Calibration of Component - use SOLT?

[oilytong]

Hi,

I am doing a project and I need to obtain the s-parameters of a chip that I have. Someone suggested that I use SOLT or TRL, but I am not sure as to which one I should use.

Would greatly appreicate it if someone could offer some pointers as to which one I should chose or is there a better method. To add, where could I find more info about this?

Thanks a lot!

 

[sergio mariotti]

TRL and SOLT are different calibration tecniques.
Almost in all cases you may choose the one you prefere.
For In-Wafer measurement, at low freq. (i.e. < 5 GHz) SOLT is more accurate than TRL.
At higher freq. (i.e. > 26 GHz) TRL may be more accurate than SOLT.

In the case you have, printed, ready to use, Cal standard on your substrate, use them in the possible manner (TRL or SOLT) and don't look for other substrates cal standard.

Generally TRL is more accurate than SOLT, but practically, most depend on the operator skilness.
TRL ask for 3 standards and 4 connections (omitting isolation)
SOLT ask for 4 standards and 7 connections (omitting isolation)
SOLT may start from DC
TRL has Fmax/Fmin<8

 

[oilytong]

Hi,

For TRL calibration, it requires a Thru, Reflection, and a line.

1) With regards to the line, how do you go about deciding on the length for the line?

2) For the Thru, is it just a transmission line of any arbitrary length?

Thanks and regards.

 

[sergio mariotti]

Hi,

For TRL calibration, it requires a Thru, Reflection, and a line.

1) With regards to the line, how do you go about deciding on the length for the line?

2) For the Thru, is it just a transmission line of any arbitrary length?

Thanks and regards.

Easy questions, long answer.
Telegraphically...
There are many combinations. The most widely adopted for lenghts are: Line= λ/4 (see note), Thru=0
Other choiches are:

1) Thru= A, Line= A+λ/4
In this case the loss of the may be corrected if the ref plane will be set by the reflect. In any case the loss of the "line" are often very small.

2)Consider also the TRM calibraion, just subsituting the Thru with a matched load.

note: in TRL calibration there are some ridondant informations; i.e you may decide that the algoritm will use the accurate lenght or accurate Zo. Rispectivelly Zo will not important and lenght will not important (but it must be close to λ/4).
How to decide for lenght or Zo? it depend on what VNA are you using. I may reply only about agilent 8722

 

[hcseu]

wish this application from agilent will help you

 

[sergio mariotti]

There are tons of Application notes, most by Agilent, i think is not the case to upload here.

Rather that, let's navigate on www.vnahelp.com , there are lots of links to very useful AN and lots of tick & trips non found on AN's.
The site is manteined by a very expert guru who work at Agilent

 

[hcseu]

to sergio mariotti
I can not enter the web,anything wrong?

 

[sergio mariotti]

to sergio mariotti
I can not enter the web,anything wrong?

I've checked now. it works.
**broken link removed**

 

[oilytong]

Hey...

Thanks all for the help.

I found this other website that has a list of the different documents too. Pretty helpful.

**broken link removed**

oily

 

[g579]

Just remarking, that some links are invalid in the above webpage!
g579

 

[loucy]

If you have high quality standards, SOLT is more accurate than TRL. TRL is involved with more assumptions that your measurement setup might not satisfy.

 

[yanziqqqqq]

There are many useful materials in Agilent, and you can make a search. Hope it would help.

 

[starchern]

how to set the TRL"reflector" parameter?

it include C0,C1,C2 and C3

 

[sergio mariotti]

how to set the TRL"reflector" parameter?

it include C0,C1,C2 and C3

When you edit, on your VNA, the Cal Kit Parameters, there are few sections:

On the Standard Definition Section, you may type in the Standard Data like C0, C1, C2,C3 for opens but also Offset Delay, Zo, Offs. Loss, etc....

While on the Class Assignament Section you may cross-assign each Standard to each type of Cal Measurement. In this Section you decide what are the Standard used for you Calibration (i.e. TRL).
In particular for TRL, take in account that the REFLECT must be the same for both the 2 cable ports.

--- Note that a PC software that help editing Cal Kit Parameters in a friendly manner exist. It's for free. But the website will be closed soon!!!
www.vnahelp.com