RocketM and intentionally spliting antenna coverage.

[JustJoe]

OK, so we confirmed that a RocketM with a dual pol antenna and 2x2 capability is supposed to coexist perfectly with BulletM which is limited to 1x1.

Cool. :)

As I see it, the way the 2x2 products are being packaged, they are going for maximum data transfer, something really important in a PtP link.

But in a rural area, I cannot yet get Internet back-haul for WISP customers that will take advantage of that. So I would rather optimize PtMP for 2 other situations. Using only 5GHz. Since I am covering average distances to customers of 5-6 miles, Nanos don't reach because of the low gain of the antennas. So I would want to use BulletMs as stations on the back of 22dBi panels. I would wonder if you guys can share your thoughts:

1) Absolute minimum power requirements at a remote hilltop POP needing 2 separate sectors for 2 separate population centers. Question is could I use a single RocketM and two independent 60 degree sectors of the same polarity (vertical, since they are cheaper) attached to it??? The intention is that they do NOT overlap at all in their coverage area.

I realize this was an absolute no-no with .11a/b/g and trying to use antenna diversity!! BUUT what does RocketM with 2x2 mimo think of it???

2) OK, again on a remote hillside, but this time overlooking a single iindustrial park with an assortment of metal and stucco buildings. What if I use a single RocketM, this time attached to two independent 90 degree vert sectors with a couple of feet of separation, but aiming at the same coverage area. (Maybe I should even try tipping sectors slightly out of vertical, one to the left and one to the right.)

For this case, I am trying to take advantage of the supposed multipath capabilities of 2x2 mimo of the RocketM, but still limited by 1x1 of the Bullets. Is the protocol and the firmware robust enough to take advantage of the best path or reflection, or do they force themselves to all use the same one sector. Or worse yet, do you degrade overall performance?

[UBNT-Mike.Ford]

Hello Joe,

1) Pointing two antennas in two different directions will not work with Mimo and will have horrible performance.

2) Outdoor Mimo does use reflections, problem is most of the reflections go into open space, not returning to a reciever, and can cause degraded performance.

Thanks,

Mike

[JustJoe]

Thanks for the quick reply Mike.

I must admit that I am confused by #1 not working though. :S

I was saying that I would set up a RocketM as an AP with 2 separate sectors pointing in opposite directions.

The stations off this AP would all be BulletMs, each with a panel polarized in one direction.

But in a perfect world, radios with horizontally polarized antennas don't see radios with vertically polarized antennas. That is why we use both and take advantage of that phenomenon.

So how is my proposed setup different from a dual pol sector off of a RocketM pointing at a vert pol panel/bulletM combo (which basically cannot see the horizontal part of the signal coming from the dual pol RocketM) ?

Is it that the 2 radio outputs of the 2x2 RocketM are not interchangeable? In other words, maybe one always acts as a master and one acts as a slave to that one when 2x2 session is negotiated?

If that is the case, do we need to watch which connector is going to which between the RocketM and the dual pol sector? And then it would mean we cannot go in the field and randomly pick horizontal or vertical orientation when installing the BulletM panels? We would have to pick the one that matches the way we set up the AP and stick with it? :P

[UBNT-Mike.Ford]

Hey Joe,

When I say it wont work, let me rephrase. It can in theory work, however the performance of link would be significantly degraded. So much so that I will say it will not work.

However, 2x2 mimo is sending the same exact data stream out of both antenna ports, but still runs into some of the same problems as Antenna Diversity and pointing the antennas in different direction.

Thanks,

Mike