when "Thru" or "Line" standard was connected between the two ports, there are 6 traces on the screen,I know 4 of them are the S-parameter of the standard, but, what are the other two traces? what do they cantributed to the TRL calibration?

" when "Thru" or "Line" standard was connected between the two ports, there are 6 traces on the screen,I know 4 of them are the S-parameter of the standard, but, what are the other two traces? what do they cantributed to the TRL calibration? "

I'm guessing that you are seeing the extra R2/R1 and R1/R2 measurments needed to compute the delta match term used in the TRL cal.

" I'm guessing that you are seeing the extra R2/R1 and R1/R2 measurments needed to compute the delta match term used in the TRL cal. "

Thank you, Joel. you're right, the extra two traces are marked with R1/R2 and R2/R1 respectively.I'd like to know what 'R1' and 'R2' mean?Are they the reflection coefficient of internal reflectometer?Another question is what 'delta match term' is? I'm very interested in TRL calibration method.I'd like to learn more of the TRL calibration theory ,would you please explain it more detailedly? Thank you for your help!

To explain it in enough detail to fully understand is 10-20 pages of information, so I would suggest that you go to the Agilent Website and search for app notes with TRL in it. The fundamental paper was: G. Engen and C. ****, "Thru-Reflect-Line: An Improved Technique for Calibrating the Dual Six-Port Automatic Network Analyzer", IEEE Transactions on Microwave Theory and Techniques, December 1979.

With reference to your particular question: The TRL method of Engen and **** assume that the source and load match at a port is the same, that is, the match of a port when it is sourcing a signal is the same as when it is not sourcing a signal (that is, when it is a load, aka "load match"). For almost all modern VNAs, this assumption does not hold. However if you can measure the difference in match between when a port is sourcing and when it is acting as a load (which we call the "delta match" or change in match between sourcing and loading), you can modify the approach above to accomidate the delta match. The R2/R1 and R1/R2 measurements are essentially measuring the delta match.

To explain it in enough detail to fully understand is 10-20 pages of information, so I would suggest that you go to the Agilent Website and search for app notes with TRL in it. The fundamental paper was: G. Engen and C. ****, "Thru-Reflect-Line: An Improved Technique for Calibrating the Dual Six-Port Automatic Network Analyzer", IEEE Transactions on Microwave Theory and Techniques, December 1979.

With reference to your particular question: The TRL method of Engen and **** assume that the source and load match at a port is the same, that is, the match of a port when it is sourcing a signal is the same as when it is not sourcing a signal (that is, when it is a load, aka "load match"). For almost all modern VNAs, this assumption does not hold. However if you can measure the difference in match between when a port is sourcing and when it is acting as a load (which we call the "delta match" or change in match between sourcing and loading), you can modify the approach above to accomidate the delta match. The R2/R1 and R1/R2 measurements are essentially measuring the delta match.

" ...is acting as a load (which we call the "delta match" or change in match between sourcing and loading), you can modify the approach above to accomidate the delta match. The R2/R1 and R1/R2 measurements are essentially measuring the delta match. "

Taking your advice, I have read G. Engen and C. ****â€™s paper, itâ€™s really very helpful. But I still not fully understand the R1/R2 and R2/R1 terms. In G. Engen and C. ****â€™s paper the calibration method calculates T-matrix based on 8-term error model (assuming source and load match at a port is the same). Using this method, 7 independent error terms of error box A and error box B are solved, thus the 4 S-parameter can be corrected. As you said, in modern VNA source and load match at a port is not the same, â€œdelta matchâ€ is measured to find the differences. Is it meaning that using the â€œdelta matchâ€ we expand the 8-term model to 10-term model, and correct the S-parameter based on 10-term model? How is R1/R2 gotten? Another question is how does the modification (using R1/R2 and R2/R1) operated, and is there some correlative equations?

