It helps give a better impedance match on the higher bands, mainly 12 and 10 meters.  Since the wire is many multiple half-wavelengths on the higher bands, the impedance transformation needs to be a bit different than the lower bands.  A 100 to 150 pF cap across the primary winding was found to give the ideal amount of adjustment through testing others have done in studying these antennas.  Without the capacitor, you would still have a resonance SWR dip on the high bands, but it wouldn't be as good, more like 4:1 to 6:1 SWR or so, adding the cap should bring it down to 3:1 or better at it's lowest point of the dip.

You'll want one at least rated for 3kV IMO, but the more the better.  In my image in the OP, I have two 220pF, 3kV rated caps in series to give 110pF, mainly because I had them in a box of misc caps..  In larger higher power builds, I use 6kV rated ones.  You can get thes kind of caps from Mouser, Digikey, etc in singles if you want, or get one of the many assortments of caps sold on Amazon/Ebay and such.

100pf 15kv on mine.

I used 2kV ~100pF on mine. I had plans to use the smith chart on the vna to determine it but in the end I winged it and it seems fine.

I have my wire cut a little long for 40 meters, coax jumper to the choke made, coax run to where I’ll feed it, and my transformer mounted to one section of fiberglass pole. Waiting on RG-400 to wind my choke with, and cooler weather to dig the hole.
I’m going up about 10 feet then slope back towards my 40 foot tower to fit it in my yard and keep the end away from the tower.

It was overcast this morning so I got done what I could, still needs a choke.

Swept it with my Nanovna with a counterpoise wire and my choke at the same distance as the counterpoise is long, and with the counterpoise attached to the radial system buried in my yard & saw no difference. I need to shorten it to bring in 40 meters, but the other 3 bands show I need to lengthen it? 40 is what I’m after so once I get it tuned there the other bands will be whatever they end up being.

How long a wire did you use? I use the table here to pick the length specifically to not be resonant.

If you have a resonant antenna then trimming the radiator works to move the dips. If you non-resonant then it doesn't really, tuning the counterpoise or unun ratio has a bigger change in the swr. I think changing the the amount of wire in the air changes the efficiency and less the tuning, as long as you didn't end up resonate accidentally.

I like to plot each band on the nano vna and look at the smith chart, you can see how well your impedance matches. If they are all off in the same direction, that specific installation might need more or less than 9:1 to match well. If they are evenly distributed then I'd play with your CP length. It doesn't change all bands evenly either like trimming the length of a resonant radiator does, I find leaving it on a spool and rolling it in/out to be the easiest way.

The other week on vacation the end fed I put up worked better on 49:1 than 9:1. It shouldn't have but it did. It wasn't supposed to be resonant (aka needing a 49:1) but the height off the ground changes things.

The use of the 49:1 transformer should require a resonant half wave of wire, or at least close to that after adjusting for final tuning, in my experience it's actually just a bit shorter than an true half wavelength.

As for your setup @my65pan a couple things to try:  Maybe raise your feed box up a bit, I always try to use it with the transformer and antenna wire starting at least 4 feet off the ground, your's looks like its right at ground level, and my experience is that shifts the resonance plot a bit.  I also don't see if you have the small compensation coil on yours, but that might help bring the higher bands lower in resonance since I know you're shooting for the CW area of the bands.  If you do already have the coil, maybe add a couple more turns to it and see what happens.

Otherwise, the harmonic are not always perfect, which is why it's a compromise when you tune these and  needing to watch how it changes on all the bands you want to work.  Mine that's cut for 40 meters is best resonant at the bottom of that band, down around 7.050 MHZ, but still below 2.2:1 all the way up to 7.300.  Then on 20/15/10, the best resonance point is more toward the middle or top of those bands, but still workable across the whole band with fair SWR (except for 10 meters, the FM portion is too high SWR to be used without a tuner).

It’s a resonant antenna for 40. Gonna tune it for phone, gave up CW years ago. If 20, 15, or 10 don’t work at all I won’t be too upset because I’ve never cared for 20 and 15 & 10 are better suited to yagis anyway.

Ah I misremembered on the CW bit.  But yea, I usually have to shoot for best resonance lower in the 40M band to get good resonance in the higher multiples.

I got it tuned for below 1:5 average SWR on 40, 20, and 15. SWR is around 4:1 on 10, but I’m in the begging stages of designing a 3 element yagi for 10 so I’ll leave it as is.
Thanks again for the inspiration, my OCFD wasn’t cutting it for me, was going to convert it to centerfed, but this was more fun, and I learned a few things along the way.

I've had a lot of fun playing with my 40 meter with loading coil for 80 meters build while operating portable. It's given me excellent results even running low power.  Beats the pants off any other portable antenna system I've tried.

Thanks again. Antennas have always been my favorite aspect of the hobby. My last 40 meter build was this magloop about 15 years ago. I plan on building another one soon, will post when I start on it. Photos of some of my other projects are on my QRZ page. AC7AZ

Bit of a necro-post, but wanted to add an interesting version I'm toying with.  I been getting great results with the shortened version for 80/40/20/15/10 meters, but was a bit bummed it left out the WARC bands.  So I did some playing around with the concept for a WARC-special EFHW.

At first look, you'd think a 60 meter half wave would also give you 30 meters, but they're a bit off from each other in actual frequency to get there without needing a tuner to touch things up, so I used the concept of the shortening loading coil to make an add-on for a 30 meter half wave element that could be dialed to fit 60 meters as well, and since these bands are so narrow anyway, the slimmer bandwidth isn't a concern.

Then, just to get one more band in there, I made a disconnect point on the 30M element to give you a half wave for 17 meters, effectively making this a 60/30/17 meter antenna.



And here are the SWR scans.





And doing a bit of testing this evening, it does indeed work well.

Here is triple core one I built a few months ago, a fun project for sure:

Bump so this thread doesn’t go into archives.

As for the matching transformer housing, any thoughts on whether it makes much of a difference between using a non-conductive housing such as a molded non-metallic outdoor-rated junction box or a metal enclosure such as a Hammond die-cast aluminum box?

Sure, the toroid "ought to" keep most of the fields confined within the core, but I'm wondering whether there may be capacitive effects from proximity to the walls of a metal enclosure that might introduce unwanted behavior, especially if the toroid windings are separated from the box walls by, say, the thickness of some black electrician's tape.

A conductive housing would simplify the ground connections.

That's it, capacitive coupling effects of the winding on the transformer to the metal box, same reason you don't want to mount the box directly against a metal pole or tower or onto metal siding, etc.

One thing you can do to make grounding a bit cleaner is stick some conductive copper tape on the inside surface of the box face where the coax connection and ground lug go, just a strip big enough to bridge those two holes and bond the connector and ground lug hardware together without having to run wire between them.