Cheers from Sweden. I watched over 10 hour of your videos during christmas in order to design a PCB that got me the highest grade possible in my PCB-design course :) Thank you for great content.
50 Ohms comes from the 30s and early use of coax cable where low loss in the cable was key. With the given cable properties back then it happened that at 50 Ohms the lowest attenuation (series resistance) was found.
@@fabianluttenberger7153 Correction: Minimum loss:77ohm Max power:30 ohm 50 ohm is a compromise between this two. It happens in solid dielectric coax minimum loss is 50ohm.
@@DavidLopez-bz4rj That is the exact reason why TV antenna systems uses 75Ohm as you need minimal attenuation without the need to delivering whole lot of powers (for Rx), whereas HAM uses 50ohm for the obvious reason to Tx high power.
60 years plus working on lightbulbs(overhead projectors) to linear accelerators and almost everything electrical in between I have never seriously considered this aspect of design. Thank you for an epiphany moment. This why I UA-cam. You know I'm going to subscribe. Thanks again.
I remember from microwave class that at higher frequencies with mismatched line impedance, a portion of signal would reflect. For max power transfer you'd always try to have same line impedance
YES, I remembered I have read related statements on why 50 ohms and why 50 ohms is so common on Eric's book, Signal and Power integrity 3rd. It's all about trade-off and 50 ohms is a good point to start tuning if need. Also there're some history and background inside too. Again, thanks. This video just brings some memory from my mind. It's good for me.
Lee Ritchey has this great story about how he was working on a submarine SONAR system and the electronic design engineers insisted on 70 ohm transmission lines. I think it had to do with a concern that the drivers wouldn't be able to switch 50 ohm transmission lines quickly enough. Well because of the increased dielectric thickness to get 70 ohms they had terrible cross talk. Distance from the reference plane is a key consideration for cross talk and cross talk is directly proportional to dielectric thickness. Lee convinced them to lower the impedance to 50 ohms and they were able to actually build a functioning system. That night after the second or third day of his course on making it to 56 Gbps when he told that story (we must have been talking about crosstalk) I flipped on TV and the Hunt for Red October was on. In the movie the Soviet Union had built a sub that had a silent drive system. The only was it could be tracked is with a very specific SONAR signature. So the Americans chase down the Soviet sub using SONAR in a Los Angels class US submarine, the very one Lee worked on. What a coincidence! Lee speaks very highly of Eric Bogatin. I notice Eric referenced "loosely coupled" differential pairs. Taking the words right out of Lee's mouth. Not saying it was Lee's idea, but he stresses tightly coupled differential pairs are not a good idea.
really liked this video. it answered a question i haven't even had to ask yet.(if that makes sense, im still learning ya see,) so now i will know the answer to that 50 OMH impedance. coz it would have come up. SO IM VERY HAPPY TO HAVE SEEN THIS VIDEO. THANK YOU MR. ROBERT FERANEC..
Great stuff! I have learned quite a lot from watching your channel! I love that you can explain everything in a easy to understand manner, and that your guests on the show do it as well.
thanks a lot, the short videos are more engaging. i could never watch full long videos although i know you are sharing great content and im missing a lot.
The impedance is the characteristic impedance (the ratio of voltage to current of a wave traveling in one direction without loss). 50 ohms is an arbitrary standard that is commonly used in rf and microwave transmission lines to prevent signals from reflecting off impedance discontinuities.
Hi, 50 ohm is not an arbitrary number. A coaxial cable of 30 ohm characteristic impedance would handle more power. A 75 ohm cable would have less attenuation with VHF and UHF frequencies. And 50 is the arithmetic and geometric mean of the two numbers. So, the 50 ohm becomes the standard for the amateur radio and the like and the. 75 ohm for the video industry. De VU2RZA
Another point of using 50 Ohms for the PCB layouts is the following: it's well known from the theory that the coax cable, which is widely used for the PCBs interconnections has (1) maximum power transport (for the fixed cross-section) if the impedance is = 90 Ohms: (2) Has minimum attenuation if the impedance is = 75 Ohms, and, finally, (3) Has minimum signal linear distortions (thus - maximum bandwidth) if the impedance is = 50 Ohms: this is the most important condition for the functioning of PCB.
