Why do people keep saying that, anyway?
I'd rather have insulating corrosion -- it doesn't have any effect on the fields!
Either one is bad for _connectors_, with one being worse than the other. But plated parts? Conductive oxide is like nickel plating on top!
Given that the silver plating is thick enough not to corrode through, of course.
Tim
-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website: https://www.seventransistorlabs.com/
Nickel-zinc ferrite might get you to 30MHz
And plated-on silver is less conductive than clean copper.
Use a pulsed laser to melt the electro-deposited crystals without alloying them with the copper, and you might actually see that 5% advantage.
A coil of high-temperature super-conductor ribbon sunk in liquid nitrogen could be whole lot better. Or you could get something fabricated with YBa2 Cu3 O7-delta thin films deposited on a LaAlO3 substrate, and sink that in liquid nitrogen.
Super-conductors don't get hot, so lots of turns in a small volume gets a lot more practical
But a coil of high temperature superconductor in liquid nitrogen would take you quite a bit further, though probably nowhere that John Larkin wants to get to.
He's not a fan of great leaps forward.
-- Bill Sloman, Sydney
The problem for litz at higher RF frequencies is its extremely high wire-to-wire self capacitance, which makes it look like a single big copper wire.
John, are you sure you need such a high inductance? How about staged values in series, the smaller ones would handle the highest frequencies and work at RF, the larger ones would handle the lower frequencies.
--
Thanks,
- Win
Maybe, but very easy to work with and what's more, they readily give free samples with web-page requests.
This air coil series is good at RF freqs to 400MHz, calculate the loss at any freq from their Q curves.
--
Thanks,
- Win
The Coilcraft 1010's are (somehow) wound from flat copper. The coils are tight, so maybe proximity effect crowds current out to the edges and increases skin losses. Pulsing, they get almost as hot as conventional wound coils, but their thermal conductivity is huge: soldering one pin melts the solder on the other. So they are candidates for heat sinking.
You could probably make some kind of decorative motor with them. Or use three or four on a PC board to levitate and spin a metallic sphere.
You could make a high-voltage transformer by putting a pair on opposite sides of a PC board.
I've got to investigate interaction with my PC board's ground plane,
30 mils below the surface. Maybe I'm heating the board with the mag field.-- John Larkin Highland Technology, Inc lunatic fringe electronics
I've seen big transmitters tank coils that are wound from water-cooled copper pipes; looks like other people have had this problem before!
-- John Larkin Highland Technology, Inc lunatic fringe electronics
eed to try some 12 gauge house wire. I've seen coils made from copper tubi ng.
t of aluminum and then plate the aluminum coil with copper. And then remov e the aluminum using Lye. And make a heat pipe or pump water through the h ollow copper. You would probably have to start with aluminum tubing so the lye solution could be circulated thru the coil. Other wise it would take forever to disolve the aluminum.
TWT's this way. Starting with electroplating copper onto aluminum wire. Then putting sections of the wire into copper pipe and swaging the pipe to cold weld the copper together. Then twisting the pipe to the same spiral a s the electron beam would have. Then slicing the pipe into wafers and etch ing out the aluminum. The result was copper grids.
convince them that they should get into making hollow copper coils.
A back-of-the-envelope scribble says just switching from #22 tinned to #18 bare copper would reduce your losses by a factor of ten.
Cheers, James
I do need something around 320 nH.
I'm thinking that it may be fundamentally impossible to make an inductor with low enough losses, in the volume I have to work in. SAo the problem becomes thermal.
I have done some work with making wideband inductors, for other projects. Staging various size inductors is a basic technique, but there have to be shunt elements, otherwise the same current goes through all of the inductors. Resonances can be a problem too. Wideband bias tees heve to deal with this.
There are some super-wideband inductors, straight and tapered, that have ferrite-epoxy fill inside. The Piconics idea, and the Coilcraft
4310. I don't know how they handle HF power.-- John Larkin Highland Technology, Inc lunatic fringe electronics
But it isn't insulating. Copper oxide is a semiconductor. Silver oxide is very thin, because silver is a noble metal. Corrosion roughens the surface and degrades the conductivity of the first few monolayers of metal, both of which increase loss.
Nickel is bad because nickel oxide forms tunnel junctions so easily. (I used to make Ni-NiO-Ni tunnel junctions for a living.)
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510 http://electrooptical.net http://hobbs-eo.com
Now, roughness is a real argument. I recall reading the AC resistance of PCB traces can be 2x or more, due to the surface treatment (oxide or just roughness) normally needed for peel strength.
But then, is a smooth patina okay? Verdigris? I suppose the metal-salt surface is always going to be rough, but it can be smooth at the wavelength in question, whatever that wavelength might be. It might be a problem at
30GHz, but at the 30MHz in this thread? The poor thing would have to be hot-rolled for that to be a problem...So aside from roughness -- /so what/ about insulators? Everyone parrots that statement, but it doesn't check out!
Even semiconductors aren't a problem, as long as their resistivity is that many times the bulk metal's, so that a tiny fraction of signal current flows in them.
Seriously, I've done this before, on a much grander* scale -- I designed an induction heater that runs at two frequencies, for stirring (at low frequencies: magnetic field penetrates the graphite crucible) and gentle heating (at high frequency).
