Integration with RF Elements Symmetrical Horn Antenna


#1

Fingers crossed your planned A5C AP would live up to all your other products but a customer perspective, Mimosa are missing a trick by not working on releasing the AP to be compactible with the RF Elements Symmetrical antenna which clearly is the best antenna for 5Ghz Urban deployment at the moment.
I know that Mimosa is aware of this as well, the question is why can’t you make the A5C connectorised to matched the twistport adapter integration platform for the RF element Symmetrical horn antenna. From an RF perspective, that is the holy grail you owe your customers.
I have made up my mind to use the RF element antenna anyway but without this adaptation from Mimosa, people like us would have to spend more money to get this done because 2 antennas per AP since Mimosa AP are 4 x 4 while the RF element interface is 2 N connectors.


MTI and KP Performance Sector Antennas
#2

Hello Beshel, it seems there are solutions for RF Elements working with Mimosa radios: http://www.rfelements.com/products/antennas/#mimosa

Don’t know which adapter needed, however.


#3

Hi Vincent, I understand that the Mimosa radio can work with the RF Elements antenna but the current design mean that, one will need to use 2 antennas per AP instead one. That is doubling the cost of Antenna; a simple solution would be getting the twistport adapter for Mimosa radio. The cost of this enhancement is less than $20 rather than you spending additional $220 plus if you factor in the cost you may pay to have additional antenna on the mast or rooftop. Since RF Elements have been talking to Mimosa about this, mimosa has to take this seriously.


#4

@Beshel appreciate the interest, but what you’re suggesting currently simply isn’t feasible. We have a number of antenna manufacturers that have nicely built “native” 4x4 sectors that are connectorized. We welcome all well designed solutions that match our design, that are also well designed for GPS sync applications. Physically there simply is no feasible way yet to have “dual” horn with a twist port. Horns, much like a parabolic antenna by their nature are limited to dual polarization, they are not panels/sectors where we can collocate in a single enclosure 4 antenna sets in an array which we need for beamforming. Furthermore, with the size of the radio, and necessity of bottom fed connectors with GPS on top of our system, it doesn’t make sense for a twist port at all. This is why RF Elements began with their N-type connector solution which makes sense for products like our B5c in multipoint. We continue to interact with RF elements and other antenna partners wherever there is great differentiation and value to be added to our radios for the WISP community.


#5

We like the RFE horns in our close-range (< 1km) city deployments because of their antenna pattern (the vertical width mainly). If we did use 2 carefully aligned horns, would this work with the A5C, or does the beamforming require a special antenna (e.g.: the kp one)? And if so, would it still work with the horns just without the added gain of beamforming?

Rich


#6

Has anyone successfully used two 2 dual-pol RF Elements horns with an A5c yet? If so, I’d love to hear your results as I’m looking at experimenting with a similar configuration. What would this do to the beamforming capabilities in the A5c?


#7

Let’s ask @Kent or @Rory regarding their experience with the horns on the A5c. They’ve both been testing recently.


#8

I’ve got 3 A5c’s running that exact configuration with 90 degree RF Element antennas covering 180 degrees. One A5c has 59 users and only 6 of them are Mimosa C5’s. The other has 39 users and all of them are C5’s. On the unit with 39 users, that was an A5-360 14 but we had too much noise and it was causing issues. We swapped it to the A5c with the RF Element antennas and voila, everyone worked much better. The third unit hasn’t started adding clients yet, new area. The one thing I would suggest depending on range, use the highest gain and most narrow antennas you can with the RF Elements. My clients are all within 1/2 mile with the first 2 A5cs so not a big deal. The third unit will have farther clients and we realized the that 90 degree antenna wasn’t our best option so I will be ordering some 30 degree units Tuesday.

Rory Conaway
Triad Wireless


2 sectors back to back on A5c
#9

What type of service plans are you offering on these A5cs? Have you seen any indication of user limits yet?


#10

We offer 10, 25, 50 and 200Mbps plans, depending on the area and backhaul capacity.


#11

Notwithstanding real world experience, Mimosa recommend limiting at 40-50 clients.


#12

Happy New Year!
Rory - can you clarify how you are connecting the A5c to the Horns? It sounds like you are connecting two 90 degree horns side-by-side per A5c to get that 180 coverage? What is happening with the beam forming in this configuration? Does this affect the GPS Sync function in anyway? Are you ditching the a5-360 14 all together or are you still finding it applicable in some situations?


#13

Nevermind - I posted a little premature. A little digging reveals the answer! thx


#14

Rory, when you’re switching out to 30° horns, what angular separation are you planning? I have a site where I need only about 150° coverage or less, and I was hoping to cover it with 2x A5C, using 4x 40° horns. However this would create a center-to-center difference (on a same-frequency pair) of about 70°. Would this work without GPS sync? (Yeah, I know it’s coming…)


#15

Haven’t done 30 degree antennas yet, only 60’s.


#16

I don’t know but using two horns side by side but their centre line at the same angle from each other as the working range (so 2 x 90 degrees fit with 90 degrees separation for 180 degrees sector coverage) would hamper the beam forming working when it comes available?

These horns are so good that a CPE connected well in the sector of one horn, would not be ‘seen’ by the other horn. After all, that was the design purpose of these horns to start with!
As I understood from the beam forming theories all four streams of the A5© are needed but if only two streams are to be used beam forming would not work…
(MU-MIMO would still work if that becomes available? Even on two separate sectors?)