75 ohms characteristic impedance corresponds to nearly minimum transmission loss, 30 ohms gives maximum power handling capability in Coaxial cables and 50 ohms is a rounded compromise number.
I think it all comes from TTL (Transistor Transistor Logic) from the 1960's which was 5V. TTL = 5v, I=V/R=5/50=100mA.. Easy numbers to work with. Same resistance at each end of a line for maximum power transfer. Industrial controls use 500 Ohms at 10V =20mA for 0-10V, 0-20mA, 4-20mA. 4-20mA being as 16mA range scales well into 16384=2^14+sign+parity=16 bits so in and integer register about 1000=1mA in 14bit
Radio transmitters often use 50 ohms, TV receivers use 75 ohms. It is likely that transmitters must transmit more power. With TV, it is important to transport low power cheaply.
Im currently facing a roadblock in my PCB design: I need a 50 ohm trace but I cant have a trace wider than 20mils (too thin) because of the 0.65mm pitch bga the signal comes from how do people deal with that . Thank you for all your amazing content its incredibly precious
One thing is to choose a layer stackup that minimizes the thickness of the prepreg between your signal layer and the ground plane below: thinner dielectric -> narrower traces for a given impedance. Other than that, you just have to neck down the traces eventually.
50ohm lines on FR4 are quite wide, especially on 0.062" material (0.105" wide). I can't imagine anyone using 50ohm lines unless you were working with 50ohm components.
Noobie question. Isn't it 50 ohms over a specific length? Wouldn't the impedance double if you double the length of the cable? Or, am I completely misunderstanding?
It doesn’t depend on the length of the cable. The impedance of a transmission-line is a result of the cable’s internal capacitance and inductance. It’s what a driver “sees” as a load when trying to change the voltage on the cable, and create a wavefront that moves down the cable. At the receiving end, you want a load resistance that matches the cable impedance. Otherwise, some portion of the wavefront will reflect back to the source. Also, at the driving end, you want the driver’s resistance to match the cable impedance. If not, the wavefront could easily keep on reflecting back and forth for several end-to-end trips.
If you add length, you have also added inductance which exactly balances the increased capacitance. The increase capacitance would tend to *reduce* the impedance (equivalent resistance) where the added inductance increases the impedance. *Adding length changes loss* that is the only thing that increases with added length. The wire has actual resistance, not much but some, and the dielectric also causes a bit of loss as the electric field causes electrons to move slightly and that absorbs a bit of power.
@@thomasmaughan4798 Aha, now I think I'm getting it, thank you. Presently my struggle is how to understand that, knowing the way the two aspects of reactance exactly cancels when we do phasor diagrams. If it does cancel, why does it even matter? How do you measure the impedance rating of a cable?
Thanks to all of you who tried to help me understand this. I spent the morning coming to a solution. For noobies like me with similar questions, you'll find help in a Wikipedia page "Telegrapher's equations", where they explain the length is assumed to be an infinitely long line terminating in a ground, with no confusing back reflections. The capacitance does not completely cancel the inductance because the inductance is in series, the capacitance is in parallel (forehead smack). It is humbling to realize this was all worked out, over 150 years ago.
Thank you Ben. I moved things around for the AltiumLive recording. I like it too ... I may use it more often, just need to figure out how to do recording when I am doing something on PC (in this setup I can't have monitor in the front of me)
Hello.. Is it necessary to keep 50 Ohm impedance for all Traces .. If we are designing any logical operations with 74HC595D(SPI Mode) is it necessary to match all the Input trace impedance! Kindly give a clarity between Hi speed design and low speed design.. When will be system considered as High speed design!? Any threshold
He mentions that higher impedance lowers current, thus lowering power consumption on other elements. why doesnt the power not dissipate on the 50 ohm impedance?
