Low performance of A5/A5c

I have 2 x A5 and 2 x A5c (+N5-360 antenna) all covering an estate that is only 600 y 400 meters.
3 of them are sync’d AP’s with 80 Mhz wide band which seems to work fine (the sync) and thus cannot go back to Wifi mode. One A5 is sort of surrounded by the others and runs on 40Mhz channel. All same gender. All 75/25

Although all my clients are allowed to download up to 200Mbps I never see that kind of traffic. The AP’s hardly ever go beyond 100Mbps and some clients wine about the low speeds. When www.speedtest.net to our own ‘in house’ Ookla server (So full network speed is available) hardly ever see speeds reaching over 50Mbps.

In the same area I have on Mikrotik Netmetal running 46 associated clients, in an 80Mhz partially overlapping (with the Mimosa’s) band and can easy push the Netmetal over 200Mbps and a single client to 150-180 Mbps … much better then the Mimosa network! This Mikrotik Netmetal also works in plain 802.11ac rts/cts mode! (Which many say is almost impossible to do…)

I went for Mimosa 4 years ago to deliver high speeds to my customers to keep the fiber guys out of my pond but now it is really failing me… And if a Mikrotik in plain Wifi mode is outperforming the Mimosa’s that cost 3-4 times as much I am really scratching my neck…
(All Mimosa clients are C5’s with 2 or 3 C5c’s. All clients are within 400 meters of their AP)

Suggestions?


This is the spectrum of the A5 running 40Mhz on his own…
Just to give you an idea of the spectrum and noise.

Let me get this straight, you have multiple A5c radios colocated in the same location and you are using omni-directional antennas and you are having noise issues.

First off, how much spacing do you have between your antennas?

Why are you not using directional antennas for the A5c radios? Honestly I would recommend some 90 degree RF Elements horns. No one can design a system for multiple omni-directional antennas to be blasting each other like that.

No, you misunderstood. Each AP is some 200-300 meters away from the other. Like minicells.
Originally I was planning to start using sectors for each cell if numbers on the omni’s would reach too high. But that would need at least 4 sectors for each of the 4 towers in the end meaning some 16 new radio’s with there antenna. Even with some sychronization that would put an evern further strain on the spectrum. Most of the spectrum is used by other operators and several backhauls leaving this same estate. Not so much to do about. I have at leat 4 other operators ‘fishin’ in this estate with their sectors/omnis aiming to us… (I was the first though 12 years ago…
In that case of 4 x 4 new units with sectors would be another 4-6 years ROI wich I believe is not realistic to work with.
I changed two of the original A5’s into the A5c’s with the N-360 antenna in the hope to increase the capacity and the performance a bit on these two of the most heavy used towers. To be honest, I hardly saw any improvement towards the original A5’s. Pretty dissapointing.

Ok, that is significantly more reasonable.

Could you post some pictures of your Client page on the PHY Stats Tab?

You are dealing with a lot of noise, Omni-Directional antennas force you to listen to any noise in the area. The only difference you would have seen would be a slight increase in signal strength, but that would increase both the customers and your noise floor so ya, you would not have seen an improvement. I would look to setting up APs that are pointed in your most densely populated directions with high isolation horns or sectors.

Another option would be increasing the gain of your customer antennas, probably a more costly endeavor, but cheaper to do on a case by case basis. The goal being to get as many of your customers running MCS 8-9 as possible.

Another option would be to look at some of the new 60 GHz radios coming out, but you are talking about ROI of years for Mimosa equipment that probably isn’t an option for you. The benefit of 60 GHz is that noise is far harder to pickup there and it’s pretty optimal for really short distances like you are talking about.


This is the ‘stand alone’ A5c with some 30 associated clients.


This is one at approx 300 meters distance, with some 29 associated clients

And its spectrum

Third AP (A5)
And its spectrum.
This one is some 200 meters from the first, and 400 meters from the second it can directly ‘see’.

Between these points we have 60Ghz backhauls running. (Mikrotik and Ignitenet) but also 5Ghz for backup.
From these 3 towers leave at least 2 more backhauls to other AP’s in the countryside.
The 2nd and 3rd tower also have another 5ghz Mikrotik AP for temporarily users.

This is our Mikrotik Netmetal at some 150 meters from the 2nd Mimosa and some 350 from the 1st and some 450 from the third. This one has at times over 40 clients associated. Partially overlapping the Mimosa AP’s.


This is our estate.
The AP’s are marked with letters.
The orange lines between them are 60Ghz backhaul links
The green lines are UBNT AF5XHD backhauls leaving to remote locations.
The yellow lines are Mikrotik backhauls leaving to remote locations.
The blue line (left) is a 2, x 40Mhz wide backhaul leaving to remote location.