In fact, if you are looking for beam forming on one A5c you’d either need a purpose designed 4x4 sector or fit 2 2x2 sectors in the same working direction side by side. (And to have it all working C5’s are needed too…)

Now, in considering going from ‘old’ style 20Mhz channel (or 40Mhz) atheros setup where two 90 degrees AP’s were needed to cover for instance a 180 degree sector these channels needed to be well separated to get nice PYI rates. (We hate co channel noise or tower interference so hence RF develloped their Horns…)

If you now change to a Mimosa A5c and double the channel from 20 to 40 (or 40 to 80)Mhz and fit the same two 90 degree antennas again we still have a better situation.

First we have fully eliminated tower self interference in-between the two sectors.
Secondly the peak load over the combined clients are now better distributed over that one radio that has much more capacity then where we came from…
Third, at short ranges (< 500meter) even with two side by side but separated in direction sector antennas we have 4x2 multi path probability (like in the 360 A5) due reflections and better change both radios (AP and station) get two good working chains. I already found out Mimosa works with two streams even if these are 10dB in difference in strength. Try that on a atheros radio! Most radio’s I know drop the weakest chain completely if it is 3dB weaker then the strongest…)

This compared with the enhancements of the chipset and the software gives already such a dramatic increase in capacity over the ‘old’ (tdma) 2x2 atheros setup you might ask yourself if the extra investment in a full Mimosa (meaning all CPE to be replace by C5’s) network is justifiable.

Its like going from an 8 year old Toyota to a Mustang where we all dream that spending 10 times more in a Ferrari gives us the ultimate in driving… Wouldn’t most of us not opt for the Mustang?


#17

I’ve read through all of this as well as a number of other posts. We, too, are big fans of the RFE horns and use a ton of them.

In regards to the A5c specifically, while you can connect two separate antennas to it (two 60’s to get 120 degrees), it is not really designed to do that, right? Both are on the same frequency since it is the same radio, correct? We have not gotten our hands on a A5c yet, just the A5-14’s. So, if you use one A5c to feed two separate antennas pointing in different directions, you loose MU-MIMO ability but I suppose GPS would continue to work fine.

If you point the two antennas in exactly the same direction, I’m trying to wrap my head around what you gain exactly. MU-MIMO at some point but anything else? Buying two RFE horns and pointing them in the same direction seems like alot of $$$ to only buy yourself some vertical beam width (and a very clean antenna pattern).

While we like the A5-14 for our mini-pops, we are looking at the A5c in some 1 - 2km deployments and are trying to wrap our head around the best antenna to do this with and I’m not sure I’m smart enough to have figured it out…


#18

Feel free to correct me where I’m wrong here - most of this is from memory of past posts.

My understanding of MU-MIMO is that you’re still transmitting to (or receiving from) 2 or more clients at the same time using the same frequency. You still need to have separation of the signals otherwise if they overlap they will interfere with each other. Therefore having an A5c with 2 antennas pointing in the same direction will not give you MU-MIMO unless the antennas implement some form of beam steering. Without that, the best you can hope for is some improved signal due to diversity.

The A5-14 can potentially do MU-MIMO because it has 4 antennas at 90* to each other, so theoretically you can be using the same frequency to talk to 2 different users on different sides of the quamni at the same time. The A5c should be able to do the same if the 2 antennas don’t overlap (so 2 x 60* antennas giving 120*) but then I think you’d need some separation to avoid issues at the edge.

All this said, MU-MIMO is not (currently) supported by the A5 software. I think it’s on their roadmap, but not sure when. So with the current software, and non-overlapping antennas on an A5c, you can have 2 clients (1 on each antenna), but they are serviced sequentially.

If you look at something like the Cambium 450m, they do the MU-MIMO this by steering up to 7 MIMO streams to 7 different clients using a beam steering antenna and a radio that can do 14 streams. In real world deployments they reckon they get around 3x-4x improvement rather than 7x as it’s difficult to get 7 precisely positioned clients in each transmit cycle. So each of these 7 steered beams are equivalent of having 7 non-overlapping antennas in our A5c example.

Rich


#19

Rebooting this thread and it sounds like @Rory might have the most experience with horns. We are looking to replace a mini pop radio with an A5c potentially. The mini pop currently has an omni but we only need about 120 degrees. Our thought is to put up one A5c and two RFE 60 degree horns.

What should I expect to happen at the edge where the two 60s meet (the center of the 120)? There will certainly be a bit of overlap. The AP with have GPS sync on.

Due to noise in our current setup, I’d rather not use an A5-14 here. The horns will be better, if I can go side by side with one radio and two horns. Totally get no MU-MIMO this way but that’s not an option now anyway.


#20

As far as I understood from readings MU-MIMO just makes use of the fact that a radio can ‘per’ chain (or per duo chain) communicate with one CPE and with another chain (or duo) with another. This way increasing the throughput over the AP since more than one CPE can be served at the exact same time.
As far as I understood there is no need for special antennas for this. Its a radio technology.

The same if for beam forming. The radio manipulates per CPE the chains that ‘speak’ with the CPE in such a way that by slightly delaying the signals per chain a ‘peak’ in signal strength increase can be achieved at the location of the CPE. Thus increasing signal and thus S/N.
At the same time in other ranges where no CPE’s are to be found the AP antenna in fact decreases its ‘hearing’ capability and thus eliminating noise coming from that range.

The more antennas (‘chains’) a radio has the better and the more precise a beam to a CPE can be formed.