I have been wondering why 50 ohms is used for so long, even in the mathmatics when they use normalized impedance of 50 ohms to then figure out of things I was always wondering but why 50?
I have a question - how come trace impedance calculators only relay on dielectric thickness / stackup and line width? What about the trace length? How is it possible that I will get 50 Ohms with X thickness and with Y width, no matter the length?
"Where did we ever come up with 50ohms" Lol, that's a question where you'll never ever get a straight answer. Another question where you never get a straight answer is assuming you chose 50 ohms as your line impedance what makes anyone assume that the impedance of the driver and source devices are anywhere near 50 ohms? For RF designs the impedance is presented to you in form of S-parameters with a whole slew of equipment for determining how things are matched. For the high speed digital folks what are the options for characterization ?
"Where did we ever come up with 50ohms. Lol, that's a question where you'll never ever get a straight answer." You will get many straight answers. some will probably be correct. as it has been answered many times in the comments I won't go into it here but it is actually a deliberate and careful choice confirmed by mathematics and real world experience.
Please explain why so many differential signal board traces are matched in geometric length by folding snakes on one side for compensation. In a thought experiment you can lay down a differential signal trace as a spiral. So at the end the signal must be in-phase instead of complentary at some frequeny following the geometric rule. But in reality they are coupled and still in 180 degree. So are the matching snakes bullshit?
Length matching is for skew not impedance. The most critical parameter of the relationship between members of a differential pair is time. The point when the signals cross is critical to signal integrity. The faster the edges of the signals, the more critical skew is.
What he said; but an ohm meter will hopefully read infinite between the conductors, whereas it also acts like a capacitor between the conductors and an inductor longwise along the length. Adding longer coax adds more capacitance and would reduce the capacitive reactance, BUT you have also just added some inductance which increased inductive reactance. When everything is designed to a specification, adding more length makes no noticeable difference to whatever is pushing a signal into it.
Cheers from Sweden. I watched over 10 hour of your videos during christmas in order to design a PCB that got me the highest grade possible in my PCB-design course :) Thank you for great content.
Thank you very much
good to hear i will take this course in a couple months too. right now still doing some theory
@@RobertFeranec en
@@RobertFeranec en
Thank you for trying out this new format! I really like it. However, now I can’t wait for the long version.
The short format is nice. Thanks
This is the exact type of videos that I was expecting from you. Great topic, short and direct to the point! Great job!
I hate the 2h long videos.
50 Ohms comes from the 30s and early use of coax cable where low loss in the cable was key. With the given cable properties back then it happened that at 50 Ohms the lowest attenuation (series resistance) was found.
50 ohms is for maximum power, 75ohm for minimum attenuation
@@DavidLopez-bz4rj the loss in a cable is referred to as attenuation = R/Z0. At 50 Ohm you get the minimum attenuation in db/m.
@@fabianluttenberger7153 Correction:
Minimum loss:77ohm
Max power:30 ohm
50 ohm is a compromise between this two. It happens in solid dielectric coax minimum loss is 50ohm.
It is 50 ohm for the amateur radio and 75 ohm for the video industry
De VU2RZA
@@DavidLopez-bz4rj That is the exact reason why TV antenna systems uses 75Ohm as you need minimal attenuation without the need to delivering whole lot of powers (for Rx), whereas HAM uses 50ohm for the obvious reason to Tx high power.
60 years plus working on lightbulbs(overhead projectors) to linear accelerators and almost everything electrical in between I have never seriously considered this aspect of design. Thank you for an epiphany moment. This why I UA-cam. You know I'm going to subscribe. Thanks again.
Some times I do not have time to watch all video. Its very nice to see key points of the videos.
I remember from microwave class that at higher frequencies with mismatched line impedance, a portion of signal would reflect. For max power transfer you'd always try to have same line impedance
We talk about this in the full interview
Correct, it’s about maximum power transmission ( impedance matching ?
