Impedance matching of power amplifier with s2p file

[mayjailomo]

Hi all,

I am new to this field. I am now drawing a PCB for a power amplifier (HMC408). Although analog device has provided the reference schematic and PCB layout, external matching circuit is required for this IC, and the matching circuit in the reference design is designated for RO4350 and CPWG while I am using other substrate and another type of transmission line (microstrip line).

I have searched in google and someone suggested to use the s2p file (they said that the s2p file is for the IC, not for the eval board) provided by analog device to do the impedance matching in ADS. I have tried it on ADS but I found that the S11 and S21 is pretty good and matched with the data provided in the datasheet even when I simply connected the 2 ports with 2 50Ohm terms. When I connected the s2p output port to the external matching components suggested from the datasheet, the performance dropped. May I ask if that's normal? Can I really do the impedance matching using the s2p file provided by analog device?

I am sorry for my poor english.
Many thanks!

Reference schematic for HMC408:

Setup in ADS: ( The parameters of the CPWG are measured from the gerber file provided by analog device)

S11, S21: (connected to the terms directly)

s11,s21 with external matching circuit:

 

[BigBoss]

Microstrip Lines were used in schematic but you measured CPWG on application PCB.
Are you sure that you have already used the substrate parameters as same as actual PCB ??

 

[mayjailomo]

Microstrip Lines were used in schematic but you measured CPWG on application PCB.
Are you sure that you have already used the substrate parameters as same as actual PCB ??

Thanks for your reply! According to the gerber files and reference pcb layout, they are using cpwg instead of microstrip lines. I have measured the length of the TLs in the gerber file, they are exactly 0.2" and 0.1".

For the substrate parameters, I set the parameters according to the datasheet of RO4350B which is the substrate stated in the datasheet of HMC408. However, the thickness of the copper and the substrate remains uncertain. So i tried all the available thickness of RO4350B, they have slightly different in performance, but as i still dont understand why the S21 is so good when the s2p is connected to a 50 Ohm term directly, I didn't put much time to analyze the results.

 

[BigBoss]

There is a s-parameter file for this circuit that will be used for Load which is seen by Output Pin thru 50 Ohm.
Have you ever checked that ?? There are some tricks for optimization in a certain band.
I think the application circuit has been optimized for 5.7-5.9 GHz band and the circuit should be examined for other frequencies.

 

[mayjailomo]

There is a s-parameter file for this circuit that will be used for Load which is seen by Output Pin thru 50 Ohm.
Have you ever checked that ?? There are some tricks for optimization in a certain band.
I think the application circuit has been optimized for 5.7-5.9 GHz band and the circuit should be examined for other frequencies.

Are you talking about the s1p file? I dont know how to use this data to do the matching. Do you know where can I find some related tutorials?

 

[FvM]

I have searched in google and someone suggested to use the s2p file (they said that the s2p file is for the IC, not for the eval board) provided by analog device to do the impedance matching in ADS.

"They said" doesn't comply with a comment in the s2p file

!Misc: Input reference is the input pin of the device, output reference is the last component on output line

It's also obvious that the s2p data are matching the datasheet plots describing the output matching implemented in eval board.

 

[mayjailomo]

Thanks for you reply. I am now sure if the replies in that forum are from analog device's supporting team or not.

But if the s2p file is describing the performance in the eval board, does that mean it is impossible to redesign a matching circuit for the board with difference substrate?

Should i just give up and choose another PA...?

--- Updated Apr 4, 2022 ---

"They said" doesn't comply with a comment in the s2p file

It's also obvious that the s2p data are matching the datasheet plots describing the output matching implemented in eval board.

Sorry i forgot to quote your reply.

 

[FvM]

I'm not quite sure which port 2 setup has been used for the s2p file. It is named "de-embedded", so we would expect a reference plane at the output terminal, "last component on the output line" sounds different.

Regarding possible modifications of the eval board, TL1 and TL2 are specified as 50 ohm lines, respectively any implementation of 50 ohm lines with the same electrical length should be equivalent.