Re: shielded inductor -> low frequency?

Simon S Aysdie wrote:

On Sep 20, 5:58 pm, Terry Given <my_n...@xxxxxxxx> wrote:

Joerg wrote:

LVMarc wrote:

Michael wrote:

Greetings gents of the SED variety - I have a couple quick questions
for you all:

I am working on a DC/DC switcher that will use the Linear LT1930A. It
switches at 2.2MHz. I need to find a ~8-10uH inductor that can handle
about .5-1A for this. I identified the Sumida CR43 series as a good
candidate. For example, there is a 8.2uH (at 7.96MHz) CR43 series with
166.2mohms DCR speced at 0.84A. It also has a very small 4.3x4.8mm
footprint. It would be perfect, except that it isn't shielded. I would
really prefer it to be shielded. I started looking at other Sumida
inductors, and it seems that all the shielded ones are speced in the
low KHz region. Am I missing something here?

Also, what is the relationship between inductance and frequency? I
googled around about that for a bit but couldn't find anything.

Thanks,

-Michael

Electrostatic (conductive) Shielding does not effect the magnetic
lines of flux you need high u material for that. and high u magnetic
shielding material only works at low < 100Kc frequency.

so placing a metal can around a coil is only partially effective in
reducing radiated emissions.

I believe what Michael meant is a closed versus open magnetic path.
Closed path is what inductor company marketeers call "shielded". For
example, a ferrite toroid would be quite good in that respect, that's
what I usually use. I found that custom fab in Taiwan/China is actually
less expensive than a fancy catalog inductor. Provided the quantities
are at least a few thousand a month.

these "shielded" bobbin cores have a ring of ferrite material placed
around them, spaced out a mm or two from the OD of the bobbin flanges.


They seem to be nearly pot-core in design, as I see them.


it reduces external field leakage from an awful lot to a fair bit. that
might suffice.

Joerg's approach (toroidal cores eg koolmu, -52 etc) is by far the best.


But it can also be the most expensive, especially if quantities are low and you can't find a catalog part.


What do you think of page 52 of Linear's AN70?



Man, that prototype on page 37 sure looks ugly. The generator on page 46 is worthwhile peeking at. I wonder how they kept it so fresh. I've got one from the 50's but it looks way more tired.

Page 52: H-field EMI can be dealt with (to some extent) with crossover turns but it is an art to get these just right for best cancellation. Almost like adjusting microwave stuff. It is even tougher to then explain all that in a fabrication procedure. T'is where even us electrical guys need to whip up a mechanical CAD drawing.

BTW since it's often said that shielding doesn't work I have to disagree. Even a Faraday type shield has saved the bacon many, many times in my EMI work. And EMI is >30% of my bread and butter.

--
Regards, Joerg

http://www.analogconsultants.com
.

Quantcast