YES, I remembered I have read related statements on why 50 ohms and why 50 ohms is so common on Eric's book, Signal and Power integrity 3rd. It's all about trade-off and 50 ohms is a good point to start tuning if need. Also there're some history and background inside too.
Again, thanks. This video just brings some memory from my mind. It's good for me.
Lee Ritchey has this great story about how he was working on a submarine SONAR system and the electronic design engineers insisted on 70 ohm transmission lines. I think it had to do with a concern that the drivers wouldn't be able to switch 50 ohm transmission lines quickly enough. Well because of the increased dielectric thickness to get 70 ohms they had terrible cross talk. Distance from the reference plane is a key consideration for cross talk and cross talk is directly proportional to dielectric thickness. Lee convinced them to lower the impedance to 50 ohms and they were able to actually build a functioning system.
That night after the second or third day of his course on making it to 56 Gbps when he told that story (we must have been talking about crosstalk) I flipped on TV and the Hunt for Red October was on. In the movie the Soviet Union had built a sub that had a silent drive system. The only was it could be tracked is with a very specific SONAR signature. So the Americans chase down the Soviet sub using SONAR in a Los Angels class US submarine, the very one Lee worked on. What a coincidence!
Lee speaks very highly of Eric Bogatin. I notice Eric referenced "loosely coupled" differential pairs. Taking the words right out of Lee's mouth. Not saying it was Lee's idea, but he stresses tightly coupled differential pairs are not a good idea.
historically the choice of 50ohms for coaxial cables was a compromise between maximum power and minimum losses.
Robert, thank you for the great content! Shorter videos are much better!
A Great video Robert, Thanks for that ! Regards from Chile
Very good and informative interview Robert! Great job! 😊
This short videos are great, thanks robert
really liked this video. it answered a question i haven't even had to ask yet.(if that makes sense, im still learning ya see,) so now i will know the answer to that 50 OMH impedance. coz it would have come up. SO IM VERY HAPPY TO HAVE SEEN THIS VIDEO. THANK YOU MR. ROBERT FERANEC..
Exactly what I needed! Thank you!
Great stuff! I have learned quite a lot from watching your channel! I love that you can explain everything in a easy to understand manner, and that your guests on the show do it as well.
Definitely like both the short and long form. Thanks!
thanks a lot, the short videos are more engaging. i could never watch full long videos although i know you are sharing great content and im missing a lot.
The impedance is the characteristic impedance (the ratio of voltage to current of a wave traveling in one direction without loss). 50 ohms is an arbitrary standard that is commonly used in rf and microwave transmission lines to prevent signals from reflecting off impedance discontinuities.
Hi, 50 ohm is not an arbitrary number.
A coaxial cable of 30 ohm characteristic impedance would handle more power.
A 75 ohm cable would have less attenuation with VHF and UHF frequencies.
And 50 is the arithmetic and geometric mean of the two numbers.
So, the 50 ohm becomes the standard for the amateur radio and the like and the. 75 ohm for the video industry.
De VU2RZA
Great interview! Very insightful and useful! Thanks for posting! Can’t wait to see the second part.
Another point of using 50 Ohms for the PCB layouts is the following: it's well known from the theory that the coax cable, which is widely used for the PCBs interconnections has (1) maximum power transport (for the fixed cross-section) if the impedance is = 90 Ohms: (2) Has minimum attenuation if the impedance is = 75 Ohms, and, finally, (3) Has minimum signal linear distortions (thus - maximum bandwidth) if the impedance is = 50 Ohms: this is the most important condition for the functioning of PCB.
75 ohms characteristic impedance corresponds to nearly minimum transmission loss, 30 ohms gives maximum power handling capability in Coaxial cables and 50 ohms is a rounded compromise number.
50 ohm significance is very well explained, Thanks.
