PDM output filter

Welcome and feel free to ask questions about class E, mosfet's, and AM.
You must be registered to post.

PDM output filter

Postby W2ZE » Thu Feb 06, 2003 10:52 am

Can anyone tell me the where and how I can find the values for the output filter :?: Also, how large do the coils have to be? I am trying to get an idea how to mount all of my components, and need general info. Thanks :)
Posts: 1
Joined: Thu Feb 06, 2003 10:45 am
Location: Shortsville, NY

Steve's site

Postby VE7KHz » Thu Feb 06, 2003 11:07 am

Try this link. Steve has an excellent write up on the topic. This is my first exporsure to PWM and is the article that got me "hooked".

http://www.netway.com/~stevec/ham/pdm_a ... state.html

It's a good read.

I used to keep an open mind.......but my brain kept falling out.
User avatar
Posts: 23
Joined: Fri Jan 31, 2003 3:01 pm

Filter calcs

Postby VK3KRI » Fri Feb 07, 2003 4:46 pm

I stumbled across this site : http://members.aol.com/maxfro/private/butter.html, which has an online butterworth filter calculator. I have no idea of the accuracy, but the results seem to make sense.
As to the phsical size of the inductors, I suspect DC ohmic losses in the conductors will dictate the conductor size.
Has anyone any experience in winding inductors with a DC current flow on ferrite rods ? As the flux should be limited by the air gap, will the change in inductance due to DC current (varying with modulation of course) be enough to cause a problem?
This would decrease the physical size of the inductors considerably.

Posts: 44
Joined: Fri Jan 31, 2003 8:32 pm
Location: Melbourne, Australia

Postby blaine » Sun Feb 09, 2003 9:47 pm

this is from Frank GFZ.

Mike ZE my filter is 2 section 32 UH connected to the modulator and 24 UH connected to the final.
I put 20 UF to ground where the 2 inductors join.
(20 1 UF / 200 Volt film caps in parallel) The output of the filter is about 8 UF made up of film and ceramic caps. It starts to roll off about 8 KHz. It is set up for 2 or 3 ohm load.
Want to double the output Z, double the inductors and cut the caps in half. Want to double the frequency of roll off to 16 KHz reduce all values to half. Easy. The 2 inductors are wound on 3 1/2 inch O.D. PVC drain pipe. The 32 UH is 25 turns of #6 insulated covering 7 inches of form. The 24 UH inductor is 18 turns of #6 insulated spaced over 5 inches of form. I used #6 solid copper ground wire that had gray insulation. I think it is used in telephone systems because the insulation isn't that thick. Testing your filter.
Sweep the modulator and watch the filter output
with a scope. A wrong LC ratio will show up as pass band ripple or a large peak just before it starts to roll off. I would think normal operation the response will show a slight peak
just before it starts to roll off. When you change your power recheck the filter. I'm guilty of this one having jumped from 200 watts out to over 300. I have never had switching crud reports from locals so think this type of configuration is a good place to start. Additional sections or traps can be tried to improve switching noise attenuation. Like any other filter layout and grounding is critical for good results.I mounted the inductors at right angles and away from metal. Ground connections are short. This design comes from Don White's
little green book on filter design. It is the same filter configuration as the one used in my Viking 2 at 1000 times higher output Z. Your question about winding filters on cores.
I tried rods, pot cores and toroids. The problem is the core changes as current increases through it. This means the inductor changes value as you talk. The V2 uses 7 - 2 1/2 Ui 60 iron cores and
even that low Ui makes them unstable. Air core inductors of that value are quite large. I wound these values 32 mH and 24 MH once and it required 6 inch forms over a foot long each. I tested them on my rig and was told they sounded better. Pot cores with gap were also tried but still like air core for stability.
This filter should be good for 50 or 60 amps of current (average). I would think the caps would be more of a limit unless you use MLCC types.
Blaine N1GTU
"The box said that I needed to have Windows 98 or better...
so I installed Linux." --- LinuxNewbie.org
User avatar
Site Admin
Posts: 98
Joined: Fri Jan 31, 2003 1:13 am
Location: CT

Return to Class E Discussion

Who is online

Users browsing this forum: No registered users and 1 guest