Post Go back to editing

HMC547ALC3 Mismatch with Datasheet Results

Hi, 
I am using SPDT switch HMC547ALC3 for X-band. While testing, it was observed that when RFC to output RF1 path was ON, not only the insertion loss was poor but it was mismatched. Similar behavior was observed in the case of RF2. Switch IC was also changed to be sure if the IC had issue. PCB layout was also verified in ADS and CST software to make sure about any issue with layout tracks. Short test was also performed. What can be the issue? 
I have attached measured S-parameters for 8-12 GHz (Insertion Loss and Return loss).


   



Query about HMC547ALC3 RF switch
[edited by: Yumna at 5:21 AM (GMT -5) on 24 Dec 2020]
Parents
  • Hi Yumna,

    From the screenshots of the test results, the insertion loss is very high, and the reflection is also very high. It could be that the IC is damaged, or the matching between the IC and the PCB is very bad.

    1. I know you have changed the IC, but did you check the current of CTRL pins A and B? If the IC is damaged, the current should be abnormal;

    2. For the PCB design, do you have a RF thru trace to measure? If the RF thru trace is well designed to to 50Ohm, the insertion loss should be very loss and the reflection should also be very low;

    3. Could that be the problem of soldering between the IC and the PCB? A DC connection cannot guarantee a proper RF performance. Is there any way you can check if the IC is soldered well onto the PCB?

Reply
  • Hi Yumna,

    From the screenshots of the test results, the insertion loss is very high, and the reflection is also very high. It could be that the IC is damaged, or the matching between the IC and the PCB is very bad.

    1. I know you have changed the IC, but did you check the current of CTRL pins A and B? If the IC is damaged, the current should be abnormal;

    2. For the PCB design, do you have a RF thru trace to measure? If the RF thru trace is well designed to to 50Ohm, the insertion loss should be very loss and the reflection should also be very low;

    3. Could that be the problem of soldering between the IC and the PCB? A DC connection cannot guarantee a proper RF performance. Is there any way you can check if the IC is soldered well onto the PCB?

Children
  • 1) Yes I have checked CTRL pins. Both the pins draw normal current which is 3.5uA. 

    2) PCB is designed on Rogers RT/Duroid 6002 with thickness 0.254mm and copper cladding of 1oz. Therefore, RF track width is 0.65mm which is of 50ohm impedance. I have attached image of top layer for your reference. I have also attached comparison of old measured results (which were shared before) and new measured results (of reworked PCB which we did this week). There is no change in insertion loss and return loss. 

    3) PCB soldering was analyzed with X-ray inspection system and IC is well soldered into PCB. 

  • Hi Yumna,

    I am wondering in your PCB design, if there is a RF thru trace for de-embedding? We do recommend include a RF thru trace in your design so that you can verify if the PCB is well designed to be 50Ohm, and also you can use the RF trace to de-embed and get the true performance of the IC.

    Do you have the measurement of the RF thru trace to share?