" Is it meaning that using the â€œdelta matchâ€ we expand the 8-term model to 10-term model, and correct the S-parameter based on 10-term model? How is R1/R2 gotten? Another question is how does the modification (using R1/R2 and R2/R1) operated, and is there some correlative equations? "

Yes, the delta match means "the vector difference between the source match at one port and the load match at the same port". Also, all the Agilent correction algorithms use a 12 term error model (10 terms as above, plus a forward and reverse crosstalk).

The R1/R2 with source at port 2, and R2/R1 with soruce at port 1 form a set of reflectormeters. If you consider R2 the signal from a reverse directional coupler that can measure the port 2 load match, then R2/R1 with source at port1 is measuring the load match of port 2 directly (well, not exactly directly, as the loss between the load element and the test port is not measured, but it is the same loss for the source match). When the TRL terms are evaluated, the measurement of the load match in this way is incorporated into the error term computation of the TRL calculation. I've gone through the calculation before, but I don't have it in front of me and it would take a bit of doing to re-formulate it.

There was an old 8510 paper by Doug Rytting that had the computation in it. Also, you might send a private message to "daveb" on this forum, as I believe he may have published a similar paper about the delta match correction, and he is our resident expert on calibration algorithm development.

NHello my name is Mouloud, i'm phd student in algeria contry. i would like your help about the TRL in practice, i.e, i realised the O4 stucures T, R, L and with transitor MACOM fixed in microstrip and coplanar. I do no how can I combine all S parametres meseaured by VNA 630 of the 04 structures to obtain only the S of the transistor (DUT).

Thanks a lot...

coordialy

M. CHALLALetwork Analyzer Error Models and Calibration Methods

" NHello my name is Mouloud, i'm phd student in algeria contry. i would like your help about the TRL in practice, i.e, i realised the O4 stucures T, R, L and with transitor MACOM fixed in microstrip and coplanar. I do no how can I combine all S parametres meseaured by VNA 630 of the 04 structures to obtain only the S of the transistor (DUT).

Thanks a lot...

coordialy

M. CHALLALetwork Analyzer Error Models and Calibration Methods "

I suspect by VNA 630 you mean 360. That would be a product from Anritsu. You could contact them for specific details. Agilent has an application note that may be useful for understanding the generalities. It is Specifying Calibration Standards and Kits for Agilent Vector Network Analyzers. Application Note 1287-11. It can be downloaded from the Agilent website.

If I understand you correctly, you have made measurements on the standards and want to use those measurements in combination with the measurements of the DUT to correct the DUT measurements. In theory that would work (provided you have all of the correct measurements), in practice calibration is a process that you do, then you make the measurements. The distinction being that the instrument firmware gathers the measurements it needs, computes the error terms and provides a calibrated measurement system. You then connect your DUT or DUTS and get error corrected measurements. Since you have your own standards you will need to provide the definition of the standards to the instrument--the application note will help you do that for Agilent network analyzers, but you will have to contact the VNA manufacturer for specifics for the VNA you are using.

I'm guessing that you are seeing the extra R2/R1 and R1/R2 measurments needed to compute the delta match term used in the TRL cal.

Thank you, Joel.

you're right, the extra two traces are marked with R1/R2 and R2/R1 respectively.I'd like to know what 'R1' and 'R2' mean?Are they the reflection coefficient of internal reflectometer?Another question is what 'delta match term' is?

I'm very interested in TRL calibration method.I'd like to learn more of the TRL calibration theory ,would you please explain it more detailedly? Thank you for your help!

Tina

G. Engen and C. ****, "Thru-Reflect-Line: An Improved Technique for Calibrating the Dual Six-Port Automatic Network Analyzer", IEEE Transactions on Microwave Theory and Techniques, December 1979.

There's additional information at:

http://www.vnahelp.com/tip18.html

With reference to your particular question: The TRL method of Engen and **** assume that the source and load match at a port is the same, that is, the match of a port when it is sourcing a signal is the same as when it is not sourcing a signal (that is, when it is a load, aka "load match"). For almost all modern VNAs, this assumption does not hold. However if you can measure the difference in match between when a port is sourcing and when it is acting as a load (which we call the "delta match" or change in match between sourcing and loading), you can modify the approach above to accomidate the delta match. The R2/R1 and R1/R2 measurements are essentially measuring the delta match.

