❤I once worked with a very smart guy (his brother won a MacArthur grant) and he told me that a key point in his college education came when he realized that he could go to the library and find other texts on the topic of a given class than the one the professor chose, and often one of those would be much clearer to him than the class text. This is a very profound insight and what IMSAI Guy is doing is presenting his view of this topic; I think this is an enormous service to the world and I wanted to call this out to people. “I can explain it to you, but I can’t understand it for you”, as another former cow orker liked to say. Let these concepts be the lamps of your eyes as you pursue understanding. Look at topics from as many angles as you can.
No video that is sharing knowledge is ever going to be too long in that wonderful digital world: one can pause and resume whenever one needs, how often one needs ;-)
This may have been a "long" video, but it was worth it. I had a number of "aha!" moments in this video and a subject that was totally confusing has transitioned to something that I now have a very basic understanding. (Very basic) I have much more to learn, but this was a very good start. Thank you
This video brought back memories of barely passing Linear Algebra. We had to use a basic 4 function calculator for exams, and I frequently got numbers mixed up when calculating sums and or products of larger matrices.
Thank you for this take on the material. While I had some abstract understanding of how use the S parameter measurements, it's always good to understand what they actually are, or mean. The math didn't hurt that much; like eating your vegetables, it's good for you even if you don't think so at the time.
Pretty thorough! I learned a lot here. I never ever heard of a full parameter test set, using a chain of couplers on the input and output side of the network under test, but it makes sense. I kinda view these devices in terms of optics, with semi-transparent mirrors that bounce the waveform entering from the output side off to a side port. Makes me wanna play with NanoVNA and RF systems for sure!
Very nice introduction and explanation. If in college the teacher would have started the course the way you did, I would have very much enjoyed it, now it was yet another boring subject to quickly master and pass the exam. In the next few days I start remembering the topic yet I mostly find those math skills more useful in a general sense. Thanks for yet another interesting vid :)
Scattering is a general physical process where one entity collides with another entity. The details of what comes out tells us a great deal about the interaction between the two entities. In this case we have rf waves interacting with an electronic device. It is also important to point out that the s-parameters are complex numbers (or functions of frequency) so that the time delay (sinusoidal phase shift) is included in this linear model. Harmonic generation and a few other possibilities cannot be modeled easily by linear systems.
The cal kit standards you use to calibrate the VNA's have theoretical values for Open S11=1, Load S11=0, Short S11=-1 and through S11=0, S21=1, S12=1, S22=0 (thats "thru" in the video is more of a "one way"). S-parameters are for voltages, not power, and therefore need impedance specified (like 50 Ohm). And they are only valid for linear systems like filters, but not i.e. mixers. If you measure devices with more than two ports on your 2 port VNA you must terminate all other ports. There is a industry standard file format for exchange called touchstone or SnP, you can find them i.e. on the Mini Circuits Website under S-Parameters. They are text files you can read with a text editor. You can import them into qucs and simulate some stuff beforehand building. I once have written a script for GNU Octave to measure N-Ports with a Nano VNA V2.
yes, S-parameters are linear only, transistors are linear over small changes. Keysight extends the math to include non-linear and calls it X-parameters.
You can measure S-parameters for nonlinear systems, but the numbers are not very useful. The numbers only describe the particular results for the one applied amplitude that was used during the test. Also, depending on the design of the test set, any energy produced in the device under test at any frequency other than the test frequency may be either ignored or may alter the measurement in undesirable ways. The results of such a measurement can't be scaled with confidence to any other amplitudes nor to any other combination of inputs nor cascaded with the S-parameters of another network to get overall S-parameters.
@@AB-xn6un There is the exception that if the non-linearity is not dependent on the signal, but other states of the device. RF-switches are one case, if they stay in their state (i.e. on or off) they are linear, however the switching between the states is a non-linear operation. Another case are linear amplifiers at different temperatures. Solution is to take separate S-parameters for the different states. In case of touchstone files, you end up with one for every state, and the state information are in the filename and comments of that file.
