Just wondering if someone can tell me why there are no 3x3 MIMO antennas on the market much greater than 20dBi ?
I have a couple of RB921UAGS-5SHPacD-NM(triple chain capable) doing about 8KM’s point to point, but limited to 2x2 due to antenna limitations(cant find a commercial 28 to 30dBi antenna with 3 connections!! )..
Cheers
Maybe because the polarisation insulation must be high enough which is not achievable while 3rd chain is turned only by 45 degree to others?
And maybe 3x60 degree is still not enough…
I hear you, that said, the isolation issue @ 45 degrees must be enough otherwise you would not design such a radio device with 3 chains.
The item below is a 17dBi, which they say is better than 26dBi. I think that’s pretty good gap as most 2x2 big 30dBi parabolic’s have approximately 30-35dBi cross polarization isolation.
If only we could do a beefed up version ~30dBi …
https://www.alibaba.com/product-detail/5GHz-17dbi-3x3-Mimo-Panel-Antenna_60349088440.html
Well. Buy it and try and come back here to tell how it works. I suspect that the throughput will be better with central chain off. But maybe it will be so good that three chain could work independently. We cannot judge before the tryout.
For 30 dBi you should look for a parabolic offset or prime focus feeder.
Something like that
http://antenewireless.ro/index.php?route=product/product&path=170_181&product_id=377
https://www.ubnt.com/airmax/airprism-sector/
This one is 22db did you check it?
Three separate antennas in one box facing each other at different direction are not usable for the required purpose.
Its tempting, but I already use a 28dBi + a 30dBi for my 7km link. So I know the gain needs to be more..
This looks interesting as an option as a feeder. I wonder if the offset so far up is for phase offset rather than traditional polarization offset or a subset of both or neither ..
SPEC’s ***** “-> Port 1 to Port 2 → 24 dBi ; Port 3 to Port 1, 2 → 45 dBi”
Either way the spec’s look ok’ish(I would have expected port 1 to port 2 to be closer to 28-30dbi ! ?) → Port 1 to Port 2 → 24 dBi ; Port 3 to Port 1, 2 → 45 dBi
Now where to find some as a complete 30dBi parabolic s with such a feeder (In Australia ideally ) for testing ???
that feeder is a bullshit. two connectors parallel to each other, no phase difference as i understand
Seriously !! Understand that you used this type of feeder.
Or do you just give it your opinion about what you see in a picture?
How did you find that there are no phase differences between the connectors?
3x3 MIMO is for local use, where you have an access point in a room or outdoor area and antennas that can radiate
in 3 different patterns. The different clients can each have different use of these patterns.
With point-to-point, 3x3 makes no sense. There are no 3 different polarizations that you can independently use to connect
the two endpoints.
When you look at commercial products for mimo e.g attached picture for a 3x3 device all antennas are(generally) VERTICALLY POLARIZED. Yet they are spaced out to receive signals that are potentially out of phase with other antenna(if one or two antennas pick up better than others it uses them), as well allowing TX on any of the antennas to conduct BEAM FORMING(Concentrating RF into directional pattern) (https://en.wikipedia.org/wiki/MIMO) .
So can a feed assembly and associated physical antenna(e.g parabolic or something like it) be designed in a way to utilize this for gain in a p2p scenario?.
That said I see some people are doing some experiments in this area.http://alog.ipduh.com/search/label/3x3%20MIMO
making and testing feeds for point to point work (with not a great level of initial success) for DUAL antenna designs.
This is a badly based experiment, the construction based only on some images, without any technical detail .
On the other hand, there is absolutely no connection that MIMO 3x3 is exclusively for indoor.
On the contrary, an oudoor link is much more feasible because there are two identical equipments.
I do not think Net Metal 5 triple chain it was created for the indoor, for example !!
WTF, I do not want to go into technical details, but if you really need to use MIMO 3x3 equipment there is a solution to turn a 2x2 feeder into a 3x3 feeder.
It’s simple but it’s not cheap.
