Phase delay between 2 different RF chains

[Maitry07]

Hello support team,

I have 2 different RF chains both can work with 2 GHz RF input but having different Balun, lumped components and matching network. So when I apply same input signal to both RF chains then how can I calculate the phase delay between the 2 chains due to non-identical components in both chains?
Is there any method to have an approximate idea regarding the same.
awaited your guidance.

 

[psach27]

Hi!

You can measure with a VNA the S-parameters of the two chains and compare the results.

With the phase of S21 (transmission between input (port1) and output (port2) of your chain you can calculate the difference in degrees. Be cautious, you have to compare the unwrap version of the phase because usually the phase have several jumps between -180º and 180º.

 

[Maitry07]

Actually , the chain I am talking about is inside the high speed transceiver eval board whose output are directly interfaced with the high speed receiver ADC pins.
That's why I can not test it through VNA.
So, I am asking for any approximate calc that I can done to estimate the phase delay.

one method I have figure out as below.
1. I am applying 360 deg phase diff signals to both the chain and I am getting I and Q components at the output of ADC+DDC.
2. Based on the received I and Q components, I have mathematically calculate Atan(Q/I) on all the received I and Q components from both the channel and then subtract the ATAN(O/I) for ch 1 - ATAN(Q/I) for ch 2
and then based on quadrant I have added the phase shifting as below.
If ch 1 I and Q are in 3rd quadrant and ch 2 I and Q are in 2nd quadrant , then subtract 180 and add 180 to the subtracted output.
If ch 1 I and Q are in 1st quadrant and ch 2 I and Q are in 4th quadrant, then No change in the subtracted output.
If ch 1 I and Q are in 4th quadrant and ch 2 I and Q are in 3rd quadrant, then 360 deg add and 180 deg sub in the subtracted output.

and so on.

So, based on that I am getting constant 81 deg phase difference.
So, though I am applying 360 deg phase diff from external, I am getting 81 deg phase diff due to non-identical chains.

Is my method correct?
If yes, then How can I approximate this change from 360 to 81 through mathematically?

 

[dick_freebird]

How do you hope to calculate with so many unkowns?
Do you have any plan to turn those into "knowns"? Bad
data (including misplaced trust) makes for worthless results.

You could arrive at statistical bounds by analysis of component
variability but even that vague analysis wants data regarding
that. And good luck.

May as well just measure.

 

[Maitry07]

So, Is there any method to figure out this phase delay in this eval board? awaited your guidance.

 

[FvM]

Sure, by applying the same input signal to both channels and recording raw I/Q output, bypassing any signal processing that might corrupt the result. Repeat with different magnitude and relative phase.

 

[Maitry07]

yes, thank you so much. I will try with these.