter the Reference, it helps, thank you.
I didn't find any requirements on the reference purity in pdf. I had 2nd and 3rd ref harmonics about -25...30 dBc (ref is also ADF4360-9 with 3rd order LPF on DIVout). Then I've tried external BPF + notch filter (2nd harmonic), the spurs were reduced slightly (3...6 dB).
Now I'll try to make a new pcb with bilt-in filter, I think it will help more.
I never saw this effect in synthesizers with external VCO.
Are there any ref purity requirements for the ADF4360 for the filter calculation?
Reference spectral purity requirement will depend on the application as the noise of the reference inside the loop is not attenuated, but is well attenuated outside the loop bandwidth. We dont spec it.
Check the spectral purity of your reference?( DIVOUT from the other ADF4360-9).
It would help if you sent me your .pll file that you used for both ADF4360-9 as well as your register writes, also a spectral plot of your ADF4360-9 reference, as well as your ADF4360-9 heterodyne