Uff - another video that's fast and loose with terminology :-s You should have given the definition of Z0: the characteristic impedance of a transmission line _under the condition of a matched load_ (or equivalently - an infinitely long line, where the reflection never returns). Then that immediately shows why terminating in ZL will give a different Zin if ZL Z0: by definition. Also, you could have spent a sentence to show that R and especially G turn out to be very small compared to j*2pi*f*C, such that it simplified to Z0 -> sqrt(L/C) (with is a Real number) for almost all PCB materials >1MHz. NB using the term "ohmmeter" for an impedance measurement device is going to confuse most newcomers, just call it an impedance meter.
Hi Laurens, Thanks for taking the time to watch. I'm sure you can understand that with bite-size videos, we only have so much time to cover all of these concepts. We are producing a series that addresses everything you brought up, and much more. Next on the list is termination and input impedance. Regarding the transmission line characteristic impedance, I have described this several times in terms of the geometry directly on Altium's website and in several peer-reviewed papers on the topic. Not to self-promote, but you can watch one of my IEEE or EDICON presentations on the topic (see my channel) if you want to learn about transforming lossless analytical models into lossy analytical models that include dispersion and roughness. This is the mathematically correct way to describe transmission line impedance and, as you'll see from wideband optimization results I perform in those videos, the impedance does NOT simplify to Z0 -> sqrt(L/C), and it is certainly NOT a real number. I even show that commercial field solvers get it wrong precisely because they take the approach you list in your comment. We have an upcoming video on this as well. You can also read this blog on the topic: resources.altium.com/p/modeling-copper-foil-roughness-altium-designers-impedance-profiler-0 Regarding ohmmeter vs. impedance meter, I'm sure you're aware that there is no impedance meter (at least not in the sense of measuring transmission line impedance). Input impedance in an interconnect or device under test (DUT) is sometimes measured with a TDR measurement, followed by converting the response to a transfer function and calculating the impedance spectrum from that. Or you calculate it using the input impedance determined from an S11 measurement using a VNA. I'm waiting for my NanoVNA to come in the mail before we start filming this!
6:12 I understand how it works for high frequencies but.... Z + ZL is true for very low frequencies? I am confused because as I understand it Z consists mainly from inductors and capacitors and they are negligible at low frequency so the sum would be not true, it would be just Z. I know this is not very important here as this video is about something else.
Please elaborate: If there is only ONE forumula for impedance calculation, why different impedance calculators' results are not exactly same and have varying degree of accuracy?
Hi Raza, I'm assuming you're referring to transmission lines, and there are multiple formulas that are used for impedance calculations. The IPC-2141 formula is known to be accurate are low frequencies, but it becomes inaccurate at higher frequencies. If you read Brian C. Waddel's Transmission Line Design Handbook, you'll find many different formulas for trace impedance. Calculators online give different impedance because some will use IPC-2141, and some will use Waddel. Some won't tell you what they use. Others will apply an etch factor, and some won't. These are some of the reasons you will get different results in different calculators. Field solvers can give more accurate results, but even field solvers don't always get a perfect calculation. Hope this provide some context! I'll be talking more about this in an upcoming video. Thanks for watching!
At DC voltage, you only have the DC resistance of the transmission line. So you just use the standard formula for resistance with a square-shaped trace. There are online calculators that can do this for you.
Hello, thanks for watching! What I meant was that you can't just sum up the transmission line impedance and the load impedance to get an equivalent impedance. This is because the formulas you use to calculate an equivalent impedance of a circuit assume that the electrical signal exists everywhere throughout the circuit. This is not the case with transmission lines and with real circuits: the signal requires some time to travel between two points, so it might only interact with just the transmission line or just the load at any point in time. Hope this helps!
I think a better topic would be the why and how of transmission lines. In a digital system, frequency is a complex issue. A square wave consists of many frequencies, not just one., how should we be thinking about this? What is the effect of reflections that can occur when a transmission line is not properly terminated? What effect does the use of transmission lines have on EMI? What effects are there on crosstalk? ETC. What does the energy pattern from a transmission line look like? How does that compare to a differential pair, for example? Practical applications are more useful than definitions.
Uff - another video that's fast and loose with terminology :-s
You should have given the definition of Z0: the characteristic impedance of a transmission line _under the condition of a matched load_ (or equivalently - an infinitely long line, where the reflection never returns). Then that immediately shows why terminating in ZL will give a different Zin if ZL Z0: by definition.
