Plate choke designs

Plate choke design.

By Matt Erickson, KK5DR.

Traditional design:

The plate choke, as used in an HF tube type amplifier in this article. The plate choke is a device which serves the purpose of feeding the DC plate current back to the HV power supply, since DC current only travels from negative to positive voltage potentials, the plate choke is in the return path for the current that has done the work of amplifying an RF signal inside the vacuum power tube. The choke must pass DC current, but block RF currents. To do this, the choke has an inductance that is high enough to block RF on the lowest frequency the amplifier will operate on. The high inductance presents a high impedance to the RF signal. This all sounds straight forward, but it is more complex than it sounds. The choke has extreme levels of RF forced against it. A small amount of RF current sometimes enters the choke, which can cause a problem if the frequency of the RF applied to the choke happens to be close to or on the series resonant frequency of the choke.

The problem with traditional plate choke design is that to get enough inductance to block RF on the lowest operating frequency, these chokes have a certain amount of distributed capacitance in the wire windings. Usually at some point the inductance and distributed capacitance act together to form LC circuit, what is known as "series resonance" in a choke. At a certain point in the choke windings the resonance will allow a large RF current and or voltage to exist, many times this will heat the wire to the point of melting, or arc to near by metal part. Either way, this is not a desirable condition.

Most traditional plate chokes in amplifiers today have a range of frequencies that exhibit this series resonance. The amplifier designer will often design a circuit that breaks up the distributed capacitance on the higher frequency bands where less inductance is needed, by switching in a capacitor to ground or jumping a section of the choke with a switch contact or relay.

Here are a few examples of chokes and the series resonate frequency;

This choke covers 1.5-30MHz (except 10.5-13MHz & 15-17MHz).

It will carry 1.5A of current.

This choke covers 14-54MHz (except 24MHz).

This choke also covers 14-54MHz, but is resonate at about 96MHz.

This is a "Layer Pi" wound choke, but can't be used as a plate choke for high power amps.

All of these chokes are wound in what is known as a "solenoid" wire winding. The choke at the top is what is known as a "Pi" wound choke, with spaces between the windings. Pi-winding helps to breakup series resonance but does not always work at all HF frequencies applied to them.

The "layer Pi" choke is used in low power circuits or as a DC removing path in single-ended tank circuits. These typically have very high inductances, and also relatively high DC resistance.

All plate chokes have a certain value of DC resistance as well as inductance and distributed capacitance. The wire used in a plate choke must be able to carry continuously the full load DC plate current. Wire that is too small will heat up and likely ruin the choke. Wire that is too large will not have enough inductance to block the lowest RF signal frequency. Typical wire sizes for traditional plate chokes are 18-24Ga. enameled/Formvar type wire.

To determine if and where a choke has resonance, the choke can be tested with a grid dip meter held near the choke while in the tank circuit, or shorted end-to-end or through a .001µF capacitor. Then the GDM is swept through the operating range of the amplifier. If the meter dips at any point in that range, it is highly likely that it is a series resonance. Hopefully, any dips found with this method are outside any amateur bands. If a dip is found on a band the amp should work on, do not operate the amp on that band until the choke it replaced with a choke that does not have any resonance points within or close to the bands it will operate on.

A NEW plate choke design:

You will notice that this choke has been wound on a hollow ceramic tube, or a tube of electrical grade fiberglass. On it's own, the choke has about 90µH of inductance, and enough distributed capacitance to make it resonate at about 24MHz. This would be a problem if the amp is to be used on the 12 meter band.

I have developed a new type of plate choke that can make this choke work. I use common ferrite rods used for filament chokes about 1/2" O.D. and at least 4" long. I slip the rod into the hollow choke tube. Chokes that I have tested and used with this method go from about 90µH up to 1200µH. But, the choke still has the distributed capacitance of the original lower inductance choke. Using a grid dipper, I found that the choke now has a resonance that has dropped down to about 500kHz, due to the high inductance provided by the ferrite, and the low distributed capacitance which causes the resonance to be very low in frequency. This creates a very low Q resonate circuit, so little RF energy would ever be present in the circuit. The grid dipper showed no dips anywhere in in the MF, HF range. I have used several of these in various amplifiers, and using various wire sizes and inductance levels, none of the chokes have ever had a series resonance problem.

