Ah thanks. I work with very high currents, very high static and dynamic fields, and at 4.5K, so was wondering if it was something I was unaware of.
We do get very interesting issues when multiple wire strands of super are involved. One cable I use is 6 around 1, and at ramp rates over 1/2 amp per second, the middle wire doesn't see current due to skinning. During ramps over that, the bundle will have 6/7ths of the wire's capability. The flat rutherford cables typically have 30 wires, and that cable is ramp sensitive especially at tight turns. There, the twist pitch vs bend radius is the important parametric, as one wire can be on the ID for most of the turn, and one can be on the OD, so ramping magfield will create circulating currents across the cable.
I've also had copper issues. When a good copper is used, it's conductivity will rise two orders of magnitude at 4.5K, so it's skin depth will drop into the tenths of an inch. Ramps can generate copper heating, and if the super is soldered to it, the heat can quench it. This because below about 50K, metals have almost no heat capacity..
If you ever recall what you saw, I'd appreciate the info. Thanks.
jn
Last edited:
Attached.
As long as only the standard electrodynamics applies, there is no reason to believe the skin effect has anything nonlinear.
However, when the solid state conduction mechanisms in metals are involved (or could no longer ignored, like in this case when the skin depth is of the order of magnitude of the free electron path), the material properties are no longer necessary constant (or even well defined!), and the linearity of the solutions could be measurably compromised, depending on the field strenghts, etc...
Maxwell equations are macroscopic, the microscopic view could be different.
Excellent...thank you very much.
It doesn't apply in general to my work at DC, but could have impact on where I connect v taps to big pieces of copper...
again, thanks..and df, thanks for mentioning it.
Well, lets see.
1. Looking for nonlinear skin effect issues in normal conductors is pretty much a waste of time, we all certainly agree on that.
2. Proximity effect also for regular conductors, but for inductors, proximity effect issues can deliver unexpected parameter changes, recall RNMarsh and how he selects L's based on Q?
3. The interleaved resistor concept eliminates all current slew errors inherent in low value resistors, this is of use for emitter stabilizers as a start (edit:individual transistor collectors could be sent through a zero ohm interleave with the emitter R's totally eliminating a lot of inductance). Splitting the interleave into two very different values allows the building of a zero error feedback ladder, removing one more error source in GNFB systems. Making them electrically independent is an extension of the design, and I've not yet had time to think about where it may be useful (other than the "so called trivial" circuit). But should I think of a new application for the design, you know darn well I'll share it with the great crew of people on this site.
4. While designing a fet switch (not jfet as you stated) capable of 1 nsec full power transitions may be trivial to you, it is not trivial to me. As a circuit "newbie", should I not ask questions out of the fear that someone like you would call it trivial? Or, should someone like me be allowed unfettered access to people who are very very knowledgeable and can assist me in areas I am weak in? While I am very good in the layout aspects, I am not so with the semiconductors and appreciate the good comments given me so far.
jn
Last edited:
I believe John Curl has been stating such for the last decade (at least).
In addition, steel chassis can do it by eddies, and a magnetic steel wall separating two low impedance circuits can cause an intermixing. This from Tom Van Doren's lecture.
jn
In the more sophisticated PCB software....
There is usually a 'fudging' factor added to the classic formulas [text book] used for difference pcb layout topologies... for strip-line, ground planes, Z etc because of the field coupling isnt perfect (fringing affects, skin effects, L, edge irrigularities, coupling to other traces etc). These included factors add up to a fudge factor in the formula. If not, you wont get the accurate results you are expecting.
THx-RNMarsh
Really, I stated years ago it is easy to measure the effect of a steel bench top on an air core filter during characterizing the AD797. I needed to improve an old Sound Technology generator because it was all I had.
Yes, years ago we made an AC power strip regulated with an internal PWM converter as a research project. The output filter was a 60Hz iron core inductor, which was right next to a steel end panel, and that panel got hot as heck.
Got a stove that works like that .
Yes, we thought about calling that a feature, and making it into a coffee warmer.
Still alive and kicking ....Parasound Shares its Strengths in San Jose | Stereophile.com
There has, it seems to me, always been an assumption that many things seen in HF or RF cant have any possible impact at AF. This isnt always true. Take this case of a ground plane on an audio pcb...... what is the path a 10Hz signal will take on the ground plane?
View attachment grnd plane 10Hz.pdf
It will travel in a direct line from input to output - a diagonal path. But what if the freq was increased to only 1000 Hz? Is the path still the same?
THx-RNMarsh
View attachment grnd plane 10Hz.pdf
It will travel in a direct line from input to output - a diagonal path. But what if the freq was increased to only 1000 Hz? Is the path still the same?
THx-RNMarsh
Customers get older, don´t they :https://www.youtube.com/watch?v=LOz6tsOOhYo&list=UUZEBNIVWh_lTlHH8vjKRlGQ
- Status
- Not open for further replies.