LMH3401EVM: Why the s-parameters of LMH3401 has small ripple？

Hello,

    Your setup looks fine to me. Can you tell me what type of cables you are using and the approximate length? I will test it on my end, and I will see if I can replicate your problem. 

Thank you,

Sima

Hello,

    I test it with the cables FL141-24SM+ ,from Mini-Circuits,with 24 inch(600mm) length.

Thank you,

Jiang

Hello Jiang,

   Thank you for the information. I should be able to give you an update by the latest mid next-week (Jan 22).

Thank you,
Sima

Hello Jiang,

    I tested your setup from the diagram you shared earlier with two types of cables: around 6 inches and around 24 inches. I did not notice severe ripple in the longer cable unlike in the scope photo you shared; but, there were partial ripples n the 24 inch cable(600mm) and none in the 6 inch (150mm) cable. It is recommended to use shorter cables; especially in very high frequency applications such as in this case (preferably less than one inch). Would you be able to redo the measurement with shorter cables, and let me know if it minimizes the amount of ripple.

Thank you,

Sima

Hello sima,

  Thank you for your answer, I will repeat the test with shorter cable after Chinese New Year holiday. Can you share the S-parameters result files in your test? The s-parameters files I got from the forum have too few points from 1G to 8G.

Thank you,

Jiang

Hi,

the 50R series termination resistor in the LMH3401EVM is formed by the series circuit of an external 40R2 resistor (Ro+, Ro-) and the internal 10R output resistance of LMH3401. But the internal 10R output resistance can show manufacturing tolerances of up to +/-2R. The same is true for the input resistance. Also this can show manufacturing tolerances. Have a look into the datasheet of LMH3401.

This can result in some mismatch of termination resistors in connection with the 50R cables, the 50R generator and the 50R scope. This would affect the transmission performance, the more the longer the cables are.

And the SMA cables aren't perfect as well. A specified maximum frequency of 18GHz doesn't necessarily mean that they are really good in the GHz range. Try to find a real good semi-rigid SMA cable. They are some more expensive than your FL141-24SM+ cable. Or, if you cannot buy such a semi-rigid cable, keep the cable as short as possible!

Kai

Hello Kai,

    Thank you for the very detailed explanation! That would explain the difference in our observation between the partial and severe ripple.

Jiang,   

    I can get you those result files by the end of this week. Happy Chinese New Year!

Thank you,

Sima

Hello Jiang,

   I have attached the data I obtained from the LMH3401EVM. 

LMH3401_SEin_SEout_FreqResp.xlsx

Note: This is a rough data collection to match your setup (with shorter cables: about 10-12 inches). Please refer to the LMH3401 datasheet for a more standard data collection figure.

Thank you,
Sima

The original datasheet plots were done with a 4 port network analyzer using an automated calibration box which will remove the cabling effects. 

The original network analyzer sweep shown here with ripples looks like a poor calibration (not oscillation).