#1456 PCB Yagi antenna for 2.4GHz

Episode 1456
I bought this off eBay. Let's test it.

[subramanianr]

Hi, I want to make some clarification here.
The yagi's driven element has the J shape and has a feed point impedance of around 150 ohms in free space. (The regular folded dipole structure is employed for a wide bandwidth (as in analogue TV receiving antenna) and has a feed point impedance around 300 ohms) Here in this antenna impedance matching (tweaking for low SWR) to a 50-75 ohm coaxial cable can be done by trimming a bit off the free end. That's why the free end of it is shorter compared to the first director.
The explanation for the working of the yagi-uda antenna itself needs clarification too.
The dipole (in free space) alone has a feed point impedance of around 75 ohm +j43 ohm and has a directivity gain of 2.15 dBi.
In the form of a yagi uda directional antenna the feed point impedance of the dipole (driven element) gets lowered to around 35 ohm or even less depending on the number of parasitic elements (reflector and the directors) you are going to mount on the boom to achieve more directivity gain in a particular direction. This is because the dipole current is getting shared among the parasitic elements by mutual coupling.
The spacing between the driven element and the reflector is worked out to around 0.2 lambda; the spacing between the driven element and the first director is around 0.15 lambda. It may be varied for a specific Front to Back ratio and cancellation of the side lobes.
The signal from the dipole tends to propagate in all directions and particularly in both forward and reverse directions in the plane of the yagi antenna. The reflector (usually being 5% larger in size) and the typical spacing result in interference with the signal emanating from the active driven element. Part of the reflected signal after a small delay constructively interferes with the signal on the driven element and moves forward. The current on the director element too constructively interferes with the signal from the driven element and all move forward as if guided along the boom. With each half cycle the wavelets alternate the polarity but keep moving forward as if a light beam (after all light itself is the visible part of the EM spectrum) is being directed by a parabolic reflector and condenser lenses etc.
The centre of the boom being at isoelectric potential can have the centre of a folded dipole or the centre of the driven element used in this antenna and all the parasitic elements' centre anchored to it.
Any small metal fragments scattered in between the elements unconnected to each other and to any of the elements or the boom would be invisible to the RF energy propagating in the forward direction. So, the letters having copper in them would not affect the performance.
This 2.4 GHz Wi-Fi yagi uda antenna must have produced the S11 pattern as found in the literature that came along; I don't know why you didn't get it right.
Moreover there's no ground plane or the need for a ground plane as such.
A dipole is a balanced device. The coaxial cable (intrinsically unbalanced device) has both the hot (centre conductor) and the cold (outer shield braid) terminals that are connected to the dipole at the center.
The half lambda dipole has maximum current in the middle and maximum voltage at the ends; the current and the voltage are 90 degrees out of phase. Because of maximum current and minimum voltage at the center the impedance is low, ie, around 75 + j43 ohm. The RF power fed to the feed point of the dipole is dissipated in the (virtual resistor) Radiation Resistance; yes at resonance the inductive reactance and the capacitive reactance are getting cancelled in the otherwise series resonant half lambda dipole antenna elements.
The voltage at the ends maybe around 2500-4000 ohm or more depending on many factors like height of the feed point above the ground, nearby RF absorbing, scattering or reflecting structures.
The EFHWA antenna is fed with a 49:1 UnUn for this reason.
De VU2RZA

[Randrew]

I found this video (and its comments) to be very informative and I thank you for it... any mistakes or incomplete explanations be damned. The video started a fine conversation and the comments have been filling in the gaps. This is all good :)

[nezbrun]

WA5VJB (Kent Britain) is an RF guru particularly when it comes to microwave antenna design.

I spoke with him about his antenna designs maybe 15 to 20 years ago. The "strange" feed mechanism is by design: it offers a balun function, impedance match, and makes it a little more broadband, at the expense of a slightly off-bore direction, which is insignificant for this relatively low gain.

Of note is that you must remember that the PCB substrate is not benign here: it has a significant effect on the antenna's performance in terms of impedance matching and size due to the relative permittivity of the dielelctric.

I've had lots of success with his many Yagi designs over ht past couple of decades, from VHF to SHF using his designs: they are the most reproducible Yagi antennas I've ever built.

Basically, he _really_ knows his onions: I'd be very careful about second guessing his expertise in this area!