The first thing you need two 30 dBi dish antenna . I recommend MTAD-5G-30D3 because their feeder has much better port to port separation than other similar products (ubnt !!).
That’s the cheap part of the problem !!
Now you need two hybrid couplers to work in the 5 to 6 Ghz range. That is the most expensive part.
You find this at various producers such as Minicircuits, Narda, Kathrein … depends on what you find in your area, or … ebay.
The rear port of the feeder is coupled to the MIMO 3x3 device via the hybrid coupler. So you’ll have 2 ports connected to your equipment. The third port connects to the front port of the antenna feeder. Now you’ve got a 3x3 feeder with the worst isolation between ports due to hybrid coupler quality. This is one of the ways to make a MIMO 3x3 feeder for 5-6 Ghz range .
I’m not sure if you understood very well what I explained, the main thing is to understand that the polarizations do not matter if the isolation between chain is big enough to satisfy MIMO protocol.
Why not to take 6 single polarisation dish antennas, 3 on each rod at the same direction, turn them 60 degrees apart and put them towards to be sure that 60 degree distance is or is not enough? You might get what you want…
A point-to-point link already uses a beam so there is nothing more to be gained by beamforming.
That is exclusively for indoor use where you have multiple stations in different directions.
Outdoor MIMO uses polarisations that are orthogonal (H and V, or L and R) to have two transmission
paths for the same channel. But you cannot have 3 paths that are “isolated” because there are
always only two orthogonal polarisations. And so there is nothing to be gained using 3.
Therefore I suggested to use 60 degree instead the 45. I expect it should provide better isolation with 3 channels then any other combination.
The crosstalk is cos(angle) related so at 90 degrees there is “theoretically zero” crosstalk (in practice those 20-30dB figures) but
at any other angle there is substantial crosstalk. At 60 degrees about half of the signal is present.
That is the biggest one I know
https://wireless-instruments.com/en/index.php?com=produkt&pg=antena&aid=56&pdf
And the reason why I try the compex 4x4 Modules with ROS, but the driver did not find this chipset
I suspect that is a correct assumption. If you were transmitting simultaneously on a Vertical & Horizontal polarization and had a 3rd element(dipole) say at the receiver end connected to a spectrum analyzer. I would assume that as you rotated the dipole from one V / H angle to the other, as you started rotating towards the other polarization you would start to see the interference of the opposite pol starting to show up as well as the reduction of the original starting pol. I would then assume at about the half way mark you would see about a 3dB reduction between each pol’s carriers, and a high noise floor caused by the ‘mix’ of both pol’s carriers.. I would say the noise floor SNR is going to be the killer at the 45 degree mark. Injecting a carrier at 45 degree’s would therefor interfere with the original V & H pol’s too. Even moving a physical feed for say a 3rd chain input(wavelengths at 5ghz) to make a phase angle change(eg>0 to 180 degrees) will still cause an interference issue.
Looking at alternates such as Right and Left hand Circular polarization. https://www.linksystems-uk.com/vsat-polarization/+
That’s going to be the same issue. Not sure how a mix of Circular and linear polarization would go??(Probably interfere at key wavelength distances)??
The only thing I can think of then is to add a 2nd linear antenna at both sites and point it off(>10degrees) towards a RF reflective surface that both sites can see. Thus giving an alternate path for the 3rd chain(which will be weaker, because of the extra distance & the reflector surface absorption). But maybe plausible in a favorable scenario !
Right and Left hand Circular polarisation are orthogonal to eachother, but not to horizontal and vertical!
The attenuation between one circular and one linear polarisation is similar to the suggested 60 degrees offset.
The only thing I can think of then is to add a 2nd linear antenna at both sites and point it off(>10degrees) towards a RF reflective surface that both sites can see. Thus giving an alternate path for the 3rd chain(which will be weaker, because of the extra distance & the reflector surface absorption). But maybe plausible in a favorable scenario !
Yes that is possible, and that is what the indoor solution is doing as well. However it is not very likely that the reflected link can maintain a stable connection.
That of course depends on local situation.