Apart from this we have 3 other providers with 5Ghz omni/sectors (4 or 5 frequencies) inside the same area trying to get some clients.
Then we have to more towers at some 1-2 km distance pointing their sectors towards this same estate from yet again other companies.
Then we have at some 5-10 km at least 6 more towers that have sectors that also cover this area.

This is a wifi scan from the Mikrotik (“E” in the map) which shows how many SSID’s we have to deal with…

From “B” we have 4 x 60Ghz backhauls running, plus 3 Mikrotik 60Ghz sectors that serve another 10 clients with high speed internet.

As you can see, in this estate we serve some 150-180 clients in total. It is a 1000 houses estate so obvious others want to fish in the same pond…
Since 3 years we also have a fiber company offering connections.
We started with Mimosa 4-5 years ago to be ahead of that in offering more then the at the time standard 15-30Mbps packages. So with Mimosa we started upgrading clients to 50 and 100Mbps and when the fiber was finally there (at slightly better pricing than us, so we had also to lower our standard monthly fees…) we made that 100 and 150Mbps (or 200mbps).

In all those years (this estate is our ‘core’, our birth place…) we’d grew little by little but now we are loosing slowly clients towards the fiber.
We live in this estate as well and we know most clients personally. It is a expat community but now with Brexit approaching more and more British leave to be replaced with locals (Spanish) that have no bonds with us and go straight for fiber.
Existing clients are pretty loyal to us since we speak their language and they know us but the ‘speed freaks’ are going to fiber, and we had to lower our pricing to not offend our clients in beeing more expensive then the fiber guys…

If now it seems Mimosa cannot cope with all the client-noise situation as expected (Where the Mikrotik over the years improved and can deal with it) and due Covid and the arriving of more young families with their kids and thus higher internet usage I can only foresee we are fighting a lost battle.

We run in total another 600 clients through 30 sectors in the countryside away from this and since here we don’t have to compete with fiber, only with other wireless, we focus on putting our money there in offering 100+Mbps for those at 1-8 km’s distance from our towers…

We can only flip the coin only once and since budget is getting more tight every month there is little justification to invest on more Mimosa. (Even Mikrotik 60Ghz line are cheaper, with the only limitation of only 8 clients per AP)

So, you tell me, how does the situation look to you now?

You are in a tough situation.

Wireless isn’t magic, and at the density that you are at Fiber starts to make more sense.

First off, if you have 30 customers on an AP that are on 50-200 Mbps plans I would be hesitant to say that you are actually going to be able to deliver that. The best performance I have seen out of an A5 was above 200 Mbps on a 40 MHz wide channel. (real world, plenty of people with test sites that have pushed the full throughput) So maybe you can do 400-500 Mbps through an AP in your situation, maybe more probably less. I don’t sell Mimosa here, I just help out. I also don’t have a lot of high density deployment experience, so maybe there is something I am missing. (My average customer is +1.5 km from my towers) So I won’t ever say that you should be seeing 1 Gbps through an A5 radio outside of a lab, maybe someone has, but not me. Now, does that mean that Mimosa is crap? No, it just means you need to figure out what your APs are actually capable of and build your network around that.

I would honestly leave your APs where they are and begin moving your high bandwidth customers to either smaller sectors sized APs (RF Elements Horns for example) or 60 GHz. I don’t have a lot of experience with Mikrotik Wireless, though I do know that they have been improving over the last 2 years, honestly you probably made a pretty good decision 5 years ago when you went Mimosa because Mikrotik wireless was in such a bad place. There is also the new Terragraph gear, which while pricey might fit your needs.

From there, you have to have something that gives you an edge over the competition. Where we competed against fiber, we had to really fight. We made sure we had amazing tech support, we built redundancy into our network so we had fewer/shorter down times compared to the fiber company, we were cheaper and we were super reliable with our speeds. (Note: we only oversubscribe our bandwidth 4x which means we know how much our CPE, APs, Routers, backhauls and our upstream providers can do and we will only sell upto 4 times the bandwidth of the weakest link in that chain in an area, most frequently we keep that to even less).

I also implemented a policy of revisiting customer’s issues, when people had issues, we fixed them even if it took us weeks of diagnostics. We then revisited the customer a week or so later and make sure they are happy with the service and everything was fixed. We made it so our customers thought the annoying internet issues they experienced annoyed us more.

Doing all that we were holding our own and frequently growing against the fiber network. It helped a lot that they didn’t have the technical expertise to maintain their network and they had outrageous prices and a boatload of fees. So, it’s doable to compete against fiber, just hard.

