9:12 is the "money shot". We can see both sides of the PCB. As I understand it, the trace connected to the center pin is ofset from center so that it overlays on top of the gnd strip on the opposite side of the board. This becomes a 100 ohm transmission line. (This is the mean impedance between each end) The line is also 1/4 wave long. The far end is shunted back to the ground side input by a 1/4 wave (90 degree) strip. This makes the line behave as a voltage type balun. Because 1/4 wave line INVERTS the impedance, that puts the far end at 200 ohms (balanced) and the SMA end at 50 ohms (unbalanced). The tiny stub near the SMA is a few pf capacitor to null the slight inductance of the far end where the transmission line takes a turn and connects to the folded dipole. A very elegant design. Traveling wave theory is very cool. 73
@@IMSAIGuy 1. I suspect the width of the directors (and the reflector) lowers the Q (increases the bandwidth) of the Yagi. I'm not aware of a formula for this. My guess is it's more trial and error as there are too many variables at play. ie. stray capacitance between elements that is altered by their widths. 2. Yes, the via stitching is vital IMO. The reflector and directors are viewed as one solid element this way as opposed to two plates of a capacitor (pcb is the dielectric). Just my thoughts. It would be great to get input from a real RF design engineer. I'm just a ham. 73 OM
The axial lead is capacitively and inductively coupled to one side of the bottom dipole fold. Once around the bifold is one wavelength. Similarly, the contact (vias) to the other end excites a wave propagating in the opposite direction around the bifold. The coupling stub on the bottom is another half wavelength. Without the top traces, these would arrive out of phase and cancel; but the top traces delay each wave by 90 deg. so that they arrive in phase and resonate.
I helped an EE put together an RF quiet room he bought in an alley somewhere. He used it to make measurements of a whole bunch of wifi antenna prototypes we were testing for robot comms links. The building we put it together in has since been torn down... wonder where the RF room ended up?
"It does good wigglies!" ☺ But you didn't mention where you found it? edit: now in pinned comment BTW without you discussing the exact frequency, I might have spent much longer not cottoning-on that most "2.4Ghz" signals are in-fact banded roughly 2.4 - 2.5ghz, so choosing 2.45ghz in the middle for an antenna probably makes more sense than 2.4 alone.
I've never been able to get signed up with Aliexpress. Their sign-up procedure just doesn't seem to work for me, and i don't know why. Could/would you provide the dimensions of the PCB pads and spacings so I can etch some antennas? I have some old FR4 that needs to be put to use. I assume your friend already measured the dimensions used, in order to produce their model of the antenna and has those dimensions handy so there's no need to reinvent the wheel. Thanks.
Hello friend, how are you? I would like to ask if you have any ideas for modifying this antenna to resonate at 2.40ghz, as my idea is to use one of these for the QO-100 amateur radio satellite. Thanks!
You said low return is the sign of a good cal, but by definition the calibration itself causes you to measure zero return, with only the noise of the system on top, as all the load calibration does is measure the return signal and subtract it. Being a bit pedantic here but low return means that the calibration procedure worked but not that it is a good calibration, you'll need to verify it using another standard to do that.
Have you looked at running your VNA via computer software (I believe there are several free ones available)? This would make screen capture etc. much easier
I have also tested this antenna within a big 2.4 GHz antenna survey where i compared the gain and S11 / SWR of different commercially available and DIY antennas and my observations are consistent, this PCB antenna is the most broadband of my candidates and meets the advertized gain of 9dBi. Not sure if i'm allowed to post a link here, or can i mail you ?
Thank a lot of! Will test this kind of antenna on my QO-100 set-up! (Amsat QO-100 geo-stationary HAM radio satelitte, sponsored by Quatar) Maybe a 'replacement' for my 'integrated 3.5 turn-helical' . Had a lot of problems to 'set-up', both the 10GHz-download- and the 2.405GHz upload- antennas. They interferre, one to the other, ... not much, but enough to "have no good results"! (Have you heard about the "POTY antenna" for this prupose?) All of that kind of 'duo-band- antennas are a less or more 'compromises', and will not give you the best results. But, here, on that delicate frequencies, and the 'long way' (36000km each) every tens of a DB is nessecary! Will Report you, what happend! 73 de Markus - db9pz - JN39fq -3mile/5km east of Lx
There is a video on this by a youtuber called hugoland3 with a similar PCB antenna. If you use the one presented here and a bigger dish, it should work with much less than the 20W. The helical antennas are a bit delicate to build the stub correctly, and the patch antennas are difficult to tune because the real circular polarization is only obtained in a very narrow range.
you should hook one to a router and a router/repeater and see how far it will work. My guess would be about ~20x the distance of a standard router whip antenna because of the directivity.
You should find a guy on RUclips named Andrew McNeil. Does a bunch of microwave antenna work and some really trick microstrip/ patch antenna. He hasn’t posted in a while so I hope he’s ok. His channel is binge worthy.
search Aliexpress for 'High Gain 2.4G Wifi Directional Image Transmission Yagi Antenna 10DB 10W'
9:12 is the "money shot". We can see both sides of the PCB. As I understand it, the trace connected to the center pin is ofset from center so that it overlays on top of the gnd strip on the opposite side of the board. This becomes a 100 ohm transmission line. (This is the mean impedance between each end) The line is also 1/4 wave long. The far end is shunted back to the ground side input by a 1/4 wave (90 degree) strip. This makes the line behave as a voltage type balun. Because 1/4 wave line INVERTS the impedance, that puts the far end at 200 ohms (balanced) and the SMA end at 50 ohms (unbalanced). The tiny stub near the SMA is a few pf capacitor to null the slight inductance of the far end where the transmission line takes a turn and connects to the folded dipole. A very elegant design. Traveling wave theory is very cool. 73
I'm not great at RF engineering but I know a bit, and I'd say you nailed it.