The S-parameter numbering system is counterintuitive...the way I remember what's happening is by saying "What happens at the first number as a result of what happens at the second." S12 - what happens at port 1 as a result of what happens at port 2. As you said you'd think it was the other way around but it's not! ;-)
Thank you this really helped me understand S-Parameters! RF is something that always confuses me and where my knowledge of electronics breaks down probably because I was a computer science major.
Idk how on earth his keeps being the case, but every time I start with a project you manage to produce a video exactly on the same subject and they're always incredibly helpful :-)
Perhaps consider exploring some of the never discussed and undocumented functions of the nanovna? The menu of functions has increased a lot, but there is no documentation or available demonstrations of how the functions are used.
Thank you for this! I have an opportunity to purchase an HP 8714ET (very similar to your 8711C). My understanding is it has transmission / reflection readings, but not S parameters. In your opinion, am I losing much by getting the ET version vs. the ES version?
I don't know much. the ES has full switching to measure both S11 and S22 without making two measurements and flipping the part. not a big deal for hobby work. I don't know if there are any other differences.
@@IMSAIGuy thank you so much for your reply. I really enjoy your videos, you have a great nack for taking really complex subjects and making them easy digestible for amateurs like myself.
I liked the video, but I got lost with the matrix. One question about the s-parameter labels. Does the first number after the S represent the output and the second number after the S represent the input? I think that I got confused with transferring the inputs and outputs to the matrix. Could you do another video about matrices? Thank you for making these videos.
I’m curious how nanoVNA’s implement a return loss bridge that operates over such a broad frequency range. May make an interesting topic for a future video.
Наука
❤I once worked with a very smart guy (his brother won a MacArthur grant) and he told me that a key point in his college education came when he realized that he could go to the library and find other texts on the topic of a given class than the one the professor chose, and often one of those would be much clearer to him than the class text. This is a very profound insight and what IMSAI Guy is doing is presenting his view of this topic; I think this is an enormous service to the world and I wanted to call this out to people. “I can explain it to you, but I can’t understand it for you”, as another former cow orker liked to say. Let these concepts be the lamps of your eyes as you pursue understanding. Look at topics from as many angles as you can.
I learned something new today, even better, I now understand the S parameters more than I did.
Thanks for an interesting lesson.
No video that is sharing knowledge is ever going to be too long in that wonderful digital world: one can pause and resume whenever one needs, how often one needs ;-)
You answered a whole bunch of questions I had left over from my Extra Class examination. Now I want more. Thanks and that was a lucid explanation.
This may have been a "long" video, but it was worth it. I had a number of "aha!" moments in this video and a subject that was totally confusing has transitioned to something that I now have a very basic understanding. (Very basic) I have much more to learn, but this was a very good start. Thank you
I had the opportunity to work around guys like Doug Rytting, Julius Botka, and others at HP in the 90s, what an education!
This video brought back memories of barely passing Linear Algebra. We had to use a basic 4 function calculator for exams, and I frequently got numbers mixed up when calculating sums and or products of larger matrices.
I think I have a headache now😀 Thanks for the informative video you increased my meager knowledge greatly.
Thank you for this take on the material. While I had some abstract understanding of how use the S parameter measurements, it's always good to understand what they actually are, or mean. The math didn't hurt that much; like eating your vegetables, it's good for you even if you don't think so at the time.
Pretty thorough! I learned a lot here. I never ever heard of a full parameter test set, using a chain of couplers on the input and output side of the network under test, but it makes sense. I kinda view these devices in terms of optics, with semi-transparent mirrors that bounce the waveform entering from the output side off to a side port.
Makes me wanna play with NanoVNA and RF systems for sure!
Very nice introduction and explanation. If in college the teacher would have started the course the way you did, I would have very much enjoyed it, now it was yet another boring subject to quickly master and pass the exam. In the next few days I start remembering the topic yet I mostly find those math skills more useful in a general sense. Thanks for yet another interesting vid :)
Scattering is a general physical process where one entity collides with another entity. The details of what comes out tells us a great deal about the interaction between the two entities. In this case we have rf waves interacting with an electronic device.
It is also important to point out that the s-parameters are complex numbers (or functions of frequency) so that the time delay (sinusoidal phase shift) is included in this linear model. Harmonic generation and a few other possibilities cannot be modeled easily by linear systems.