(*Grand, in the literal sense that the frequencies in question are >= 3 orders of magnitude lower. Power level is also higher, by a few decades I guess, but the system is linear enough not to care, so that doesn't matter.)
That too. But just on the matter of Q and power handling, it's awful because of its permeability. You don't use nickel underplating in waveguides (at least, not without an unusually generous thickness of silver on top of it).
Tim
-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website: https://www.seventransistorlabs.com/
The Q vs freq of the inductors looks good enough, I'll be your introducing huge eddy-current losses in the PCB below the inductors. Cut radial slits.
--
Thanks,
- Win
An excellent point -- electroplated copper (the bulk of PCB traces) is around 50% worse than pure (annealed) copper. Haven't seen numbers on silver, but it's reasonable to assume a similar outcome.
Ag-Cu alloy is worse than either one alone, so it must be done quickly indeed!
Hmm, can you even quench-anneal silver? I suspect so. Probably still not ideal because you won't be getting much grain growth in that time.
I suppose the much better industrial method would be simply cold cladding, then drawing.
Depends. (Most?) Commercial wire isn't made for AC operation: it's copper clad, so it doesn't turn into an insulator when it warms up.
Is there pure-superconductor cable that's made for AC operation? Dunno, I see very little on superconductors in the normal course of business, sadly.
Type II superconductors can also be pretty awful away from DC. Flux pinning is a hysteretic loss effect. I don't know if that's characteristic of the toy bricks, or if they've solved that with better methods.
Even the best (e.g., Nb at 2K) has a high but not infinite Q (on the order of 10^7). Though for Q's that high, I wonder if they're stretching the capability of their instruments to measure it at all.
Neato.
They claim "very high quality factors", but happened to leave a number out of the abstract...
Anyway, even without superconductors, regular copper (it has to be very pure -- low RRR) has quite good performance at low temperatures, I'm sure enough to do the job here. But then you still have the problem of a LN2 dewar sticking out of the thing, or however many watts a cryopump needs to do the same job.
Tim
-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website: https://www.seventransistorlabs.com/
The two Coilcraft parts that I've used (initially 4 of the maxi-springs, then tried two of the 1010VS's) both get real hot, and they are nice pure enameled copper.
It would be nice to use a tape-feed pick-and-place part. Winding and plating stuff would not make me popular downstairs.
-- John Larkin Highland Technology, Inc lunatic fringe electronics
Or buy from CoilCraft and send them out to a plating house.
You might try some of the smaller powdered UHF balun cores, with only a turn or two. They have low permeability and offer moderate surface area with good thermal conductivity, if the winding makes good physical contact (varnish or epoxy).
RL
In case you have not checked, you can buy copper tubing in lots of sizes. 1/4 inch, 1/8th inch, 1/16 inch. capillary tubing for refrigeration in a number of sizes ( WW Grainger lists 17 sizes ) and then there are metric sizes too.
Dan
Strip the enamel first?
There would be a small or probably no advantage to silver plating copper. Not worth the considerable cost or hassle.
I was looking at the 1010VS-141. It's 146 nH, Q=140 at 50 MHz, DCR=3 milliohms.
At 50 MHz, reactance is 23 ohms and the equivalent resistance is 164 mohms, about 55 times the DCR (which is my problem!)
6 amps RMS at 50 MHz will dissipate about 6 watts. Two in series makes 12 watts.The thermal resistance end-to-end is about 450 K/W, so both ends need to be well heat sunk. Theta from the middle out to both ends is about
110 K/W, so the middle will get very warm.The numbers in the Coilcraft data sheet suggest a theta of 34 K/W, which sounds high to me. 34*6 would be 200 degrees C rise!
Gotta heat sink them.
We put gobs of effort into the thermal management of the semiconductors and messed up the inductors!
-- John Larkin Highland Technology, Inc lunatic fringe electronics
You don't think a plating house can do that?
But the reactance doesn't heat the coil.
I thought your design reviews covered everything. ;-)
Mouser does EPCOS/TDK RM5 and RM6 K1 (nickel-zinc ferrite) cores.
The RM6 part
offers 40nH+/-3% per root turn, so three turns would give John 360nH.
It's gapped core (0.8mm gap) whence the tight inductance spec.
The material is good to about 200mT and about 100A/metre. The path length around the core is about 30mm, so John's 6A is too high. Three or four in parallel might work, or two two-turn inductors in series ...
The coil or coils would be very short lengths of what could be quite fat wire almost 3mm OD would work - or one could machine a helix out of a chunk of copper and clamp the RM6 core around it - so the resistive losses could get right down.
It's possible that you could get hold of 2.5mm OD copper tube to make the three turn coil, and pump water through it, but you couldn't get much a flow rate in tubing that small.
There may be bigger K1 cores around, but they'd be hard to get hold of.
-- Bill Sloman, Sydney
You are trying to put too much current in too small a wire. You could make a flat spiral inductor with ~400 nH inductance from 4 turns of 8 ga magnet wire. It would take about 1.5 inch dia and have a RF resistance at 30 MHz of several milliohms. It would barely get warm. Unfortunately, you box yourself in trying to work in too small a space.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.