What are the switch terms?

1,234 views
Skip to first unread message

Michael Wollensack METAS

unread,
Dec 1, 2014, 11:06:51 AM12/1/14
to vnat...@googlegroups.com
  1. What are the switch terms?
  2. Which standard (Thru, Match or High Reflect) should I measure?
  3. Is it necessary to have those switch terms for my R&S VNA?

Michael Wollensack METAS

unread,
Dec 1, 2014, 12:19:26 PM12/1/14
to vnat...@googlegroups.com
Hi,

  1. The switch terms describe the imperfect load when the source is switch to the other port.
  2. A thru connection or an adapter with low insertion loss.
  3. The switch terms are necessary for all VNA with more than one port.

Regards
Michael

Ondřej Morávek

unread,
Dec 2, 2014, 5:30:40 AM12/2/14
to vnat...@googlegroups.com
Hi,

is the point 3. applicable also for the four-sampler VNAs without the source switch? I think at least the 4-port ZVA67 with K4 option certainly has 4 separate generators and with sampler for the reference wave quantity. Additionally, If I perform first-tier SOLT beforehand and then measure the (SOLT) corrected S-parameters which would be used later for secondary TRL calibration as raw data, I assume the switch terms aren't necessary in this case.

Regards
Ondra


Michael Wollensack METAS

unread,
Dec 2, 2014, 7:46:53 AM12/2/14
to vnat...@googlegroups.com
Hi Ondra,

yes. You can only measure the switch terms if you have 2N receivers for a N-Port VNA. If a port is in receive mode (source is driving another port) then there is an imperfect load at the end which is described by the switch term.

Two-Tier calibration:
Your assumption is right, but you can not use error corrected data from the first-tier SOLT calibration as raw data input for the second-tier TRL calibration in VNA Tools. The reason for it. You can measure the switch terms and compute directly the TRL calibration.

Regards
Michael

tbu...@maurymw.com

unread,
Sep 21, 2016, 8:18:31 PM9/21/16
to VNA Tools
Hello Michael,

You said that the switch terms must be a thru connection or an adapter with low insertion loss. However, when I look in the example folder of VNA tools II called "01_Basic_Example_N(f-m)" and display the S parameters of the so called "Switchterms_01.sdatb" file in the data explorer, S21 and S12 magnitude are both extremely small indicating no transmission between ports 1 and 2. Why is that?

Michael Wollensack METAS

unread,
Sep 22, 2016, 10:34:24 AM9/22/16
to VNA Tools
Hi,

the switch terms are stored as sdatb file inside VNA Tools II. The a look at the Switch Terms in the raw data folder 'Measurements_01'.

S11: a1/b1_p2, switch term of port 1 when port 2 drives
S21: 0
S12: 0
S22: a2/b2_p1, switch term of port 2 when port 1 drives

the switch terms should be measured with a thru connection.

> ... "Switchterms_01.sdatb" file in the data explorer, S21 and S12 magnitude are both extremely small indicating no transmission between ports 1 and 2. Why is that?

I'm assuming your looking in the error corrected folder at the switch term file. The error corrected file of the switch terms makes no sense at all. Take a look in the raw measurement folder. For the error corrected S-parameters of the Thru take a look at Thru_01.sdatb in the error corrected folder.

Hope this helps.

Regards
Michael

Tasnim Khanam

unread,
Sep 25, 2016, 2:52:01 PM9/25/16
to VNA Tools
Hello Michael,

Would you be please elaborate the part 
"S11: a1/b1_p2, switch term of port 1 when port 2 drives
S21: 0
S12: 0
S22: a2/b2_p1, switch term of port 2 when port 1 drives"    where S21=0. I don't understand why that happens?


-Thank You,
Tasnim


Michael Wollensack METAS

unread,
Sep 26, 2016, 3:57:05 AM9/26/16
to vnat...@googlegroups.com
Hi,

SwitchTerms_01.sdatb (contanins the switch terms of port 1 and 2 stored as S-parameter S11 and S22. S21 and S12 contains no data)
S11: a1/b1_p2, switch term of port 1 when port 2 drives
S21: 0, zero by definition
S12: 0, zero by definition
S22: a2/b2_p1, switch term of port 2 when port 1 drives

Thru_01.sdatb (contains the S-parameter of the trough connection)
S11: b1/a1_p1
S21: b2/a1_p1
S12: b1/a2_p2
S22: b2/a2_p2

Both files ('SwitchTerms_01.sdatb' and 'Thur_01.sdatb') are measured using a through connection between port 1 and 2.

Regards
Michael

Dao Pham

unread,
Sep 29, 2017, 9:11:20 AM9/29/17
to VNA Tools

Hi, 


I have a question about the switch-term corrected S-parameters for the 4-port VNA measurement.

In fact, with the 2-port VNA model, we can measure 2 switch terms (in forward and reverse directions) then apply the switch-term correction as following:

In other way, we can apply the switch terms to the measured S-matrix (Sm) under the matrix manipulation as below:

For the switch-term correction on a 4-port VNA system, I’m thinking about applying the same matrix manipulation as the 2-port VNA case

But the problem is that I haven’t found down the references/publications to prove this approach.

I would be grateful if any experts here could give me some ideas or references regarding to taking all switch-terms into account for the S-parameter measurements with 4-port VNA.


Thank you in advance for your help.


