according to https://www.hittite.com/content/documents/operating_guide/pll_operating_guide_rf_vcos.pdf there shall be no spurious at frequencies other than multiples of PD frequency in Integer Operation. Unfortunately, that is not my case, see the attached plot of single sided power spectrum of my HMC1033 output, where some spurious are at offsets of several (tens of) kHz. The power of 100MHz carrier in the attached figure is about 8dBm. I designed a PCB with HMC1033 inspired by respective evaluation board using HMC1060. Moreover, I have another board with HMC1032 using different low noise voltage regulators, but with similar result (various carrier frequencies tried). As already mentioned, the PLL is operated in integer mode with PD frequency of 1MHz and passive filter from datasheet. I do not see any suspicious signal on power supply.
Do you have some tips, what could cause the spurious?
Thank you for the tips. I use the 10MHz reference from the spectrum analyzer. When measured by the spectrum analyzer itself, no spurs are visible. Not sure, if this can be considered as a proof of spur-free source. I assembled another board without the voltage regulators, just HMC1033 (and microcontroller for HMC1033 programming, held in reset during measurement) supplied directly from a battery. All nearby electronic devices was turned off, but the spurs are still there. Maybe really some interference from environment; unfortunately, I have no possibility to test it inside an EMI chamber. I tried to enclose it inside a box of 2mm aluminium which is probably not sufficient shielding.