Thanks. If you don't mind
1. how do you pick the width of the directors
2. do you think the via stitching is necessary?
@@IMSAIGuy 1. I suspect the width of the directors (and the reflector) lowers the Q (increases the bandwidth) of the Yagi. I'm not aware of a formula for this. My guess is it's more trial and error as there are too many variables at play. ie. stray capacitance between elements that is altered by their widths. 2. Yes, the via stitching is vital IMO. The reflector and directors are viewed as one solid element this way as opposed to two plates of a capacitor (pcb is the dielectric). Just my thoughts. It would be great to get input from a real RF design engineer. I'm just a ham. 73 OM
The axial lead is capacitively and inductively coupled to one side of the bottom dipole fold. Once around the bifold is one wavelength. Similarly, the contact (vias) to the other end excites a wave propagating in the opposite direction around the bifold. The coupling stub on the bottom is another half wavelength. Without the top traces, these would arrive out of phase and cancel; but the top traces delay each wave by 90 deg. so that they arrive in phase and resonate.
I helped an EE put together an RF quiet room he bought in an alley somewhere. He used it to make measurements of a whole bunch of wifi antenna prototypes we were testing for robot comms links. The building we put it together in has since been torn down... wonder where the RF room ended up?
"It does good wigglies!" ☺
But you didn't mention where you found it? edit: now in pinned comment
BTW without you discussing the exact frequency, I might have spent much longer not cottoning-on that most "2.4Ghz" signals are in-fact banded roughly 2.4 - 2.5ghz, so choosing 2.45ghz in the middle for an antenna probably makes more sense than 2.4 alone.
Where can we find this? No details on who makes these.
I've never been able to get signed up with Aliexpress. Their sign-up procedure just doesn't seem to work for me, and i don't know why. Could/would you provide the dimensions of the PCB pads and spacings so I can etch some antennas? I have some old FR4 that needs to be put to use. I assume your friend already measured the dimensions used, in order to produce their model of the antenna and has those dimensions handy so there's no need to reinvent the wheel. Thanks.
Nice, who makes this one?
Hello friend, how are you? I would like to ask if you have any ideas for modifying this antenna to resonate at 2.40ghz, as my idea is to use one of these for the QO-100 amateur radio satellite. Thanks!
More magic and wizardry. But
always interesting to watch. I
always learn a tidbit here and
there. :-)
What about 4+ layer boards?
Merci.
You said low return is the sign of a good cal, but by definition the calibration itself causes you to measure zero return, with only the noise of the system on top, as all the load calibration does is measure the return signal and subtract it. Being a bit pedantic here but low return means that the calibration procedure worked but not that it is a good calibration, you'll need to verify it using another standard to do that.
Will be nice if you share measurements for experimentation.
I 3D printed a yagi to my 2.45 endpoint and it works.
One should use 2 of them and measure the transfer characteristics versus distance with the VNA to see if the antenna has really a gain of 10dB.
Have you looked at running your VNA via computer software (I believe there are several free ones available)?
This would make screen capture etc. much easier
I love this series about antenna's 😋
I have also tested this antenna within a big 2.4 GHz antenna survey where i compared the gain and S11 / SWR of different commercially available and DIY antennas and my observations are consistent, this PCB antenna is the most broadband of my candidates and meets the advertized gain of 9dBi. Not sure if i'm allowed to post a link here, or can i mail you ?
Can you post a title of the article/+ author but thanks for doing the research for us! I'm looking for a cheep higher gain antenna for 2.4ghz wifi.
How do you know it's a good yagi? A good VSWR is a necessary but not sufficient condition. We would also like to have a decent pattern...
you are right, but at least it got one thing right
There's an Application Note DN034 from Texas Instruments describing a Yagi 2.4 GHz PCB antenna. Please search for that paper and DIY !
The photons go in the direction of the arrow - except on receive.
great 👍
Thank a lot of!
Will test this kind of antenna on my QO-100 set-up! (Amsat QO-100 geo-stationary HAM radio satelitte, sponsored by Quatar)
Maybe a 'replacement' for my 'integrated 3.5 turn-helical' . Had a lot of problems to 'set-up', both the 10GHz-download- and the 2.405GHz upload- antennas.
They interferre, one to the other, ... not much, but enough to "have no good results"!
(Have you heard about the "POTY antenna" for this prupose?)
All of that kind of 'duo-band- antennas are a less or more 'compromises', and will not give you the best results. But, here, on that delicate frequencies, and the 'long way' (36000km each) every tens of a DB is nessecary!
Will Report you, what happend!
73 de Markus - db9pz -
JN39fq -3mile/5km east of Lx
There is a video on this by a youtuber called hugoland3 with a similar PCB antenna. If you use the one presented here and a bigger dish, it should work with much less than the 20W. The helical antennas are a bit delicate to build the stub correctly, and the patch antennas are difficult to tune because the real circular polarization is only obtained in a very narrow range.
If you can find a brand name it might be fun to reach out to the inventor and see if he wants to do a follow up video with you.
There is a lot of microwave theory in use here. More than I can explain simply.
you should hook one to a router and a router/repeater and see how far it will work. My guess would be about ~20x the distance of a standard router whip antenna because of the directivity.
You should find a guy on RUclips named Andrew McNeil. Does a bunch of microwave antenna work and some really trick microstrip/ patch antenna. He hasn’t posted in a while so I hope he’s ok. His channel is binge worthy.
How could you not know about it's directivity? I had an arrow RIGHT THERE on the PCB 🙂