Best Regards,

Dao PHAM




Michael Wollensack METAS

unread,
Oct 4, 2017, 8:40:43 AM10/4/17
to vnat...@googlegroups.com
Hi,

your posted equation looks allright. 



I've some additional comment. 
Γ1,2 Γ1,3 and T1,4 are the switch terms of port 1 when another port is driving. They should be all the same. 

How to come up with this equation, see the following steps:

Sm = B x A^(-1) = B x Adiag^(-1) x C^(-1) = S'm x C^(-1)

and

C = A x Adiag^(-1)

The matrix C is similiar to your matrix above with the switch terms.

For a four port VNA it would look like that:



It can be shown that the ratio a_px/b_px is independent of x. That means you've one switch term per port.

I remove the switch terms in VNA Tools with the following code:

        /// <summary>

        /// Remove Switch Terms from N-Port

        /// </summary>

        /// <typeparam name="T">Real Number Type</typeparam>

        /// <param name="a">N-Port S-Parameter Data</param>

        /// <param name="swt">Switch Terms</param>

        /// <returns>N-Port S-Parameter Data</returns>

        public static SParamData<T> RemoveSwitchTerms<T>(SParamData<T> a, SParamData<T> swt)

            where T : IRealNumber<T>, new ()

        {

            int nports = a.NPorts;

            int nfreq = a.NFreq;

            if (nports != swt.NPorts) throw new System.Exception("Number of ports must agree");

            if (nfreq != swt.NFreq) throw new System.Exception("Number of frequency points must agree");

            SParamData<T> swt2 = ChangeZr(swt, a.PortZr);

            SParamData<T> b = new SParamData<T>(a.Frequency, a.Ports, a.PortZr);

            Parallel.For(0, nfreq, (int i0) =>

            //for (int i0 = 0; i0 < nfreq; i0++)

            {

                Complex<T>[][] temp = new Complex<T>[nports][];

                for (int i1 = 0; i1 < nports; i1++)

                {

                    temp[i1] = new Complex<T>[nports];

                    for (int i2 = 0; i2 < nports; i2++)

                    {

                        if (i1 == i2)

                            temp[i1][i2] = 1;

                        else

                            temp[i1][i2] = a[i0, i1, i2] * swt2[i0, i1, i1];

                    }

                }

                //b[i0] = LinAlg.Dot(a[i0], LinAlg.Inv(temp));

                b[i0] = LinAlg.Dot_A_invB(a[i0], temp);

            });

            return b;

        }


My switch terms are stored in a diagonal matrix.

Dao Pham

unread,
Oct 4, 2017, 9:49:02 AM10/4/17
to VNA Tools
Thank you very much for your kind and detailed reply.

Mertcan ASLAN

unread,
May 28, 2024, 2:05:25 AMMay 28
to VNA Tools
Hi

I am trying to do switch terms measurement here (see pic2)

But the message appears (see pic1)

I can measure S1,1  S2,1 and others but not switch terms measurement.

Can i have your help?

My VNA device is HP8753D with 2 port.

Best regards,

Mert





1 Aralık 2014 Pazartesi tarihinde saat 18:06:51 UTC+2 itibarıyla Michael Wollensack METAS şunları yazdı:
pic2.png
pic1.png

Michael Wollensack METAS

unread,
May 28, 2024, 2:41:54 AMMay 28
to VNA Tools

Hi,

 

with a HP8753D you cannot measure the switch terms, because it has only 3 receivers and not 4.

 

Therefor you can only perform only SOLT calibrations and not Unknown Thru for example.

 

Regards

Michael

Message has been deleted

Mertcan ASLAN

unread,
May 28, 2024, 10:01:21 AMMay 28
to VNA Tools
Hi,

Thanks a lot...

And what are the consequences of this limitation for me? Is there any alternative way to go beyond of this limitation?

Or  basicly can i use SOLT calibration for switch terms measurement? 

And how can i take Thru measurement with SOLT calibration?

I am very new to calibration measurements. So, sorry if my questions are nonsensical. I just want to know my device's capabilities&limitations and then i am planning to get education for VNA Tools but before of that i am trying to learn as much as i can do by myself. Do you provide any education?

I got a lot information from this group already by reading. Thanks for your support here.

Best regards,

Mert

28 Mayıs 2024 Salı tarihinde saat 09:41:54 UTC+3 itibarıyla Michael Wollensack METAS şunları yazdı:

Juerg Ruefenacht

unread,
May 31, 2024, 7:30:48 AMMay 31
to VNA Tools

Hi Mert,

This is are very good reason to attend our VNA Tools training course including the additional free VNA Expert Day (see:  
METAS courses ).

Unfortunately, our workload doesn't allow us to go into more detail, especially since VNA Tools is available for free. But there are many textbooks available, or check out the free resources on the internet.
Our goal is to keep the VNA Tools as a free software and better visibility with a larger number of course registrations help us to get support from our management to allocate the necessary resources. Therefore, your support will help us that we can further develop and improve our free VNA metrology application VNA Tools.

Thanks,
Juerg


Mertcan ASLAN

unread,
Jun 4, 2024, 2:51:31 AMJun 4
to VNA Tools
Hi

You guys are doing a great job. 

Keep it up!

Thanks to your work and support here. 

Have a great time.

Best regards,

Mert

31 Mayıs 2024 Cuma tarihinde saat 14:30:48 UTC+3 itibarıyla juerg.ru...@metas.ch şunları yazdı:
Reply all
Reply to author
Forward
0 new messages