Need advice on setting up WISP


Been looking at starting a WISP for my neighbors on and off for the past 2-3 years, but have recently gotten a little bit more serious about it since many of them expressed interest in it. They are all sitting on ADSL with Fiber not even planned in the nearest 2-3 years.
We are talking maybe 20 people on the low end, and 100 on the high end.

I’ve got Fiber and talked to my ISP, cleared a 5-10gbit/s connection depending on my needs.
I’m thinking of offering 100/25mbit/s and 50/20mbit/s packages, i see many WISP’s offering 100/100mbit/s, so my package shouldn’t be a problem? This is where i’m getting lost, what hardware i need etc. Could need a gentle nudge in the right direction.

I live on a small hill, with direct LOS to most of the potential customers, some places there are a couple branches in-between. Each of them are situated about 200-600m tops from the tower.
As it stands, my house is a very tall one, i could probably get the radios 13-15m above ground, which would put them at a ~15m advantage over the CPE’s, since the house is located on a small hill.

Any help, advice or direction is appreciated. Been trying to research this for some time now, but haven’t been able to arrive at a good conclusion yet.

First off, you should make sure your internet provider allows you to “resell” your internet service. Most companies have a clause in the contract you sign that prevents this and is grounds for termination. People have run into this before with terrible consequences.

Beyond that, there are a number of considerations, which I will have to get back to you on later today.

Thanks for the heads-up! This was actually the first thing i did when i started considering this. As i only had a quick chat with them, they estimated i would pay between 150-450$ per month for a 10/10gbit/s link depending on other unspecified needs. But 450$ is the absolute max according to the ISP, this would include reselling rights!

Awesome! That’s great, and an even better your price is outrageously cheap. Glad you already have that handled, it’s one of the bigger pains in the neck to figure out and learn about.

I was paying 3x your maximum for 1 Gbps, now I am a little bit higher for 2 Gbps. Luckily I have another 1 Gbps upstream that I don’t have to pay anything for.

I will get back to you about the rest of your questions, there’s a lot involved and I don’t write succinctly very well.

Oh wow… I guess sometimes it’s good to live in Sweden :slight_smile: We have a standardized “contract” for these types of leases to avoid monopolizing the ISP market.

Looking forward to your next reply! :slight_smile:

I strongly recommend you to watch How to Start A Wisp video series here:

1 Like

The Mimosa Starting a wisp series is actually pretty good info.

I use a lot of technical jargon below, largely because I’m lazy. If you don’t know what I’m saying, just ask for clarification and I’d be happy to explain.

To be fair, there’s about 50 miles of buried fiber justifying that $1,500 a month. I’m located on the road to nowhere Colorado, so internet costs a bit more then usual. Also, all my amounts are in US dollars, dunno what the conversion rate is or how costs may vary over there.

Sounds like you are in a pretty prime location, is there an HOA that might cause you trouble, or a wife who won’t like the radios on your roof? Otherwise, I like what you are saying as far as height above the surrounding area.

Mimosa has a really good link planner tool for free, just use your account and go to Mimosa’s Network design Tool and play with it. You can get some really good ideas about how different things like channel width, antenna gain, noise and TX power effect what you can do. I would recommend spending a good bit of time there just learning how the various radios and antennas perform differently.

So, first off, the question is going to be what kind of profitability are you aiming for? It’s great and all to be doing a community service, I work for a co-op, it’s fun, but it doesn’t pay. Also, what’s your end goal? If fiber could be 4 years out then you are running against a clock (sure it could be 20 years out, but you don’t know so it should at least be considered).

Next up you need to figure out what kind of cash your wanting to invest. If you are just using your house you can reduce your costs significantly. As long as you and your wife are willing to live with the eyesore of radios and antennas sticking up off of your roof. If not then you will have to look at poles/towers… Figure out how much you can charge for what kinds of service so you can have some clarity in how long it’s going to take you to make your money back.

Alright, now you should have some ideas for how much you can spend, which is a great guidepost for building something IMO. My back of the head guess is you’ll need between $2,000 to $4,000 to get started depending on your preferred system and tools you need to buy. (1 router, 1 Access point + antenna, 5 customer radios, good cable and accessories) Now, this could easily get to the $15,000 range in total investment, but that’s not what you need to start. (With the full 100 customers you are looking around $25,000) not that there are many other options out there, but I am kinda directing my recommendations to this budget.

