We are using ADF4355 as an LO of 5900 MHz in an upconversion design using mixer HMC6505A and baseband from ADRV9361 at 2450 MHz. At RF out of the mixer, we are seeing spurs/harmonics ~500 kHz from the 8350 MHz as shown in the image-IMG_0533.jpg.
We believe it's a similar issue to ADF4355-2 close-in frequency independent spur problem. For the most part, we are following the reference design from ADF4355 datasheet Figures 44 and 45 but using a 100 MHz reference oscillator and using the registers to write to ADF4355 from ADF4355 eval software. Please see the block diagram and ADF schematic for the board we have designed for the above application.
Unfortunately, we do not have any RF taps on the design, so we cannot measure the reference or PLL output on the spectrum analyzer, but we have evaluated the performance of this design using eval boards, and do not observe spurs on there. We use a few switching regulators on this board, but their switching frequencies are 1, 1.2, and 2 MHz, but the harmonics we are seeing are at 500 kHz. We have tried checking if any of the switching regulators on the board are causing these harmonics, for that we checked the output on the spectrum analyzer around 500 kHz, and did not see anything, which makes us think it is not coupled from the switching regulators, and rather bring generated from the PLL itself.
We found some conflicting information about Rset (Bias Current Resistor) values in the datasheet: suggests 5.1kΩ, but the reference design in figure 44 and 45 show a 4.7kΩ, and we ended up using 4.7 kΩ resistor on the board design, but that is not very different which makes us think that is not it. Also on the ADF4355 evaluation software I saw the button for Step size, what is that step size for, and what would that change?
We would really appreciate any help to resolve this. Please let me know if you need any further information.
Thank you in advance,