[Amps] Ferrite-loaded plate choke

[Will Matney]

> Hi,
>
> I am currently building my first HF amp using a couple of 4cx 250b.
>
> I have found in the literature that the plate choke is quite a critical
> component and I found a few kinds of plate chokes designs.
> Recently I found this text (see below) on the Internet.
>
> Now I saw that the use of ferrite cores is classical in filament chokes,
> but I never saw a ferrite core recommended in a plate choke.
>
> What do the experts on the list think about using a ferrite core in a
> plate choke?
>
> Plate choke: High RF current rules this circuit, and makes the design of
> the plate choke much more difficult. The trends have been to reduce the
> inductance to the bare minimum required to block RF out of the B+ line.
> However, this trend in design has serious drawbacks due to the tendency
> of the plate choke to resonate at certain frequencies within bands that
> the amp operates on. I have found that increasing the inductance to very
> high levels by using a ferrite core works well. The tough part of using
> ferrite in a plate choke is insulation, enough to prevent an RF arc to
> the core from happening.


 From my experience, If I recall, ferrite is non-conductive (it's 
actually a ceramic). The only coated or plasticized cores that I know of 
are iron powder and strip wound.


> If enough insulation is installed, a ferrite
> rod will give a large amount of inductance with a relatively small
> number of turns. The lower number of turns for a high inductance gives a
> much lower "distributed capacitance", which is a major factor in choke
> resonance at high frequencies. My experiments have proved that for an
> amp operating 160-10 meters, a plate choke with an inductance of
> 800-1200µH is entirely stable on all HF ham bands. The problem with a
> choke with this level of inductance and able to carry the high level of
> DC current would be very large, and have a high level of distributed
> capacitance. A hollow core of an insulating coil form can hold a
> 1/2"dia. ferrite rod about 3-4" long. This rod will bring the typical
> inductance of a 90µH choke, up to 1200µH, with the distributed
> capacitance of the original 90µH choke. However, the higher level of
> inductance would cause any resonances to be far out of any HF ham bands.


Ferrite and iron powder cores are frequency selective. Also, they 
saturate at low flux densities, say around 1800 to 2700 gauss. This then 
means the max DC current flowing through the choke has to create that 
amount of flux or less. Generally, it's better to figure for a lesser 
amount, say in this case 250 to 500 gauss less. This way you know the 
core will not saturate at it's maximum current rating. The material 
selection used in most RF applications is numbers 61, 63 and 43. Their 
permeabilities are 61 = 125, 63 = 40, and 43 = 950. Their maximum 
saturation flux densities are 61 = 2350, 63 = 1850, and 43 = 2750. These 
have all been used in broad band transformers and would be used the same 
in RF blocking chokes. They recommend that 43 material be used for 
frequencies between 2 and 35 MHz, and the other two between 6 and 30 
MHz. The iron cores mostly used is number 2 mix (color red) and 6 mix 
(color yellow) with the following characteristics. 2 mix, u = 10, and 
freq. = 1 to 30 MHz. 6 mix, u = 8, freq. = 10 to 90 MHz. As you can see 
the iron cores have a low permeability. Circuit Q will be degraded if a 
core type is applied below its recommended frequency. However, cores can 
be used above their suggested operating frequency if the designer is 
willing to sacrifice permeability and use extra turns of wire. The main 
thing is to use the proper formulas to determine both the number of 
turns and the core size.

Number of turns;

N = 100 x sqrt of uH / Al

N = number of turns, Al = inductance index by the manufacturer.

Maximum flux density (B max);

B max = E rms x 10^8 / 4.44 x f x N x A + N x I x Al / 10 x A

E rms = applied voltage
f = frequency in Hz
A = core area in square centimeters
N = number of turns
I = DC current in the choke
Al = inductance index

The area (A) can be figured by simply rearranging the equation and 
placing B max in place of A. The area can then be built up by stacking 
toroids, and adding the areas of each cross section, or finding a core 
with the correct solid size. Keep in mind that this area is the area of 
the cross section of any core where its length has nothing to do with 
its operation. On a rod, it is the area of it's circular part or 
section, not the length. The use of a toroid is actually better than a 
rod because the magnetic path is made whole, or connected, and not open 
as a rod is so the rod has higher losses. You will see that the 
inductance index (Al) of a rod is much different than a toroid of the 
same cross section.

>
>
> 73
>
> Angel Vilaseca HB9SLV


Will Matney