Starting off, you will need the basics as far as cabling stuff goes, RJ-45 crimper, power drill and other stuff for running wires. It sucks, but you are going to spend a lot of time doing installs if you want them to look decent. I would recommend talking to an electrician so you can make sure you are following your countries building codes. Get good quality outdoor rated cable, I recommend Shireen, get the stuff that is Plenum, burial rated and UV proof, it’s more costly then normal Cat6, but it’s worth it to avoid as many cabling issues as possible. (I use Cat5e, but I’m cheap and don’t follow Mimosa’s recommendations, lol)

Wireless equipment is kinda the big question though:
Since your posting in their forum, I’m assuming that your looking at using Mimosa equipment. The A5x is probably right up your alley, although the A5 is a close second in my book. I have not run the A5x yet, but it’s small, cheap and simple all of which I imagine are positives in your book. If you go with the A5x then you will need an antenna to go with it, an omnidirectional antenna isn’t a bad way to start, but you will run into issues with noise eventually. If your customers will be in a group then you are in luck, check out RF Elements and their horns, specifically their 90 twist port I’m using it with great results. If you want to build with an eye towards growing towards your 100 customer limit then you might consider an A5c with a N5-360 antenna. The A5c is definitely the expensive route, but as you grow you can move the A5c to a smaller beamwidth antenna. Client radios, I would only recommend the C5x. Grab an assortment of the types of dishes, but you should be fine with mostly the low gain antennas. Don’t mess with the C5 or C5c, C5x is the way to go.

As much as I like Mimosa, I feel I would be dishonest if I didn’t mention some other options…
Ubiquiti, you can basically buy all your equipment and manage it together in a unified system. From the client router to your router they have something that will fit your bill. I use them, they work well. But I am not a fan of some of the companies actions in the last few years. If I could I would move away from them. (Airmax AC line or their new LTU line, can’t make a recommendation on their router line)

Mikrotik (60 GHz), Mikrotik isn’t known for their wireless for good reason. But they have a great 60 GHz product that I’ve been using in small deployments like yours to great success. If I was doing this system I would probably go this route… Sorry Mimosa. The main benefit would be that you would entirely avoid interference which is a problem in urban deployments. There are negatives, for most of your potential customers it’s not an issue, but 60 GHz experiences serious rain fade issues, beyond 500M you will have to have good alignments and test in your are to make sure you won’t run into issues. Slight issue, max 8 clients per AP, but they are pretty low cost and don’t take up much room. (Just look up Mikrotik 60 GHz and you will find their whole range)

Cambium is kinda the high class radio manufacturer in the industry. Most people who run them remind me of Apple snobs, and they still have issues. I don’t run them, besides for customer routers, but they are an option. They do have a management system that is pretty slick though. They are expensive. (I use their CnPilot home WiFi routers, otherwise I can’t make a great recommendation, talk to a rep?)

If you don’t go the Ubiquiti/Cambium route, router hardware is the next question. You will need a router, for diagnosis, general network setup and isolation, all sorts of reasons all lead to “get a good networking router, don’t use a consumer router”. I like Mikrotik they make good routers that don’t cost a lot and have lots of features. If your familiar with Cisco (or any of the big names) you can find used hardware online that will work, but I can’t give much guidance there. My preference is to buy cheap new over used old. Ubiquiti is another company with routing hardware, they might be the simplest to setup, but have limitations that shouldn’t be a problem for you if you stay under 300 odd ppl. If you are still wondering about why your consumer router won’t work, just trust me and start learning networking.

Next up is switching hardware, this is debatable though. If you spend a bit more money on a router you should have enough ports available you won’t need a switch which both saves you money and simplifies your network in some ways. Getting a PoE switch will help a lot when you have multiple Access Points and don’t want a million wires going everywhere.

Finally, antenna mounts, if you go Mimosa get their antenna mounts. Otherwise you should be fine with whatever J-pole you want to use.

Gonna check out that video series for sure!

I actually went out today to scout the neighborhood properly, and found a couple issues. I started up on my roof, and realized that i was going to run into quite a few LOS issues with trees and other stuff. I’ll attach an image and show what i mean to give you some context:

The blue area is my would be customer-base. All 1 story buildings.
The red circled area is my basepoint, and where my stuff would be located. The red stripes are LOS issues with big trees, other houses and hills.
The white circle at the top, is a ~30m or so high antennamast i found, and the white circled area is the places from which you can see it, ROUGHLY. Though i don’t know if i could get a good fiber uplink there.

I read all your comments about gear, and still feel determined to go with Mimosa. Feels like a solid choice and judging by everything i’ve read on other forums etc, i think it’s the best choice.

My main concern in all of this, is whether or not i would be able to deliver my determined speeds. It feels like a lot of people just sigh at me when i say 100/25mbit/s and 50/20mbit/s packages, and that’s where i get lost. On one hand, we have people saying they achieved it and are offering these speeds to customers, and on the other we have a whole bunch of naysayers…
How many A5c’s and A5’s would i need to deliver rocksolid 100/25mbit/s packages to, say 50 clients?

Thanks for all your time, really appreciate it! This is a lot of fun tbh, learning tons of new stuff.

It all is really cool stuff and there are a lot of moving parts.

I just recently went through a bunch of stuff about delivering speeds: Client Number and Speeds, it’s not exactly what you are looking for, but it covers a lot of the issue of over-subscription which is the biggest deciding factor in how much you will be spending and how reliable your service feels to people.

