Do you agree with Grover, that the inductance of a finite, straight cylindrical wire is not a linear function of its length? Grover says that inductance is super-liner (grows faster than length-to-the-first-power) vs length. Is he mistaken?
(Here is a link to Grover's book)
Do you agree with Grover, that the inductance of a finite, straight cylindrical wire is not a linear function of its length? Grover says that inductance is super-liner (grows faster than length-to-the-first-power) vs length. Is he mistaken?
(Here is a link to Grover's book)
I am reminded of Barlow's law for resistance. Widely accepted at the time.
So now I have to read a bit to see if it is super-linear for wire inductance or linear as I have assumed along with JN.
Jneutron, are you saying that even with a Kelvin connection, varying the loop area of the test leads when measuring wire inductance will change the inductance measured?
As I understand since core and shield are magnetically coupled the DM output will reject CM noise. But if you drive current through the shield, you will measure voltage on the core between the ends of the cable. Surely this implies magnetic coupling?
As I understand since core and shield are magnetically coupled the DM output will reject CM noise. But if you drive current through the shield, you will measure voltage on the core between the ends of the cable. Surely this implies magnetic coupling?
Last edited:
Same superlinear formula for inductance-vs-length was published in 1908 by Edward B. Rosa, and again in 2009 by Priyanka Patel. Google found both of them (.pdf files, yes I realize there was no .pdf in 1908, this was a scan-and-convert job) for me, I bet Google will find both of them for you.
Too late. Was at clock school till 8 tonight, then went to the bar for a chicken sandwich.
I do not understand the distinction between "not a linear function" and "super liner", so do not know how to respond.
If you examine terman's equation for example, you note that he has three terms to the inductance of a wire pair. First is the inductance due to wire spacing, second is due to internal inductance (the 15 nH per foot modulated by a "delta function" related to skin effect, with some kind of lookup table but no Bessel functions), and a third component related to aspect ratio. The third is a consequence of magnetic field enhancement when the length is very short. I suspect the Grover Equation is detailing the very short effect. Once the aspect ratio exceeds 10:1, I would expect linear.
I note that I have had problems with aspect ratio when I was measuring 60 picohenry widgets, then you have to worry about how the current spreads from the test connection to the bulk. Due to that spreading, best I could test to was quarter nanohenry.
Yes. Also, how the other end of the wire returns to the meter. Claiming that a wire has "inductance" without considering the entire current loop is quite bogus. It is very important to consider the entire current path. Remember, even though it's a kelvin connection, at each connection there must be zero net current..so it is important to also consider the path of the current in the test leads. Didn't some dead guy postulate net zero current at nodes? Norton, or Thevenin was it? Gettin old..for some reason, I keep thinking Ralph Cramdon when I hear "Norton". Wonder why??
absolutely, it is magnetic coupling. But, when you connected your measuring device to both ends of the core, how did you prevent the shield magnetic field from coupling to the test leads? What is being measured is the time varying field of the shield magfield inducing a voltage in the loop formed by the core and the test leads, not any induced voltage in the core.
I never claimed that this E/M stuff was either easy, nor obvious.
Jn
Last edited:
If you have a coax that has 15uH from end to end, will it not behave like a 15uH common mode choke? Are you saying that it would behave differently from a 15uH twisted pair in this regard?
I understand the test leads being part of the loop, but saying there is no magnetic field in the core of coax is like saying there is no mutual inductance between the shield and core, unless you are saying that neither the shield nor core have any significant inductance.
I understand the test leads being part of the loop, but saying there is no magnetic field in the core of coax is like saying there is no mutual inductance between the shield and core, unless you are saying that neither the shield nor core have any significant inductance.
A coax inductance is a consequence of the magnetic field that resides only in the space between the shield and core, with the slight addition of 15nH per foot due to the core's internal magnetic field at DC of course.
What I said is that the shield produces no magnetic field inside the shield. Between the core and shield, the magnetic field (the inductance) is only due to the core's field. From the shield outward, the core and shield fields are equal and opposite, so cancel. A coax energy storage is confined to the space between the shield and core, while the twisted pair has an external dipole field with a helical twist.
Until there are external connections, there is no coupling core to shield...once external connections are made, you can integrate the total time varying flux through the loops.
As a thought experiment, consider the core circuit with a very small external loop, and the shield with a huge one. Current in the shield will produce a field, but the core loop picks up only a small portion of it. But current in the core, the shield loop picks up a whole bunch. Assymmetric.
Jn
Last edited:
Have to give an engineering presentation in sept at work on motion control theory, practice, and application..For the intro, I am going to list as hobbies, teaching antique clock repair, and face planting in front of large audiences..Reminds me I did say I would start a thread on audio grounding. As soon as I have enough material to be thoroughly lampooned I will do so.