David Casler also has an interesting video on why coax is 50 ohm.
great bedtime story! thanks Robert
You should keep doing these short clips, they are more likely to be seen than 1 hour :)
I think it all comes from TTL (Transistor Transistor Logic) from the 1960's which was 5V.
TTL = 5v, I=V/R=5/50=100mA.. Easy numbers to work with. Same resistance at each end of a line for maximum power transfer.
Industrial controls use 500 Ohms at 10V =20mA for 0-10V, 0-20mA, 4-20mA. 4-20mA being as 16mA range scales well into 16384=2^14+sign+parity=16 bits so in and integer register about 1000=1mA in 14bit
Sorry, nope not from TTL days. Impedance and resistance are different.
I love your videos to watch.. much informative they are… 🙏
wow this is very educational. thank you for the video!
Very informative video, reminded basics and gave them a deeper perspective. Thank you!
I just wanted to know an accurate, insightful or basic formula that explained 50 ohms. To say "tradeoffs" and end it there 😅
Thank you Sir, your tutorials are helping me a lot.
Radio transmitters often use 50 ohms, TV receivers use 75 ohms.
It is likely that transmitters must transmit more power.
With TV, it is important to transport low power cheaply.
Atif is here ur student sir... welcome to 2022 from india ...
Im currently facing a roadblock in my PCB design: I need a 50 ohm trace but I cant have a trace wider than 20mils (too thin) because of the 0.65mm pitch bga the signal comes from how do people deal with that . Thank you for all your amazing content its incredibly precious
One thing is to choose a layer stackup that minimizes the thickness of the prepreg between your signal layer and the ground plane below: thinner dielectric -> narrower traces for a given impedance. Other than that, you just have to neck down the traces eventually.
Does apply for analog and digital signals the same concept (use lower impedance for noising or higher for power purposes?
50ohm lines on FR4 are quite wide, especially on 0.062" material (0.105" wide). I can't imagine anyone using 50ohm lines unless you were working with 50ohm components.
Thanks Robert, your the man!!!
What should be the termination resistor for a 50ohms track when connecting to a port of microcontroller? Spi, signals for example
Very well explained
Thankyou Robert!
Noobie question. Isn't it 50 ohms over a specific length? Wouldn't the impedance double if you double the length of the cable? Or, am I completely misunderstanding?
It doesn’t depend on the length of the cable. The impedance of a transmission-line is a result of the cable’s internal capacitance and inductance. It’s what a driver “sees” as a load when trying to change the voltage on the cable, and create a wavefront that moves down the cable. At the receiving end, you want a load resistance that matches the cable impedance. Otherwise, some portion of the wavefront will reflect back to the source. Also, at the driving end, you want the driver’s resistance to match the cable impedance. If not, the wavefront could easily keep on reflecting back and forth for several end-to-end trips.
@@kc9scott Thank you, that helps.
If you add length, you have also added inductance which exactly balances the increased capacitance. The increase capacitance would tend to *reduce* the impedance (equivalent resistance) where the added inductance increases the impedance.
*Adding length changes loss* that is the only thing that increases with added length. The wire has actual resistance, not much but some, and the dielectric also causes a bit of loss as the electric field causes electrons to move slightly and that absorbs a bit of power.
@@thomasmaughan4798 Aha, now I think I'm getting it, thank you. Presently my struggle is how to understand that, knowing the way the two aspects of reactance exactly cancels when we do phasor diagrams. If it does cancel, why does it even matter? How do you measure the impedance rating of a cable?
Thanks to all of you who tried to help me understand this. I spent the morning coming to a solution. For noobies like me with similar questions, you'll find help in a Wikipedia page "Telegrapher's equations", where they explain the length is assumed to be an infinitely long line terminating in a ground, with no confusing back reflections. The capacitance does not completely cancel the inductance because the inductance is in series, the capacitance is in parallel (forehead smack). It is humbling to realize this was all worked out, over 150 years ago.
special thank to you
Dankie Robert
are you using a new camera? you look / it looks great!