I did mention complex: 24:42
Linear algebra was my best subject 2nd year university. Love matrices!
The cal kit standards you use to calibrate the VNA's have theoretical values for Open S11=1, Load S11=0, Short S11=-1 and through S11=0, S21=1, S12=1, S22=0 (thats "thru" in the video is more of a "one way").
S-parameters are for voltages, not power, and therefore need impedance specified (like 50 Ohm). And they are only valid for linear systems like filters, but not i.e. mixers.
If you measure devices with more than two ports on your 2 port VNA you must terminate all other ports.
There is a industry standard file format for exchange called touchstone or SnP, you can find them i.e. on the Mini Circuits Website under S-Parameters.
They are text files you can read with a text editor. You can import them into qucs and simulate some stuff beforehand building.
I once have written a script for GNU Octave to measure N-Ports with a Nano VNA V2.
Are you sure about S-parameters only being valid for linear systems? Can't transistors also be characterised by S-parameters?
yes, S-parameters are linear only, transistors are linear over small changes. Keysight extends the math to include non-linear and calls it X-parameters.
You can measure S-parameters for nonlinear systems, but the numbers are not very useful. The numbers only describe the particular results for the one applied amplitude that was used during the test. Also, depending on the design of the test set, any energy produced in the device under test at any frequency other than the test frequency may be either ignored or may alter the measurement in undesirable ways. The results of such a measurement can't be scaled with confidence to any other amplitudes nor to any other combination of inputs nor cascaded with the S-parameters of another network to get overall S-parameters.
@@AB-xn6un There is the exception that if the non-linearity is not dependent on the signal, but other states of the device. RF-switches are one case, if they stay in their state (i.e. on or off) they are linear, however the switching between the states is a non-linear operation. Another case are linear amplifiers at different temperatures. Solution is to take separate S-parameters for the different states. In case of touchstone files, you end up with one for every state, and the state information are in the filename and comments of that file.
The S-parameter numbering system is counterintuitive...the way I remember what's happening is by saying "What happens at the first number as a result of what happens at the second." S12 - what happens at port 1 as a result of what happens at port 2. As you said you'd think it was the other way around but it's not! ;-)
Thank you this really helped me understand S-Parameters! RF is something that always confuses me and where my knowledge of electronics breaks down probably because I was a computer science major.
“This is expensive”, I had to laugh, Ya buddy you ain’t kidding… Nice vid, Thx.
Idk how on earth his keeps being the case, but every time I start with a project you manage to produce a video exactly on the same subject and they're always incredibly helpful :-)
Google shares all of your data for us RUclips contributors.
just kidding 😎
Thank you very much
Perhaps consider exploring some of the never discussed and undocumented functions of the nanovna? The menu of functions has increased a lot, but there is no documentation or available demonstrations of how the functions are used.
Very useful. I always struggled with vna parms
very helpful
MERCI! A great video, thanks a lot.
Thank you for this! I have an opportunity to purchase an HP 8714ET (very similar to your 8711C). My understanding is it has transmission / reflection readings, but not S parameters. In your opinion, am I losing much by getting the ET version vs. the ES version?
I don't know much. the ES has full switching to measure both S11 and S22 without making two measurements and flipping the part. not a big deal for hobby work. I don't know if there are any other differences.
@@IMSAIGuy thank you so much for your reply. I really enjoy your videos, you have a great nack for taking really complex subjects and making them easy digestible for amateurs like myself.
My HP8753B runs to 6GHz, the setup at work runs to over 20GHz 🙂 I love my toys 🙂
I liked the video, but I got lost with the matrix. One question about the s-parameter labels. Does the first number after the S represent the output and the second number after the S represent the input? I think that I got confused with transferring the inputs and outputs to the matrix. Could you do another video about matrices? Thank you for making these videos.
The number is just row, column
“The wiggle”… A nice scientific term. 😊
I’m curious how nanoVNA’s implement a return loss bridge that operates over such a broad frequency range. May make an interesting topic for a future video.
on a nano you might have 401 points of frequency. each point is separately calibrated.
Palomar Engineers really needs to correct their units in the catalogs. Really hard for me to take them seriously when they flub the basic SI.
Very confusing!