I dunno what your prices are or your costs for connectivity, but you might look at seeing if there are better deals you can get. We just dumped one of our upstream providers for another who is selling us 10x the bandwidth for the same cost of our old one.

In short, cut down the number of customers per AP, you don’t need to have 4 radios at each location, build smart and optimize your costs. Have great customer service, don’t sell what you can’t provide and be willing to say “no, I can’t guarantee that I will consistently provide that” people respect you when you stick to your guns and when they run into issues with another provider they frequently will come back to you because of better “service”. Have better uptimes and help out people with technical issues, know your network like the back of your hand and learn everything there is to know about it. All of these have a pretty high cost, but so does fiber.

People are buying an experience of their internet connectivity, most don’t care what the numbers mean as long as they can watch their YouTube videos, Facebook works properly and when they have an issue there is someone who fixes it.

In the regard of the plans we sell, it’s mainly showcasing. Our clients are 70% pensioneers and only 30% families or younger people. Since there is little room to position ourselves with better prices then the competition the only way to attract new customers is by offering the highest speeds.
How do people test this? Well, running www.speedtest.net to our own Ookla server which basically always give good results. (We also use it for troubleshooting connections)
As long as the client indeed can see he/she receives close to or more then 100mbps they are happy and we are.
That they then in 99% of the time never use their connection with more then 20-40Mbps is nice for the network and thus for us and the client still has the ‘feeling’ he has a super fast internet.
On my desk I continiously see the top heavy users in the PPPoE interface list in real time from my gateway router and I rarely see people having connections running above 40-50Mbps…

So the high packages is merely a selling tool then we really need to supply this to many customers at the same time. Most probably this setup would not work in a highschool/University campus town…

My issue with Mimosa is more that when we installed it they had a high promise into the future and indeed 4 years ago the where outperforming a lot our Mikrotik based networks…
But Mikrotik improved a lot since where Mimosa has only been fulfilling slowly the promises of new technologies that they said 4 years ago were ‘around the corner’ at the expense (to us) of high investment costs and very high failure rates of their hardware.

Our rates are basically 25€ (that is 28USD I think) net (without the VAT) per customer per month which is marginal but by far not the cheapest you can find. Fiber cost the same and we have wireless operator offering connection for as little as 10€ a month… how they make money I don’t know…
Every 2-3 year we are hard negotiation with several uplink bandwidth providers and I think we have very good deals. Momentarily pay 1700€ for a Gig connection per month. 3 year contract though and still some 2 to go… Most providers charge 2000€+ plus for Gb connections. We were only lucky to have Vodafone (cel operator) trying to setup a new business in supplying their surplus bandwidt to business…

We recently finished the last changes of rural customers and AP’s from 802.11’n’ into ‘ac’ networks and upgraded some of our backhauls so 80% of our clients are able to reach 100mbps of speeds on their connections.
Next we are planning is to do more upgrades on some of the backhauls and replacing several sectors or 1st generation RF-elements horns for the better 2nd generation A-symmetric horns.
And I am looking into possible migration towards another platform for the Rural. Because the 100mbps we offer is now sufficient for most clients and our AP’s but the average usage is slowly growing and the Mikrotik AP’s are the limiting factor now. But where to migrate too? Every migration means a heavy investment in both AP’s and client devices with ROI’s of 2-3 years in a market that is under pressure from more fiber deployments, G5 roll-out and maybe sattelite (The new Elon Musk’s networks?).
So here we are, I am not willing to invest thousands of euros again into a product line were it is highly unsecure if I will ever earn that back. At the same time, quitting is not an option neither since we have 3 families depending on our business (including myself!) and without it my employees get only a very low support (Spain doesn’t have luxery unemployment benefits) and I am without any income at all!

One good tip from you I am going to use is to setup an even better ‘post repair’ client contact register as you’d mentioned. It will definately help to give our customers a good feeling we are there for them…

In our estate I am planning to setup more 60Ghz AP’s to off load the Mimosa’s (and the Mikrotik) but it goes slowly as we have little funds available for buying these…

One thing we are also doing now is offering for a reduce fee customers on of the latest in home Wifi. Since all our CPE’s are now having Gb ports and 75% of the customer issues are wifi related and we see an increase in IoT usage with many devices the standard basic 10€ Wifi routers we used to supply are in many cases the limiting factor. Tenda has now a series of very good, very high capacity (2,4 + 5Ghz, bandsteering, Mu-Mimo, beamforming 6-7 3dBi antennas, Gb ports, 1000mhz CPU) wifi routers that already saw many of our clients to cheer at!
I believe in most cases the bottle neck of our clients internet lies in the cheap wifi so here is something to gain against little costs.

It’s only 7 more years to my pension so we see how far we can manage to get this boat going! :slight_smile: