I am using a custom PCB to evaluate the HMC830. The board has been designed using the HMC830 evaluation baord as a reference (using an HMC1060 LDO and the same loop filter configuration). After countless hours oft trying to achieve stable operation, I would appreciate your help on my problem. At first I am fassing issues regarding heat generation which I suspect to be the reason not to lock. The part heats up pretty fast after powering it up. I cannot give exact temperatures but I estimate it to be more than 50 or 60 °C, so that it starts to hurt my finger after a few seconds when touching. (I placed thermal vias on the exposed ground pad). Could you guess any cause for the heat generation which I assume to be way too high?
However, I am perfectly able to program the PLL as well as the VCO subsystem using SPI open mode. I am getting RF output which is far away from being stable (possibly raw VCO output) although output frequency is "close" (in terms of a few MHz...) to the desired output frequency (I assume due to AutoCal). Effects are the same, regardless which frequency I want to generate or if operating in fractional / integer mode or fundamental / divider. I monitor the output spectrum on an spectrum analyzer and I am using the LD pin to monitor lock detection.
I observed that the VCO tuning voltage is constantly at 5V. It does fall to ca. 0V when inverting phase detector polarity (which also causes the frequency to change - CP out to Vtune trace is okay...). Register settings are like described in the operating guide and I compared them to the HMC830 evalution kit software register settings. Performance is not influenced by the CP current or the offset current which I used EQ 9 from the guide for to calculate.
I am assuming that the PD / CP does function properly, maybe as a consequence from heating up... Could you recommend further steps?
By the way I am using a 40 MHz LVCMOS TCXO.
Thanks in advance.