I'm looking to replace some rather expensive Motorola PTP equipment with Mikrotik-based solutions at all the tower sites at my company.
One of the first links I plan to replace has a distance of 15.7 miles, with a clean LOS:


This community seems to be well seasoned in this field and I wanted to get some feedback on a proper setup
before we buy the equipment and start testing.

Ideally, we're hoping for at least 100Mbit/s throughput, but can live with 80.

This is the list of hardware I have come up with so far:

1x RouterBOARD RB800

1x Ubiquiti SR71-15 MiniPCI
or
1x MikrotTik R52Hn MiniPCI

1 maybe 2x Laird PA58-24 24dBi 5.15-5.825 GHz

I chose the RB800 because of it's ~280k PPS/1.9Gbps capability.
This will be plenty for our highest-traffic link, yet allow for future increases in traffic to/from.

Would multiple antennas in MiMo be a necessity for higher throughput?
Does anyone have an opinion of Laird antennas?
Any other tips or things I should look in to?

Any guidance in the right direction is greatly appreciated.

Thanks,
Bryan

RB800 is a good choice for max performance. I learned the hard way about how routerboards can bottleneck your system. Thankfully it was long before i ever needed to put one on a tower.

First, avoid Mikrotik wireless card. The SR71-15 is the premium 802.11n card in my opinion. I'm just one voice but I've had too many problems with Mikrotik's N series cards (and A series for that matter.) I have not used the R52Nh (or whatever) but I have used the single chain version and it has not been pretty. 3 out of the first 6 I ordered were bad. One thing to know about SR71-15 is that RouterOS does not accurately report their default power levels so you will need to research this issue so that you can get them working at the levels that Ubiquiti's spec sheet claims they can operate.

Second, stick with Mikrotik and nstreme for your board solution as it seems you are doing. Ubiquiti makes slick products and I love their form factors and price point but Airmax is inferior in pretty much every way to Nstreme for long-distance 802.11n links. Airmax (and nstreme) are always in development so this could change but as of today I have basically eliminiated my use of Airmax for PtP and do not intend to revisit the issue any time soon.

Also, it sounds like you are replacing Motorola PTP gear with Mikrotik stuff. Realize that no matter how well your mikrotik stuff works it's not in the same class as Motorola stuff. You will get more bandwidth but it will also be a bit buggier. So depending on your application you may want to stability of a motorola solution over the performance of a Mikrotik solution. 802.11n is new tech and after having done about 7 or 8 long PtP links using the technology I can safely say that it is a work in progress and it's hard to predict how your links will work before you install them. In no way can you compare a Motorola to a Mikrotik so be mindful of this as you are contemplating this change. I have very high regard for Mikrotik and obviously choose to use them (except for their wifi cards which are mediocre) for my links over Motorola but I also know that if my job depended on these links working at 99.999% levels I would not be using Mikrotik for the solution.

Yes, MIMO will double your throughput. I have not yet implemented a MIMO solution at this time but it does help bandwidth. So you know on a single chain 802.11n link of 26mi at MCS7 (HT8-40) I can consistently squeeze 60 megs of aggregate traffic out of a link split evenly between up and down. I can probably push it up a bit more but the bandwidth test start bouncing around a bit and the ping times really spike which indicates to me that I am maxing out my link. even on shorter (18mi) and longer (36mi) links I can really only get this performance so I'm thinking that is about the limit at this time for me. Since the link itself syncs at over 100000kbps I am hopeful that further nstreme development will increase my real performance. I have seen reports that MIMO doesn't perfectly double your throughput so you may not reach 100 megs and you will need to have a link in the mid to low 60s to maintain the fade margins necessary to run at MCS7 even through the sensitivity at that rate is -74. Given your link distance and stated taste for high-gain antennas this should not be an insurmountable obstacle for you.

Yes, Pac Wireless makes nice antennas and you could do much worse (or better if you want to go with expensive Andrew units.) But I suggest you look into the Ubiquiti RocketDish as they are well regarded dual-pol solution.

colebert, thank you for your very excellent and thorough reply.
I really appreciate the wealth of insight and tips.

I'm glad you should mention Andrew antennas as I had my eye on the FPA5150-23PM-1 panel,
the RocketDish looks great though, and it's cheaper.

We use dual polarity waveguide fed Andrew dishes but I'm not convinced that the price difference buys you much improvement in overall signal strength vs a Rocket for example.

I suspect what it gets you is greatly improved build quality and durability.

you can try the antennas from Cyberbajt

My company has since been using, with excellent success, the RocketDish antennas.
Thank you for your input though

we have tried about 30 link with the Lanbowan (http://www.lanbowan.com) antenna, and it's work great...

you can try the antennas from Cyberbajt

in fact, i have interest on their parabolic antenna, but just wonder know how is the performance...
does it have the Asean distributor?