So, trees… Ya, they are the bane of all wireless services. If you are wanting to deliver the speeds you have mentioned, you will need very clear Line of Sights and short shots. That everything is under 1 Km will help you a lot.

As far as the tower goes, is there LOS between it and your location? If so you could use a backhaul radio (a PTP like the B5 or B24) to get internet to the tower and then send it out through PTMP Access points on the tower. You would need power at the tower, but that is probably easier done then getting fiber ran.

As I said earlier, there are a lot of moving parts in wireless internet. Each one will effect the end performance of the equipment. If you properly set things up, yes you can get really great results. If you don’t or if issues come up that are not taken care of, then performance will be degraded.

Not all of the issues are going to be wireless related either, a slow router or an improperly setup switch can me just as bad as other issues.

I think i have a good idea of how this will have to go, thanks for the link about speeds btw!
The tower has no LOS to my house, but i’m going to reach out to the mast owner to see what kind of possibilities there are over there.

What confuses me now is the prices of Mimosa gear, many resellers list prices very high, and other ones, like for example, are on the low end.
Since Mimosa don’t seem to have any official prices i don’t really know what to look for etc.

I’m thinking of trying a B5lite backhaul, over to an A5 into C5x CPE’s. Will try to start at 5-10 customers and then see if this concept works.

You’re not going to get 100/100 or even 50 upstream in a typical situation. Individual users may burst at 100 down but total capacity is limited. Most PtMP WISPs set the upstream to be smaller than the downstream because that’s how most demand is.

I wouldn’t touch the B5-Lite. It’s plastic and my trial set failed quickly in the field. The C5x does point to point though and it is the only radio in its class to have a metal case. It is very solid and has a choice of screw-on antennas. Of course it is really designed as a client and is good for that too. The B5© is a good backhaul if you need that two-frequency capacity. I’ve got some running over 10 km very nicely.

Trees are of course a big deal on 5 GHz and even more so on higher frequencies, if they are available in your country (I think 60 GHz is). On a short path a few trees won’t block you, but how much impact they have depends on the trees. In the US we have a narrow 902 MHz band which penetrates trees nicely, and there is also TV White Space, though that might be expensive and I don’t know if your country (.se) has it.

Yeah, 100/100 is a no-go for me, at least until i get a good hang of this. I know a company here in Sweden which specializes in delivering 100/100 via wireless using a combination of Mimosa gear and some crazy-ass backhauls.
I’m aiming for 100/10 or 100/25, as most of my customers ain’t that needy on the upside.

I’m going to take your advice on the B5-lite and stay far away from it. Read a lot of negative reviews…
You’re saying either go B5 or try it with C5x? Going to need a couple C5x’s anyways, so might try that as an inexpensive first test.
Any advantages going with the B5c over the B5? What would be a good antenna for a 220m backhaul shot with B5c?
For the C5x backhaul shot i assume the 25dBi dish is the best choice.

Lots of questions today… :slight_smile:

100 downstream is still a stretch on a PtMP system. It’s possible, especially if you can do 40 MHz channels, but it takes a very good path. Do some experiments before selling it. It’s better to surprise people with how much better it is than how much worse it is.

The B5c is the same radio as the B5 but with antenna connectors instead of a built-in 25 dB dish. So you use the B5c when 25 dB isn’t enough. For 220m, the plain old B5 is plenty. The C5x is a 2x (one frequency MIMO) radio; the B5 is a 4x (two frequency MIMO) radio thus with twice the capacity. Also, if you need DFS radar detection, and radar hits, a B5 will shut off that channel but still have one running so the link stays up. But the MCS of both frequencies is the MCS of the weaker of the two links.

When buying Mimosa products I would stick with their distributor page,

At 100 Mbps downstream you will need to be careful with your over-subscription especially with so few clients. In larger networks you can balance things out more easily, but you will want to make sure you have VERY good wireless connections. Nothing will kill your AP’s total throughput more then a couple poor connections. (by very good wireless connections I mean better then 32 dB SNR) And you will need to use 40 MHz wide channels to even have enough bandwidth to work with. That said, you have very short paths to shoot so your biggest concern is going to be trees.

It may be that you will have to tell people that you can’t deliver good service to them until you can get different points for your APs. It’s hard to pass up the cash flow, but it’s better to have a good name and fewer happy customers then a bad name and lot’s of unhappy customers.

Been reading and planning like crazy, decided to go with the B5 for backhaul.
Going to start with a 200m shot from my house to my A5 AP location, from there i will onboard a couple customers, each within 100m from the AP.
Going for C5x’s with the x25 dish across the board, just to be on the safe side, might be overkill, but the price differences are that huge, so might aswell grab the big one.

What i’ve stumbled across a couple times though, is something called “feature key” and i need to “unlock” the higher speeds… what is this about?

The “Feature Key” for the C5x is a relic of the past, an attempt by Mimosa to try out some different models for pricing. It is no longer pertinent or necessary, just use the latest firmware and you will get all the goodies Mimosa baked for the C5x.