How's your offspring doing at the job?
Jn
I could say "path of least impedance", but that would be just a really out there wag...
Did I pass?
Jn
Ps.. You've never seemed to ask me in order to learn, but only to trip...If I err, I apologize.
Last edited:
I could say "path of least impedance", but that would be just a really out there wag...
Did I pass?
Jn
Ps.. You've never seemed to ask me in order to learn, but only to trip...If I err, I apologize.
In non-physic terms.... wouldn't the path of least impedance be one with lowest inductance, also?
and in terms of flux density, would lowest L be caused by lowest flux or highest flux density?
Just checking.
THx-RNM
At high frequency, the path of least impedance is primarily the path of lowest inductance.
For a twisted pair and a coax, the current centroids are trying to get closer.
If the core is off center, at low freq the shield current will be uniform so there will be an external mag field. But as freq goes up, the shield current will redistribute such that it's center of "mass" as it were, will center on the core's centroid. That reduces external field.
Your question, core current on outer surface and shield current on inner surface, is simply the currents shifting as close to each other as possible.
Jn
She's finished now and home for the holidays before regressing back into being a student. The year has been really good for her and she won an award at some Institute of physics do for best presentation by a student last week. She'll go back next summer for another couple of months.
She has also realised that she could go back there after degree, get sponsored for phd and wake up one day to realise 20 years had passed and she had been borged without realising it. I can't help her as I have no idea even what I want to do if I ever grow up
There is some research showing that the best way to find out what it would be like to do something like getting a PhD is to find some people who have done exactly what is contemplated and ask them how they like it.
It also matters a whole lot who the faculty adviser for the PhD is. They will have great influence over what the candidate is likely to be working on for a long time, sometimes long after having obtained the degree.
Probably there are a few other people who around here who can offer a few tidbits that might be worth passing along.
You mean like the people she spent the last year working with who all went there straight after graduating, got PhDs and are still there...Yup.
She want's to do a PhD (luckily not bothered about a DPhil) but she isn't sure if she wants to do it at the x-ray laser facility. She is both more intelligent than me (proud) and more scared of the outside world than I was at that age, but not by much. I saw around me when I was at uni people who had hung around the area and were still behaving like students at 40 as that was easier than becoming an integrated part of society. I decided to drag myself away to avoid that. Looking back not entirely sure it was the right move, but hey ho.
I should also note that I was put off academia* when I discovered my supervisor, in 20 years had never actually made a whole transistor in the lab. Just bits of them. I'm just not stoic enough for that.
*plus wasn't cleva enuf
She want's to do a PhD (luckily not bothered about a DPhil) but she isn't sure if she wants to do it at the x-ray laser facility. She is both more intelligent than me (proud) and more scared of the outside world than I was at that age, but not by much. I saw around me when I was at uni people who had hung around the area and were still behaving like students at 40 as that was easier than becoming an integrated part of society. I decided to drag myself away to avoid that. Looking back not entirely sure it was the right move, but hey ho.
I should also note that I was put off academia* when I discovered my supervisor, in 20 years had never actually made a whole transistor in the lab. Just bits of them. I'm just not stoic enough for that.
*plus wasn't cleva enuf
JN- superlinear means "grows faster than a linear function"
The standard examples in the EECS curriculum are N**2, N*log(N), and exp(N). All of these grow faster than a linear function, they are superlinear.
And, according to Grover + Rosa + Patel, the self inductance of a straight cylindrical conductor of diameter 1mm and length 300mm, is more than three times the self inductance of a straight conductor of diameter 1mm and length 100mm. Inductance is a superlinear function of length.
The standard examples in the EECS curriculum are N**2, N*log(N), and exp(N). All of these grow faster than a linear function, they are superlinear.
And, according to Grover + Rosa + Patel, the self inductance of a straight cylindrical conductor of diameter 1mm and length 300mm, is more than three times the self inductance of a straight conductor of diameter 1mm and length 100mm. Inductance is a superlinear function of length.
She has also realised that she could go back there after degree, get sponsored for phd and wake up one day to realise 20 years had passed and she had been borged without realising it.
My wife did it, there was always an element that it would be an achievement in itself. We only had 4 kids and occasionally 2 Japanese/Korean exchange students, one of which literally wrapped her arms around my legs and had to be dragged to the cab home saying she wanted me to be her father.
She did a thesis on Krylov subspace techniques and then an internship at Cadence with Ken Kundert himself but never pursued it after.
hmm, I'd have to see the derivation, something is up.
It seems like they are talking about circumferential (azimuthal) currents, not axial.
jn
- Status
- Not open for further replies.