Also, you could have spent a sentence to show that R and especially G turn out to be very small compared to j*2pi*f*C, such that it simplified to Z0 -> sqrt(L/C) (with is a Real number) for almost all PCB materials >1MHz.
NB using the term "ohmmeter" for an impedance measurement device is going to confuse most newcomers, just call it an impedance meter.
Hi Laurens,
Thanks for taking the time to watch. I'm sure you can understand that with bite-size videos, we only have so much time to cover all of these concepts. We are producing a series that addresses everything you brought up, and much more. Next on the list is termination and input impedance.
Regarding the transmission line characteristic impedance, I have described this several times in terms of the geometry directly on Altium's website and in several peer-reviewed papers on the topic. Not to self-promote, but you can watch one of my IEEE or EDICON presentations on the topic (see my channel) if you want to learn about transforming lossless analytical models into lossy analytical models that include dispersion and roughness. This is the mathematically correct way to describe transmission line impedance and, as you'll see from wideband optimization results I perform in those videos, the impedance does NOT simplify to Z0 -> sqrt(L/C), and it is certainly NOT a real number. I even show that commercial field solvers get it wrong precisely because they take the approach you list in your comment. We have an upcoming video on this as well. You can also read this blog on the topic: resources.altium.com/p/modeling-copper-foil-roughness-altium-designers-impedance-profiler-0
Regarding ohmmeter vs. impedance meter, I'm sure you're aware that there is no impedance meter (at least not in the sense of measuring transmission line impedance). Input impedance in an interconnect or device under test (DUT) is sometimes measured with a TDR measurement, followed by converting the response to a transfer function and calculating the impedance spectrum from that. Or you calculate it using the input impedance determined from an S11 measurement using a VNA. I'm waiting for my NanoVNA to come in the mail before we start filming this!
You are the best..!! you explain things very clearly.!! thanks..!!
And how the software can calculate the characteristic impedance just by changing the width of the trace ..? and why is it independent of frequency ..?
6:12 I understand how it works for high frequencies but.... Z + ZL is true for very low frequencies? I am confused because as I understand it Z consists mainly from inductors and capacitors and they are negligible at low frequency so the sum would be not true, it would be just Z. I know this is not very important here as this video is about something else.
Please elaborate: If there is only ONE forumula for impedance calculation, why different impedance calculators' results are not exactly same and have varying degree of accuracy?
Hi Raza, I'm assuming you're referring to transmission lines, and there are multiple formulas that are used for impedance calculations. The IPC-2141 formula is known to be accurate are low frequencies, but it becomes inaccurate at higher frequencies. If you read Brian C. Waddel's Transmission Line Design Handbook, you'll find many different formulas for trace impedance. Calculators online give different impedance because some will use IPC-2141, and some will use Waddel. Some won't tell you what they use. Others will apply an etch factor, and some won't. These are some of the reasons you will get different results in different calculators. Field solvers can give more accurate results, but even field solvers don't always get a perfect calculation.
Hope this provide some context! I'll be talking more about this in an upcoming video. Thanks for watching!
@@Zachariah-Peterson Thanks! waiting for your next video.
You're awesome Zach
How you calculate z when you are operating on dc voltage?
At DC voltage, you only have the DC resistance of the transmission line. So you just use the standard formula for resistance with a square-shaped trace. There are online calculators that can do this for you.
Did you mean if there is a Transmission line there will be no equivalent impedance?
Hello, thanks for watching! What I meant was that you can't just sum up the transmission line impedance and the load impedance to get an equivalent impedance. This is because the formulas you use to calculate an equivalent impedance of a circuit assume that the electrical signal exists everywhere throughout the circuit. This is not the case with transmission lines and with real circuits: the signal requires some time to travel between two points, so it might only interact with just the transmission line or just the load at any point in time. Hope this helps!
I think a better topic would be the why and how of transmission lines. In a digital system, frequency is a complex issue. A square wave consists of many frequencies, not just one., how should we be thinking about this? What is the effect of reflections that can occur when a transmission line is not properly terminated? What effect does the use of transmission lines have on EMI? What effects are there on crosstalk? ETC. What does the energy pattern from a transmission line look like? How does that compare to a differential pair, for example? Practical applications are more useful than definitions.
Hi Robert,
Thanks for taking time to watch! We have an entire series on these topics coming up, so stay tuned!
Very nice topic!!!!
Thank you!