Long Range Wi-Fi – Pushing 2.4 GHz Wi-Fi to the limits (2019) | Hacker News Reader

Long Range Wi-Fi – Pushing 2.4 GHz Wi-Fi to the limits (2019)

phidgets.com

93 pointsrzk66 comments

"Ideal" isn't doing as much heavy lifting as I thought it would be in this article. This is a basic wardriving setup on flat ground.

Standard wifi is phy limited to ~1600m unless you can increase ack response time limit, otherwise retransmissions will be sent by the sender while ack's are being transmitted by the receiver.

Then, per friis equation, with enough power and antenna gain (it's the antenna gain that really matters) the limit is only obstructions/terrain. Get yourself high enough (mountains, planes, satellites, moons, etc), or bounce of an opportune reflector (planes, moon, aurora, meteors, ionosphere, rain, hail, etc), and there is no real limit, besides how big of an antenna you're willing to build in orbit :P

or just make the throughput really really really REALLY slow and ice your receivers with liquid He.

Hams make a game of it: https://www.arrl.org/files/file/WA50-Standings/Distance-Reco...

For discussion purposes just "WiFi" isn't doing much. I doubt typical AP<->phone 802.11ac on 5GHz can achieve >1km let alone 802.11ax/be on 6Ghz carrier, even 802.11n can only reach a few hundred meters with line-of-sight, sunny day, and no interference

Long range WiFi is a "solved" problem with 802.11ah with sub-Ghz carrier and can reach more than a km depending on bandwidth requirements

All that assuming tx power is within regulation, no enterprise grade antenna, and all with commercially available parts

I've been running a 1.1km 5Ghz 802.11ac wifi link between two highly directional antennas for three years now. They have line-of-sight, albeit just barely. The kit I got cost $100 at the time. It was quite easy to setup even and supposedly can go much further than I'm using them for.

https://www.amazon.com/dp/B083JRFQ24

With so much benefit we’re getting from wireless connectivity, one hopes that more frequency bands would be opened up, especially in the sub-1GHz range, to enable long distance connections.

And it should be open to public use. Not yet another carrier frequency they will charge us an arm and a leg to use, some of the cell phone connectivity charges are insane, hundreds of dollars for low bandwidth iot sims.

It is immensely frustrating that there is so little unlicensed spectrum available vs the massive amounts of spectrum available to militaries and very rich network operators.

Worse is that the unlicenced spectrum that is available is often so heavily encumbered with restrictions in many parts of the world. Wi-Fi HaLow for example, in ETSI regions, is limited to a paltry 14dBm EIRP in no wider than 2MHz channels and an unacceptably punishing duty cycle restriction in most cases that render it barely able to make a TCP connection (some mitigations exist that involve channel hopping on some products but nothing that works for things that require fixed channels, e.g. IBSS or 802.11s).

> Worse is that the unlicenced spectrum that is available is often so heavily encumbered with restrictions in many parts of the world

It's one of the great things about the UK. As long as you're not causing interference to anyone that's actually paying for a licence, you can pretty much do what you like.

Is it a practical thing as in they won't catch you, or does the law actually say that?

It's not possible to investigate it, and there's no-one to report it to.

Ofcom have about two dozen engineers for the whole of the UK, none of whom are equipped to deal with it.

In general, though, in the UK if you're not making a nuisance of yourself the laws are pretty optional, and this is a pretty good example of it.

The licensed spectrum is not empty. It is scarce because it is fully allocated! Chances are that most of the allocated users are not using it in a given geographical area, so there's space to argue about allocation policies, however.

You can also consider trying to get a license. Costs and probability of success vary widely, but you may as well find out what they are before complaining.

Some European hacker events get temporary licenses to run their walkie talkies and even their own 4/5G cellphone networks. I assume they are cheaper and easier to get than permanent licenses.

In the UK it costs 75 quid for 5 years to licence a pair of 25kHz channels in 68-88MHz, how "sub-1GHz" would you like to go?

I suspect with modern modulation schemes you'd be able to squeeze quite a chunk of bandwidth out of that.

Ubiquiti do 5.8GHz bridges that work well over long distances.

I used a pair to link from my house to a nearby block of flats, around 9km LOS, and back to my friend's house (we didn't have LOS to each other) around 15km.

They worked well at full link speed, around 80Mbps. These were just bog standard off-the-shelf ones, although for space considerations I "mix'n'matched" the links using one with a 10dB dish at one end and the patch antenna at the other.

Here in the UK at least, you can use as high an EIRP as you like, so there's nothing to stop you taking a cheapy USB wifi or bluetooth dongle and 3D printing an adaptor to mount it on the boom of a cheap satellite dish.

They address that they were looking at the maximum unlicensed power.

In HAM circles some people do use WiFi in an arguably legal manner by making the SSID their callsign and either making the network open or putting the password somewhere public. This is because you are not allowed to use public bands for encrypted communications (with a carveout for satellite control, for obvious reasons).

I'm unsure what ranges can be hit with that setup, but I know it can be far enough that LoS due to curvature of the earth becomes a factor.

I do think just getting a wireless P2P bridge is easier and smarter and why this hasn't seen more use.