[kbritain]

You keep drawing the driven element as a full wave antenna, IT IS A HALF WAVE! The plots on the data sheet were done with a Rhode & Schwarz FSL analzyer. Letters are only 1% of a wavelength wide. Way to small to affect an EM wave. Make the center conductor of your coax shorter and Return Loss / SWR will look better. Kent WA5VJB

[byronwatkins]

The ends are voltage (E field) anti-nodes and current (B field) nodes. The center is an E node and B anti-node. The total distance across is half of a wavelength and not a full wavelength. As a first approximation, the reflector acts like a mirror and places an image of the dipole (and directors but with 1/r^2 intensity reduction) behind itself. This mirror changes the charge sign and the distance delays the phase, however.

[electronics.unmessed]

There is not the one and only dimensioning for Yagis. You can optimize it for best match, maximum gain, maximum front to back ratio. And the dimensioning is always a little different.

[JumpingJack-wl]

Active element looks like end fed antenna. Also coax feeder may transmit/receive and shape of feeder may impact antenna's parameters if PCB path do not balance line enough.

[BalticLab]

0:51 "At least S11" -> What other s-parameters could he have supplied on a single port device?

[ats]

It's really easy to get PCB boards built. It would be interesting to go though the design process for this type of antenna. As I recall, there are charts that give the spacing and length of the elements, but I'm not sure how much the FR-4 dielectric will influence these numbers. Of course if it was influencing the design, a lot of it could be milled out from the board...

[danajohnson]

That was a good explanation of Yagi antennas and an interesting video. Thank you.

[willthecat]

Why is the dipole part of the yagi shortened? ... well... WA5VJB's design is using... AFAICT... a "gamma match" (the folded part)... and that bottom leg (the parallel part) adds reactance. One can sometimes reduce/compensate for that, by shortening the antenna length. Seems like that is what WA5VJB did. It's hard to get a gamma match just right.

[RealRobotZer]

5:14 Isn't the center suppose to be at max amplitude, if it is zero that means you have no signal? (when receiveing)

[KeepEveryGuessing]

1. Reflectors don't, in general, need to be grounded. Like directors, they couple just fine floating. Usually all elements are grounded on lower frequency Yagi-Uda antennas to prevent static build up.
2. The driven element here is an end fed half wave with an approximately 1/4 wave impedance transformer to produce the high impedance feed.
3. The length of the parasitic elements go according to the Yagi formulae, because that's all that controls their resonance, and thus, phase shift. (reflectors lead and directors lag so that the traveling wave in the desired direction passes each element in phase with what that element is re-radiating. The driven element need only be resonant, and in combination with the drive impedance, is whatever length this requires, not necessarily in the Yagi sequence (it comes closer to that with the typical gamma match used at lower frequencies).
4. If your hand was anywhere near the antenna when you took your readings, such as holding it by a corner of the board, or by the COAX at the edge of the board, then your measurements are worthless.
5. Don't confuse voltage with field. There is, , for instance, E field near the center of the element, it's just parallel to the element rather than radial.

[jakubniemczuk]

Are those "wrong looking" dimensions of the dipol and the directors not for impedance matching? An ideal dipol doesn't have 50 Ohm. Also the folded radiating element is not making any balun thingi so the coax shield doesn't act as part of the antenna?

[RechargeableLithium]

It would be interesting to see a Smith chart for this antenna. 73

[roboticus]

The dipole doesn't appear to have the same centerline as the first reflector ... it seems to be shifted to one side a bit. (Can't be certain as I used measurements from my screen and the angles are usually oblique, but it certainly appears that way in multiple views.) I wonder what *that* would do to the pattern. (I'm not an RF guy and have no expertise there.)

[ovalwingnut]

Antenna theory: "the gift that keeps on giving". Must see TV is back. Thank you for the gray matter transfer.

[philoso]

If unsure. Do take a base line measurement with open/short/load standards and test the antenna again.

If still unsure, recalibrate the measurement with same standards then base line test with just the load, to expect -50dB or better before testing the DUT.

Hope this helps.

[electronics.unmessed]

Another issue is that the PCB materials epsilon_r is not exactly the same in each charge. That might be a reason that the matching is not exactly as it is promised.

[danishnative]

As I recall, Kent Britain is a RF EE by career, with amateur radio as a hobby side. He was actively involved with the North Texas Microwave Society (NTMS) for may years. Why not reach out to him by email?

IMO, a simple Yagi constructed on a dielectric like G10-FR4 is a compromise antenna, requiring empirical experimentation to optimize it.