Thank you Ben. I moved things around for the AltiumLive recording. I like it too ... I may use it more often, just need to figure out how to do recording when I am doing something on PC (in this setup I can't have monitor in the front of me)
Hello.. Is it necessary to keep 50 Ohm impedance for all Traces .. If we are designing any logical operations with 74HC595D(SPI Mode) is it necessary to match all the Input trace impedance! Kindly give a clarity between Hi speed design and low speed design.. When will be system considered as High speed design!? Any threshold
Please make such short vedios
Break long vedios in short
He mentions that higher impedance lowers current, thus lowering power consumption on other elements. why doesnt the power not dissipate on the 50 ohm impedance?
Please clarify the Trace length Calculation w.r.t speed of clock & High speed signals Tolerance
It would be nice to have an indication of the dimensions involved to get the 50 ohms impedance.
I hope if you could make video with embeddee software enginners, we can't find somethings like that on youtube
Very cool.
50 ohm impedance? I was not aware it was a balanced impedance. In the high frequency pulse it would make sense.
I have been wondering why 50 ohms is used for so long, even in the mathmatics when they use normalized impedance of 50 ohms to then figure out of things I was always wondering but why 50?
I have a question - how come trace impedance calculators only relay on dielectric thickness / stackup and line width? What about the trace length? How is it possible that I will get 50 Ohms with X thickness and with Y width, no matter the length?
because of the model :
Schematic of Heaviside's model of an infinitesimal segment of transmission line.
en.wikipedia.org/wiki/Characteristic_impedance
@@danielfranchette5547 Wow, it's quite simple now that I read about it, thanks a lot man!
"Where did we ever come up with 50ohms"
Lol, that's a question where you'll never ever get a straight answer. Another question where you never get a straight answer is assuming you chose 50 ohms as your line impedance what makes anyone assume that the impedance of the driver and source devices are anywhere near 50 ohms? For RF designs the impedance is presented to you in form of S-parameters with a whole slew of equipment for determining how things are matched. For the high speed digital folks what are the options for characterization ?
"Where did we ever come up with 50ohms. Lol, that's a question where you'll never ever get a straight answer."
You will get many straight answers. some will probably be correct. as it has been answered many times in the comments I won't go into it here but it is actually a deliberate and careful choice confirmed by mathematics and real world experience.
Please explain why so many differential signal board traces are matched in geometric length by folding snakes on one side for compensation.
In a thought experiment you can lay down a differential signal trace as a spiral. So at the end the signal must be in-phase instead of complentary at some frequeny following the geometric rule. But in reality they are coupled and still in 180 degree. So are the matching snakes bullshit?
Length matching is for skew not impedance. The most critical parameter of the relationship between members of a differential pair is time. The point when the signals cross is critical to signal integrity. The faster the edges of the signals, the more critical skew is.
Can someone please explain why thick dielectric means higher impedance? (2:10)
Decreases parasitic capacitance due to a the dielectric being thicker, so the impedance becomes higher
If it's always Ohm, why use "impedance" term instead of using "resistance" term?
Because when you model the line it actually also consist of the characteristics of some Inductance and Capacitance so it's not purely only resistance
@@lukmanhdrjaya Thanks for the clarification.
What he said; but an ohm meter will hopefully read infinite between the conductors, whereas it also acts like a capacitor between the conductors and an inductor longwise along the length. Adding longer coax adds more capacitance and would reduce the capacitive reactance, BUT you have also just added some inductance which increased inductive reactance. When everything is designed to a specification, adding more length makes no noticeable difference to whatever is pushing a signal into it.
Роберт, ты сейчас в Америке или в Словакии живешь?
We are currently in Slovakia
Danger!! 500,000 Ohms! Resistance is futile.
Leaving us with a nail biter
Is u got the answer ??
I don't !
IBM. 66? 666?
yyyyyyyyyyyyyyyyyyyyoooooooooooooooooooooooooooohhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh