The AM Forum
December 12, 2024, 12:35:24 PM *
Welcome, Guest. Please login or register.

Login with username, password and session length
 
   Home   Help Calendar Links Staff List Gallery Login Register  
Pages: [1]   Go Down
  Print  
Author Topic: Core Materials for PWM Filters - Some Experiments, Test Data and Pics  (Read 17557 times)
0 Members and 1 Guest are viewing this topic.
steve_qix
Contributing
Member
*
Offline Offline

Posts: 2598


Bap!


WWW
« on: March 12, 2016, 06:56:17 PM »

For a while now, I've been using toroids in PWM filters in pulse width modulators used to modulate class E transmitters.

The advantage of toroids over air wound coils is obvious - space.  The DIS-advantage of cores is that the permeability changes with current, and this causes the value of the inductor to change with current - specifically, while modulating.

So, in order to have a successful inductor, the value of the inductor cannot change very much at all over the modulation cycle from max to 0 (100% negative).

There are good core materials that have very stable AL values (initial permeability) over a wide range of ampere-turns (DC bias) values.

I did some experiments with 4 different core materials - measuring the inductance of a single layer toroidal inductor with 0A of current and with 12 A of current flowing through the inductor itself.  The findings are interesting, but not unexpected.  The better the core material, the more stable the inductor.

Results:


Core Description......................................Inductance No DC Bias....Inductance 12A DC Bias....Percent Inductance Change (drop)
.
Micrometals T300-40 3.063 O.D.63uH53uH16%
Magnetics 58867A2 3.063 In. O.D.(Hi Flux)55uH54uH2%
Magnetics 78867A7 3.063 In. O.D.(X-Flux)53uH51uH4%
CWS CK1650060 6.5 In. O.D. (MegaFlux)40.3uH40.2uH.25%

The first three inductors are the same size, would on a "777" size core - about 3.063 inches diameter, and about 1/2 inch high.  These are commonly available cores.

Obviously, the best (and of the most expensive) material is the Hi Flux.  X-Flux is also quite acceptable, and is less expensive.  The T300-40 core is UNACCEPTABLE as material for an inductor with varying DC bias.  Micrometals does make a very good high flux core in the 777 size - I just don't have any of these here.

The CWS core is a _HUGE_ core, and it is VERY heavy.  I have a bunch of these around, so I thought I'd measure it just for yuks.  The Megaflux material is very similar to X-flux, and is acceptable for PWM filter inductors.

NOTE: Using a toroid as the input inductor of a PWM filter is only acceptable for relatively low current applications (10A or under at carrier).  The "Ripple Current" causes core heating and loss.  I recommend using an air core inductor as the first inductor for anything over 10A.  Stacking cores helps, but not much as the input inductor.

I have a 1kW pulse width modulator/power supply at Rattlesnake Island that uses a 6 pole filter (3 inductors).  The first inductor is a small (in inductance) air wound inductor, and the remaining two are each made using 5 X-Flux cores stacked.  This modulator delivers 26A of current at carrier, and the inductors are stable over the entire modulation capability of the system.

An inductance variation of 10% or less, over the entire modulation cycle is acceptable.

Micrometals and Magnetics both have very good inductor design software (free off their web sites), which will show you inductance variations, core losses, wire losses and other important data for their particular cores.  I use this software all the time.

The Picture shows the three 777 size cores, one of which is setting atop the very large CK1650060 core.  In the picture, the CK1650 core is the one wound at the moment.  I wound the same wires around each of the smaller cores with the same number of turns for the experiment.  The inductance differences between the inductors are the result of variations in the AL values of the different materials.  The very large air inductor in the background is my "isolation" inductor, so I can make inductance measurements without the power supply "shorting out" the inductor during tests.  The large air inductor is 350uH.




Logged

High Power, Broadcast Audio and Low Cost?  Check out the class E web site at: http://www.classeradio.org
n1ps
Contributing
Member
*
Offline Offline

Posts: 338


websdr http://sebagolakesdr.us:8901/


« Reply #1 on: March 12, 2016, 07:39:37 PM »

Great and very helpful especially for those looking at the 24 pill decks.

p
Logged
Opcom
Patrick J. / KD5OEI
Contributing
Member
*
Offline Offline

Posts: 8301



WWW
« Reply #2 on: March 14, 2016, 01:27:10 AM »

This is really valuable info. Thanks!
Logged

Radio Candelstein
John K5PRO
Member

Offline Offline

Posts: 1028



« Reply #3 on: April 17, 2016, 01:07:27 PM »

Have you dealt with Ceramic Magnetics Inc? They are nice folks, and do a lot of BIG custom ferrite cores for science projects like particle accelerators. They have been around for ages and seem to know their stuff. They made some toroids for me a decade ago that were 0.5 meter OD! They might make toroidal cores with the right characteristics, like the CWS you liked.
Logged
steve_qix
Contributing
Member
*
Offline Offline

Posts: 2598


Bap!


WWW
« Reply #4 on: May 02, 2016, 08:48:25 PM »

Thanks!  I don't know about Ceramic Magnetics, but I'll definitely check them out.    It would be kind of neat to get one core that would do a filter inductor, rather than having to stack smaller cores.
Logged

High Power, Broadcast Audio and Low Cost?  Check out the class E web site at: http://www.classeradio.org
DMOD
AC0OB - A Place where Thermionic Emitters Rule!
Contributing
Member
*
Offline Offline

Posts: 1806


« Reply #5 on: May 17, 2016, 07:18:23 PM »

Thanks for the data Steve.

Phil - AC0OB
Logged

Charlie Eppes: Dad would be so happy if we married a doctor.
Don Eppes: Yeah, well, Dad would be happy if I married someone with a pulse.NUMB3RS   Smiley
KQ6F
Member

Offline Offline

Posts: 124


« Reply #6 on: November 23, 2016, 08:56:35 PM »

Steve -

Based on your experiments I have ordered sample cores from Magnetics..both Hi-Flux and X-Flux.  I also tried to find your MegaFlux core on the CWS website but came up short.  Couldn't even find any reference to MegaFlux.  I wonder if I'm on the wrong website...they are calling themselves CWS Bytemark now..  Huh

Rod KQ6F
Logged
steve_qix
Contributing
Member
*
Offline Offline

Posts: 2598


Bap!


WWW
« Reply #7 on: November 23, 2016, 10:08:36 PM »

http://www.cwsbytemark.com/index.php?main_page=index&cPath=206_238

The CWS site can be a bit hard to navigate for sure!   Here is a link to the "MegaFlux" cores.

The x flux is very good for the money.  I use this type of material in PWM filters quite a bit.  It is a stable, cost effective material.

Since I've done these experiments, I have identified a core material that is suitable for the input inductor of KW class pulse width modulators.  I am using one of these inductors in my 24 FET transmitter's modulator with excellent success.  With this, one can build a modulator with ONLY core based inductors, further reducing the size of the unit.

The material is Micrometals type 2.  I got a bunch of T300 (the size) - (2D the material, type 2), (D= double thickness) T300-2D is the part number.  This is a low perm material.  The permeability is quite stable over a wide range of DC bias (current) values. In the case of this inductor, the inductance varies less than 10% from 0A to 75A of DC bias (current)!  

I was able to wind (the wire barely fit - #10) an input inductor suitable for a 6 pole Butterworth filter designed for a 2 ohm load.  I used a single T300-2D core.  Did not stack.  No core heating, and the filter behaves properly.  I don't remember the exact number of turns.

Good stuff!
Logged

High Power, Broadcast Audio and Low Cost?  Check out the class E web site at: http://www.classeradio.org
Pages: [1]   Go Up
  Print  
 
Jump to:  

AMfone - Dedicated to Amplitude Modulation on the Amateur Radio Bands
 AMfone © 2001-2015
Powered by SMF 1.1.21 | SMF © 2015, Simple Machines
Page created in 0.038 seconds with 18 queries.