G. Engen and C. ****, "Thru-Reflect-Line: An Improved Technique for Calibrating the Dual Six-Port Automatic Network Analyzer", IEEE Transactions on Microwave Theory and Techniques, December 1979.

There's additional information at:

http://www.vnahelp.com/tip18.html

With reference to your particular question: The TRL method of Engen and **** assume that the source and load match at a port is the same, that is, the match of a port when it is sourcing a signal is the same as when it is not sourcing a signal (that is, when it is a load, aka "load match"). For almost all modern VNAs, this assumption does not hold. However if you can measure the difference in match between when a port is sourcing and when it is acting as a load (which we call the "delta match" or change in match between sourcing and loading), you can modify the approach above to accomidate the delta match. The R2/R1 and R1/R2 measurements are essentially measuring the delta match.

Taking your advice, I have read G. Engen and C. ****â€™s paper, itâ€™s really very helpful. But I still not fully understand the R1/R2 and R2/R1 terms. In G. Engen and C. ****â€™s paper the calibration method calculates T-matrix based on 8-term error model (assuming source and load match at a port is the same). Using this method, 7 independent error terms of error box A and error box B are solved, thus the 4 S-parameter can be corrected. As you said, in modern VNA source and load match at a port is not the same, â€œdelta matchâ€ is measured to find the differences. Is it meaning that using the â€œdelta matchâ€ we expand the 8-term model to 10-term model, and correct the S-parameter based on 10-term model? How is R1/R2 gotten? Another question is how does the modification (using R1/R2 and R2/R1) operated, and is there some correlative equations?

Yes, the delta match means "the vector difference between the source match at one port and the load match at the same port". Also, all the Agilent correction algorithms use a 12 term error model (10 terms as above, plus a forward and reverse crosstalk).

The R1/R2 with source at port 2, and R2/R1 with soruce at port 1 form a set of reflectormeters. If you consider R2 the signal from a reverse directional coupler that can measure the port 2 load match, then R2/R1 with source at port1 is measuring the load match of port 2 directly (well, not exactly directly, as the loss between the load element and the test port is not measured, but it is the same loss for the source match). When the TRL terms are evaluated, the measurement of the load match in this way is incorporated into the error term computation of the TRL calculation. I've gone through the calculation before, but I don't have it in front of me and it would take a bit of doing to re-formulate it.

There was an old 8510 paper by Doug Rytting that had the computation in it. Also, you might send a private message to "daveb" on this forum, as I believe he may have published a similar paper about the delta match correction, and he is our resident expert on calibration algorithm development.

my name is Mouloud, i'm phd student in algeria contry. i would like your help about the TRL in practice, i.e, i realised the O4 stucures T, R, L and with transitor MACOM fixed in microstrip and coplanar. I do no how can I combine all S parametres meseaured by VNA 630 of the 04 structures to obtain only the S of the transistor (DUT).

Thanks a lot...

coordialy

M. CHALLALetwork Analyzer

Error Models

and

Calibration Methods

I suspect by VNA 630 you mean 360. That would be a product from Anritsu. You could contact them for specific details. Agilent has an application note that may be useful for understanding the generalities. It is Specifying Calibration Standards and Kits for Agilent Vector Network Analyzers. Application Note 1287-11. It can be downloaded from the Agilent website.

If I understand you correctly, you have made measurements on the standards and want to use those measurements in combination with the measurements of the DUT to correct the DUT measurements. In theory that would work (provided you have all of the correct measurements), in practice calibration is a process that you do, then you make the measurements. The distinction being that the instrument firmware gathers the measurements it needs, computes the error terms and provides a calibrated measurement system. You then connect your DUT or DUTS and get error corrected measurements. Since you have your own standards you will need to provide the definition of the standards to the instrument--the application note will help you do that for Agilent network analyzers, but you will have to contact the VNA manufacturer for specifics for the VNA you are using.

Hope this helps.

cordially