HMC966LP4E Image Rejection

Hi all,

Our customer evaluate the HMC966 RF performances.

Frequency is ,

・LO=10GHz , +6.5dBm

・RF=18GHz

・IF=700MHz

Following the measurement setting is Typical Application Circuit at the datasheet 12 page.

SG(RF) - isolator - HMC966 - 90 deg Hyblid(IF) - filter - Spectrum Analyzer

SG(LO) - isolator -

Our customer measurement data of Image rejection is the data sheet specifications as well as the 15dB, but it is minimum spec.

Customer target spec is 20dB(min) w / design margin.

What we are thinking that ,

Is it possible to improve the performance by adjusting the line length between the IF1/IF2 output and the 90 deg Hyblid ?

Do we have other way to improve the performance of the Image rejection ?

Best regards,

sss

  • 0
    •  Analog Employees 
    on Jun 27, 2016 5:27 PM

    Hi,

    Could you please double check the RF,LO,IF frequency? I think there are some errors in the frequency.

    Thanks,

  • Hi huangzeyan - san,

    Sorry for the missing freq information.

    Frequency is ,

    ・LO=10GHz , +6.5dBm >>> 16.99~18.99GHz(8.495~9.495GHz x2)

    ・RF=18GHz >>> 17.7~19.7GHz

    ・IF=700MHz >>> 710MHz

    Best regards,

    sss

  • 0
    •  Analog Employees 
    on Jul 1, 2016 12:30 AM

    Hi,

    We measured the conditions at LO=18G and the image rejection is about 25dBc.

    I think the low image rejection may come from the filter they had before the SA. To verify the mixer performance, it is recommended to measure the IF signal without the filter, connected to the SA directly or using a IQ analyzer without hybrid. If the result is better than with the filter, you may need to redesign the filter.

    Otherwise, adjust the path length is an option. What needs to be careful is that the parts vary from each other. Every part may have different path length adjustment. And one solution might only work for one fixed operating frequency.

    Thanks,