1. Home
  2. Electronics Design
  3. Planar inductors with cores on ICs
  4. Page 2

Planar inductors with cores on ICs

If you're shielding to the outside world, not sure it makes a difference. For example, I couldn't find much difference between metal thickness A and two sheets of metal thickness one half A. Unless you separate the sheets quite a bit. As shielding with the high perm it can 'spot' saturate and thus have fields punch through, recommend adding a layer of mumetal to catch remnants.

If you make a toroid out of it [careful, work hardening by bending 'destroys' the goodness, too]; make a 3 winding core, primary, secondary, and 'control'. Drill a hole through the wrap so you can wind a figure 8 coil to the outsides and through that hole, call it, 'control winding'. Then locally saturate the core using the control winding with a high current, low voltage drive. Think like you are mechanically creating a huge 'air gap' at the saturation region. Just like physically opening and closing a gap. With the primary and secondary winding on the rest of the toroid, you end up with a DC-DC transformer [if you do the supporting electronics right]. Makes for a very interesting little isolated interface.

When I say work harden, I mean it. I had perm go from 1Meg down to 100k with just flexing less than ten times. However, in the torroid application, no biggie.

From memory a toroid with its distributed air gap did not have as good a high frequency performance as different type of structure. The skin effect eats you alive at every turn [pardon the pun] with this material when it's used as 'gross' material.

I tried to get both vendors to make 2 micron 'beads' by splattering directly into the liquid nitrogen, passivate the surface of the spheres, then 'gently' sintering into larger core material to get high perm, no coercivity, and nonconductivity, but they were too busy making huge cores for the power industry to be interested. Even little wires would be better.

Any interest for your firm to do it? Or, somebody out there? we'd make a fortune! maybe as much as $10k, or even $100k. Boy has out-sourcing changed MY expectations! ;)

Reply to
RobertMacy
Loading thread data ...

YES! and their dimensions?

Reply to
RobertMacy

Teensy cubed.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net
Reply to
Phil Hobbs

The precise technical value not sure of, but EXACTLY! that's why they work into the 100's MHz

From memory, that ??permalloy stuff has around 10 times the coercivity of metglas stuff. And I think its permeability has 'direction' too. Which can be a bit awesome to use.

Reply to
RobertMacy

In a tape-wound toroid the air gaps are in parallel with the magnetic path, rather than in series like they are in the ordinary gapped cores, or in 'distributed-gap' cores such as iron powder. I have hard time seeing why, in the first order, a parallel gap would matter.

Calculating from the datasheet numbers, the skin depth at 10 kHz would be 6um, in the ballpark of half-thickness of my foil. This sort-of-agrees with what I see in my crude homemade toroid: Re(Z) begins to rise above the copper Rdc when f > 20 kHz. Im(Z) begins to drop above 200 kHz, which OTOH seems to contradict the skin-depth-eats-away-cross-section hypothesis. A significant part of the core is air, it's difficult to make the the wires 'hug' the core with such small dimensions.

In LHe half the initial mu and lower Im(Z) corner frequency.

Maybe there is, I don't know. You know the composition? I'm wondering how would one be personally able to cash on it, if there is no protected IPR to sell, except by starting a company by oneself.

Regards, Mikko

Reply to
reg

Keep in mind skin depth assumes semi-infinite thickness, i.e., current flow over (and below) the surface of an infinite plane with solid material on one side, vacuum on the other. The case is more interesting for finite geometries, such as cylinders, pipes and plates. In the cylindrical case, Bessel functions arise; unfortunately, I forget what the plate case gives.

Since the current gets phase shifted (due to the bulk L/R time constant) as it goes deeper, waves from opposite sides interfere to make peaks and nulls. Under certain conditions, induction heating is capable, not only of heating a material in bulk, but actually putting more heat (higher power density) into deeper regions than at the surface. (I don't know if that's (for the case of cylinders) radial power density, which matters more than volumetric in matters of heat transfer. It seems remarkable even to me that the absolute power density could be higher. But I can't find a reference, unfortunately.)

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
Website: http://seventransistorlabs.com
Reply to
Tim Williams

the material I used was made as a 'special run', modification of one of their alloys, but again from memory, one of their standard materials had a skin depth of 3 microns at 1MHz. Yours must be pretty conductive to shallow out so fast.

I know I'm preaching to the choir, but skindepth is a nice reference, but it's related to planar waves and surfaces. When the field has sharp gradients... or when the material is in a magnetic field; things get a bit dicey. Surprising effects start happening.

I assume you're using some type of finite elecment analysis program to explore your possible architectures? Especially take a look at the ribbon toroid configuration. Surprised me just how much that was NOT a good use of the material. The resistance started up very early. Had to do with how the fields must propagate 'along' the ribbon in order to dart across the air gap, and it was that traveling along the ribbon where eddy currents thrived. I finally came to the conclusion, supported by Hasegawa's comments, that the only way to improve the material was not concentrate on the perm, but lower the conductivity, which he did in ?? material, saturated at 1T, lowered to somewhere around 0.7 MS/m.

The one thing that kept surprising me was how the optimized structure for MEMs inductor kept coming back to a single wire 'wrapped' with a core and NO shared flux, and the optimized structure for a multi-winding transformer kept coming back to conductors 'wrapped' with a core. Everything else had less performance.

Sort of 'turns your thinking inside out' Should have remembered, magnetics are LOW impedance. Look at SQUIDs.

Reply to
RobertMacy

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.