Common Mode Choke Question

That they have nonzero leakage inductance? ;0

Cheers

Phil Hobbs

They are cheap, saturate with tiny differential-mode currents, and the ferrites are lossy at high frequencies. Lotta skin loss, too.

Ahh, I get it! I had to remove two, one at each end.

Thank, Mikek

(Megahecto what?)

Yeah, you can tell those are unsuitable, from the fact that they aren't antenna feedline pair wrapped around a core. The ball of wire has a high impedance (regardless of coupling, which is poor, because, again, there's two of them, not a transmission line), starting around, fractional MHz I guess, and extending up (with peaks and valleys) to 10s of MHz. Over that range, it centers around Idunno, 600 ohms or so.

Last time I measured a CMC that style, it had Zo ~ 600 ohms and Fc ~ 3MHz or something like that, I forget exactly.

Transformers are bandpass components with characteristic impedance; a 2nd order model of which looks like so:

Consider using a data-line choke instead, which however usually isn't rated for much power, but is usually made of twisted pair, so the DM mode cutoff is much higher (100s MHz?) and centers around an impedance of 100 ohms or so.

Or use a custom part, with Zo and Z_CM as needed. That is, feedline wrapped around a high-mu core big enough for the wire and turns. Remember to space the feedline so it's not just overlapping, else turn-to-turn impedance will be lower than differential impedance already; this shows how difficult it is to make good high impedance transmission line transformers, and better performance will be had at say Zo ~ 100 ohms (where you can use regular twisted pair). If this is feeding a resonant dipole say, consider putting it between the balun (or I guess, balbal?) and elements, so it can be made to that impedance (~75 ohm?). Or use a balun to coax feedline, and put some coax through a core.

Tim

Specifically coax, which ideally has zero coupling between the normal (odd) mode and adjacent turns. Wouldn't work as well with twin lead.

Cheers

Phil Hobbs

Twisted pair is OK too, as you say, provided that either the pitch of the twist or the diameter of the winding changes rapidly enough to prevent significant coupling.

Cheers

Phil Hobbs

Just slap a clamp on ferrite, (suitable material), around the coax or feed line or run a turn or two thorugh a ferrite toroid. Of the proper material of course.

Quick and easy

I don't know the details of frequency and such, but I understand a lot of hams make a common mode choke by wrapping a few turns of feedline around a 2 liter soda bottle.

I wonder what the loss/frequency characteristics are for soda bottles ? Might be good ? :)

Ferrites for the purpose works well.

For instance Fair-Rite 10-300 MHz (44 material)

Frequency Rloss

So many to choose from !

Works well enough for other capacitors that you can buy 'em off the shelf. Polyester is usually either precisely the same stuff (polyethylene terephthalate) or a similar relative (naphthenate something or other I believe is another common one).

It's not particularly low loss, but the electric field path is mostly through air. It might melt if you stuck it in your /microwave oven/, but

Yep, #43, 44, or higher mu since this is still pretty low frequency. 77 is absolutely fine; if you're using it for signal power, do check to make sure Bpk is low. And anyone else's equivalents, 3C90, 3F3, P, uhhh whatever numbering Laird uses, etc.

Note that 10s of ohms won't do much for a common mode choke, hence single or couple turns, or beads, won't do much either (low 100s ohms). You're looking at dozens of turns to make an effective (~kohms) "current balun". Toroids big enough to hold enough plain old feedline are the most promising option.

Tim

Yeah, downside to coax being that because its shield is relatively big, it's not doing any favors to the common mode impedance. Think of making a solenoid or toroid out of solid wire the same O.D., versus thinner wire but spaced with the same pitch. Same goes for the common mode on twisted pair, the distance between turns relativel to whatever counts as the average conductor surface. Space between turns keeps the winding CM impedance high.

Upside to the twist rate stuff, it's not a big deal unless the winding's electrical length is a significant fraction of the intended (pass and stop bands) wavelength. Just do whatever. If it's marginal (the ol' 1/10 to 1/4 wave range), it's going to make some difference, and keeping them spaced (between turns, and between twists) will help greatly.

(Or, of course, if the SWR or impedance or whatever has to be very closely matched. Those also get more important the closer to critical wavelength you are. I'm assuming that's not a big priority here, but it is definitely a consideration for say, baluns for balanced mixers.)

Tim

These are pulls form new old stock, or just randomly busted up? (the one in the left foreground is visibly damaged, the loner looks suspect)

leakage inductance? inter-winding capacitance? lossy dielectric in the spool?

I think the point is that most plastics are not high loss and a soda bottle is mostly air anyway. I believe I saw one made of PVC pipe, something a fairly good diameter, like 6 or 8 inches. Lots of air in that too.

I suppose if you want to get seriously extreme you could use a few thin slats of wood connected by supports on the ends. Lots of air in that too.

Just a followup of what I'm doing. Beverage antenna on the Ground. A box with a 3.3 to 1 matching transformer, (then I had a CMC, since removed.) After the transformer I have a pair from in a CAT5 cable. In haste I connected the pair directly to my radio. I had lots of feed line signal ingress. I was told to put an isolation transformer at the radio. This worked wonders. A 230 ft,

Yup. I made a CMC for our city-ARES ham station a few years ago... RG-58 or RG-8 (I forget which) run through the center of a series of hefty ferrite toroid cores. I was able to do three complete loops around the ferrites before the core centers ran out of room. Put the whole thing in a short length of PVC water pipe, with end caps and SO-239 connectors.

The common-mode impedance was so high at HF that my MJF meter saw it as an open circuit. The through-impedance wasn't detectably different from 50 ohms and losses were negligible.

Doing this with a twin-lead would be much more difficult.

The twin-lead field is so far outside the wires that it is probable that it's not going to work. Coax keeps the energy inside the shield.