Help with network analyser

[Element7k]

Hi all,

I need some help with changing the reference plane on a network anaylser to measure impedance. Deembedding? I have a HP 8722 Network anaylser. Can anyone help please?

Cheers,
Element7k

 

[flyhigh]

I am not quite sure for the particular NA you have, but for plane shifting only without de-embedding you can use electric delay to come to the reference you need. It is beneficial if you know your delay exactlly, but it is adjust easily so you can play a bit to get to the reference plain you need. Be carefull, if you introduce delay phi for S11, you should introduce 2*phi for S21!

FlyHigh

 

[Element7k]

Hi flyhigh,

thanks for the info.

Is there a formular to do the estimation?
Do I use Effective velocity = c / sqrt(Epsr of coaxial cable)?

How will I know if I got the right refernce plane after adjusting the electric delay?

Will very much appreciate your help.

Cheers,
Element7k

I am not quite sure for the particular NA you have, but for plane shifting only without de-embedding you can use electric delay to come to the reference you need. It is beneficial if you know your delay exactlly, but it is adjust easily so you can play a bit to get to the reference plain you need. Be carefull, if you introduce delay phi for S11, you should introduce 2*phi for S21!

FlyHigh

 

[flyhigh]

Hi Element7k,

For the first questiion, yes, your formula is OK. The second one is a bit of a story. As with other things concerning NA you need a calibration. I assume that you have to eliminate the lengh of a cable. Do the usual callibration. Then attach the cable and leave the other end open. Look on the phase of S11. Adjust the electric delay until you get the phase of an ideal open (0deg). Your reference plane is then shifted to the end of the open cable. If the length of a cable is greater than wavelength you can account for that roughly and then fine-tune the delay. This can be scaled to the length of the same cable type and delay of any length be known, but you have to do it once on an open cable if you don't know Er of the cable.

FlyHigh

 

[Element7k]

Hi flyhigh,

pardon my ignorance in this but how do I account for the length of the cable if it is > wavelength? Also when you say scale it for use on any same cable of any length, do you mean by adjusting the electric delay?

Many thanks for you help. It is great help to me.

Cheers,
Element7k

Hi Element7k,

For the first questiion, yes, your formula is OK. The second one is a bit of a story. As with other things concerning NA you need a calibration. I assume that you have to eliminate the lengh of a cable. Do the usual callibration. Then attach the cable and leave the other end open. Look on the phase of S11. Adjust the electric delay until you get the phase of an ideal open (0deg). Your reference plane is then shifted to the end of the open cable. If the length of a cable is greater than wavelength you can account for that roughly and then fine-tune the delay. This can be scaled to the length of the same cable type and delay of any length be known, but you have to do it once on an open cable if you don't know Er of the cable.

FlyHigh

 

[flyhigh]

Hi Element7k,

think this paper will describe anything you need much better than I can tell. Pay attention to pages 7-13 and 7-14, port expensions are the same as electric delay.

FlyHigh

 

[flyhigh]

Hi Element7k,

think this paper will describe anything you need much better than I can tell. Pay attention to pages 7-13 and 7-14, port expensions are the same as electric delay.

FlyHigh

 

[hyuang]

are you think in calibration

 

[harkonnen]

Actually a website that is very helpful is www.vnahelp.com
It has valuable information for deembedding on a vector network analyzer.