[TowerTalk] Litz wire

[Jim Lux]

At 11:02 AM 3/14/2005, Tom Rauch wrote:
> > Ah yes.. but manufacturing technology has advanced quite a
>bit since the
> > days of Terman.  It's all well and good to assert that
>1000kc was the
> > practical limit in Terman's day.  However, I wonder (not
>too much, but just
> > curious) what today's practical limit might be.
>
>The nice thing about my job is I've always been paid to
>measure things like this, and manufacturers and vendors give
>me sample parts.
>
>As I've said, I've never found a case above 300kHz (the
>lower limit of my RF impedance test sets) where Litz wire
>has decreased impedance, ESR, or increased Q in the same
>form factor when replacing solid copper wire.     Never.
>
>I can't imagine stranded wire being better than solid wire
>for ground impedance when the diameter is the same, because
>it never is better for RF from the BC band on up. It is
>never better at DC either for a given OD.

All true, if you're OD limited.  But there are non-OD limited cases that 
are of substantial economic importance, so that must be driving the use and 
manufacture of litz wire (otherwise, they'd not be making it in industrial 
quantities).



>As a matter of fact when I modified a KW Hustler mobile
>loading coil for 75M by rewinding it with the same diameter
>solid enameled wire instead of the Litz wire and make no
>other changes Q increases about 1.6-1.7 times.
>
>I've seen Litz wire used in small LF switching transformers,
>but never measured one A-B with enameled wire the same
>diameter for ESR. So I don't know where Litz wire improves
>things. I do know the DC resistance (for a given wire
>diameter) is higher with Litz wire and the 300kHz and higher
>resistance is higher in any application I've tested, but
>there might be a sweet spot in some applications.


There seems to be a fair amount of interest in it.
Here's a paper by someone proposing something better than Litz (or foil), 
but it shows that Litz wire is better than plain round copper.
http://www.schottcorp.com/news/technical_papers/apecpaper.pdf

Most of the application data for new PWM chips refer to the use of Litz 
wire in the magnetics.

These are all typically at several hundred kHz and higher.
A reference from TI talks about using Litz wire to maintain a reasonable 
Rac/Rdc, trading off the decreased Rac for the increased Rdc (due to less 
copper in a given cross section).  I think a lot of the design process  has 
to do with what the relative levels of ripple current and DC current are, 
which in turn has a lot to do with the dynamic range required of the power 
converter.

Everybody seems to cite the same reference (Unitrode Design Seminar SEM-400)


>My transformer book claims it only helps reduce eddy
>currents in the copper, nothing else. I tend to thing that
>is correct based on what I've measured.


In a transformer application, the inductance of the wire is immaterial, of 
course, since the permeability of the core will dominate. Then it would 
come down to a AC loss vs DC loss issue of eddy currents and skin effect, etc.

The lower inductance would be relevant for isolated conductors (not wound 
in a coil).. perhaps in a fast pulse discharge system (but there, you might 
be better off using a transmission line)