HMC7044 输出时钟有个几十KHz的杂散

Hi,

If you publish the question in English, maybe I can help with the spurs. 

From my limited understanding of the translation,  you are observing some spur issues at the output. You are using both PLL1 and PLL2. am I correct? 

I don't recommend a filter to the output as it will degrade the performance of the output clock. You can decrease the charge pump current of PLL1 and PLL2 to reduce the spur. It may require some fine-tuning with charge pump values. You can try to change the PFD frequency of PLL1 and PLL2 to see if the spur is shifting out of the band.

Thanks,

Emrecan

Hi,

If you publish the question in English, maybe I can help with the spurs. 

From my limited understanding of the translation,  you are observing some spur issues at the output. You are using both PLL1 and PLL2. am I correct? 

I don't recommend a filter to the output as it will degrade the performance of the output clock. You can decrease the charge pump current of PLL1 and PLL2 to reduce the spur. It may require some fine-tuning with charge pump values. You can try to change the PFD frequency of PLL1 and PLL2 to see if the spur is shifting out of the band.

Thanks,

Emrecan

Hi Emrecan,

Your understanding is correct. The customer found some spurs at the output and is looking for suggestions to improve the spur performance. Thank you for your suggestions.

Thanks

Elvin

 emrecangidik , thanks very much.

I use HMC7044 with PLL1 and PLL2 enable.

I have tried to change the charge pump current of pll1 and pll2, but have no improvement.

PLL1 charge pump current now set 0x5, PLL2 set  0xF

I have change the PLL1 PFD frequency from 10M, 5M ,2.5M ,1.25M .... it have no improvement.

PLL2 PFD frequency now is 100M , when i change the PFD freq to 200M , PLL2 is unlocked.

when i change the PFD freq to 10M , the spurs become very big.

After HMC7044  finish config, the HMC7044 ref clock in can observe  26K spurs, but when i set HMC7044 to sleep mode. the spurs

in ref clock disappears.

I suspect the PCB ground of HMC7044 have some problems, but have not find the proof. 

Hi Jasper,

If the charge pump and PFD frequency are changed and the spur size does not change, it indicates that the spur is not originating from within the PLL loops. I suggest investigating the VCO supply and Output buffer supplies for any potential issues. How is the supply for HMC7044 being provided? Is a switch-mode power supply or an LDO being used?

Thanks,
Emrecan

I use LTM4644 and LT1963 to provide the 3.3V supply of HMC7044; now I find LTM4644 output have some nosie of 26KHz;i bypass  LTM4644，the spurs disappear

Yes, I used EVB to configure and test according to your parameters, and there was no obvious nearby noise. The purity of the power supply noise is also very important.