The ferrite rod MUST be insulated from the HV and RF, at least 1/16th" of insulating material like the ceramic tube, fiberglass, or other good tubular insulating material will do the job. The rod must also be secured inside the choke coils, if it moves, even a little, the inductance will change radically.

One might ask, " does the ferrite heat up?" Well, no, because there is no RF current flowing in the coils of the choke, nor in the flux core of the ferrite rod. I have found that even at 1500 watts output, the choke simply does not heat up at all. This was proof to me that there is NO RF current in the choke, this is key to preventing series resonance. The high level of inductance effectively blocks any RF current from entering the choke. The resonance point of this choke of 500kHz, will not likely ever produce any problems, this is mainly due to the fact that no RF current is in the coils to start with, but also because resonance tend to go up in frequency, not down. A downward resonance would be very lossy, which would then "kill" the RF current before it could begin. Tube type amps are very narrowly tuned circuits, so RF energy outside the range of the tuned circuit would be greatly attenuated.

One of my most recent choke construction units was a 1" O.D. x 8" Delrin® rod. I bored the rod out to a depth of 5.5" with an I.D. of 33/64th", just over 1/2", this would hold a 5" long ferrite rod nicely with room to seal the open end with silicone rubber. I wound it with 22Ga. Teflon® wire (don't ask me how many turns, I wound about 6" of coil). The wire is silver plated and stranded for good flexibility on the free ends. Before I installed the ferrite, I measured the inductance at 40µH. Then I installed the ferrite and measured the inductance again. The second time it is 365µH. This is more than enough inductance for the entire HF spectrum from 1.8-30MHz. I then used the grid dipper to check for resonance in that range, first with the end wires shorted together, and second with a .001µF cap connected to both ends. I did not find any resonance in the spectrum either way. After I install the choke in the amplifier circuit, I'll run the test again to be sure before operating the amp.

How to build your own:

Use one of the following forms; A hollow ceramic tube at least 6" long with an ID of at least 1/2". An O.D. of 1" or more. The wall thickness should be no less than 1/16th", more is better, but not thicker than 3/8ths". The same parameters apply to materials like electrical grade fiberglass, Delrin®, & Teflon®. The form used will have to handle some heat generated by the DC current flow in the wire used. Remember, the full input power of the amplifier will be flowing through this small wire. The use of the ferrite rod will allow the builder to increase the size of the wire, which will lower the DC resistance, which in-turn will reduce the heat generated. However, this will also lower the total inductance of the choke.

Secure the rod inside the form with some hot glue, or silicone rubber. The wire will need to be secured to the outside of the form as well. I use a section of heat-shrink tubing of the proper size to shrink down tightly over the wire windings. Silicone rubber could also be used. What you use is up to you, but you must remember that the material used must be able to handle some heat, and insulate from HV as well as some RF voltage exposure.

The method of securing the choke in the amplifier near the tube and tank circuit is up to the builder. I prefer drilling and taping a hole in the bottom end of the form, which will allow it to be bolted to the chassis, standing up vertically. I take care not to allow the mounting screw to touch the ferrite rod, this may provide a path to ground for an HV arc to follow. Should HV ever find its way to the ferrite (which is electrically conductive), it is less likely to happen if the path is of a very high resistance, on the order of several hundred Meg-Ohms or more. We are talking very high voltages on the plate choke, so think insulate, isolate, and spacing.

The wire size used will be determined by the DC current level it will need to handle. I have found that 22-24Ga, will work fine for up to 1.5A. The wire must at least be insulated with enamel, but can be Teflon® insulated, but this will cause less turns to be used per inch.

The level of inductance your aiming for is 200µH or more, preferably much more. I have found that 100 turns on a 1" O.D. form will get you in the range needed. That many turns usually covers about 4-5" of the form.

To make the job easier, you would need a grid dip meter, and an LCR meter. These meters are highly valuable for amp building work, making the job much, much easier than without them. I purchased them years ago since I do so much of this type work.

I have developed a more advanced test procedure for plate chokes, feel free to check it out.

Good luck.

73 de Matt KK5DR