You can buy consumer wireless repeaters basically but the latency sucks

> In HAM circles some people do use WiFi in an arguably legal manner by making the SSID their callsign and either making the network open or putting the password somewhere public.

And presumably also no SSL when browsing?

Part of why I never really understood the appeal to getting a ham license. Things like no encrypted communications really limits the utility.

Anything outside of talking to randoms on the open air, and you'll better served by the computer networking infrastructure where encryption is allowed, you can speak to anyone in the world anyway, and the medium (IE physical device) you're using is irrelevant.

Thats the arguably legal part. Using a VPN or SSL over it violates the law, so...

I don't see the point personally. I'm just aware of the arguments. The "best" use case I saw was some hunting and lake cabins extending coverage.

There are several legal digital formats (e.g. PSK31, FT8, RTTY) that, while not encrypted, will render 'reading the mail' hard for anyone not aware of those formats and/or the right equipment and where to get it.

Most hams can enjoy the hobby without encryption. The limits discourage uses that endanger the hobby.

I always find the anti-encryption arguments frustrating because all the digital modes are going to have data whitening anyway, and are therefore (if you are not aware of the exact protocol) indistinguishable from encrypted transmissions for all practical purposes. We already have a number of common systems that are basically encrypted because the codecs are proprietary.

The actual mechanism preventing abuse is the requirement to broadcast a callsign in a way that casual observers can interpret it.

The difference between whitened data and encrypted data is that you can decode the whitened data if you have the correct receiver. Encryption uses a key so that you also need the correct key.

You may not have a receiver for say FT8 that I may be transmitting, but you're also transmitting in Morse code which I never learned. Nothing is stopping either of us from receiving the other's transmission if we want to.

I'm told there was been an explosion of digital modes and it is a real concern that if there are so many undocumented or hard to identify ones, it's effectively encryption. Perhaps a future rule change will address this.

Proprietary modes have never been illegal though. You can buy a receiver just like I can study morse code.

At the end of the day nobody gives a shit though. The technical hams care more about their freedom to do cool stuff and the majority would just as fast try to decode their own power supply spurs because they're messages from aliens.

Well, ham radio is an EXTREMELY diverse field. I mean, some people are using kilowatts and arrays of yagi antenna to transmit earth-moon-earth. Others are using minimal power (e.g. less than a typical bicycle lamp) to reach someone at the other side of the earth. Some are designing nice sound compression algorithms or algorithms for software designed radio in FPGAs on their self-designed board. Other try to build a full transceiver with less than 40 components total. Some are seeking contacts "rare" countries like other seek rare post stamps.

And most of this isn't really done to communicate. For that, most ham amateurs would use cell-phones, Facebook or whatnot. It's done for the sake of doing it. Or, in other words, to have an excuse to tinker with technical things.

That you can actually communicate is often just a consequence, a by-product. And the "communication" e.g. in contests to seek rare countries is actually really, really bland and boring. Translated to english: "Here is Mike, Calling Reunion, calling Reunion, calling Reunion, Mike here" - "Here is Jaques, Renunion, answering to Mike. I can your you very bad" - "Mike here, to Jaque. I can your you quite nice. And thanks for the talks ... calling Mongolia, calling Mongolia, here is Mike, calling Mongolia ..." --- and it's on PSK31 or CW even shorter.

You see, the ACTUAL communication is boring. The kick is probably to be able to manage and adjust your transceiver, wires, antenna, technical skills to actually be able to connect to these countries.

Or to be able to program and use in-real-life a sound compression algorithm with machine-learning and to test that with alike people on e.g. HF, Meteoscatter, via Satellite (which all produce very different channel characteristics).

Or to produce your own 6-layer PCP with an FPGA that actually does nice SDR algorithms.

It's a means to self-gratification.

The only two types of communications I have ever experienced in ham radio are (a) about ham radio or (b) extreme right-wing nutjobbery.

> Things like no encrypted communications really limits the utility.

How so?

Well they want to abuse ham radio spectrum to extend their home internet an extra kilometer, but they've just found out they can't, and that's annoying.

Which is for good reason. If everyone did that, it would be so congested that nobody could do that.

Why would you piss about using amateur radio for that when you can go 20-30km with cheap off-the-shelf microwave links?

In which frequency band? 2.4GHz?

2.4 or 5.8GHz - take a look at the NanoStation 5ACs. They cost pennies and have about 15km range.

If you want further you can spend about a ton and get a pair of LiteBeam 5ACs instead.

With a ham license you could use up to 1500W, which is around 200x what those have. If you were on top of a mountain and had a friend on top of a mountain 150 km away you could establish a reliable, stable point to point WiFi connection.

Over shorter range, but still more than what those low power devices can do, you could stream multiple 4K videos simultaneous where the low power devices would be crawling.

Not many people will need to do this, especially with the limits on encryption, but that's OK. Ham radio is not supposed to substitute for other radio services. It's supposed to be about learning about radio and providing a backup when other systems fail during disasters.

Well, no, you couldn't, because that's kind of not how it works.

Power doesn't make it go faster or further.

You can easily go 150km with a few hundred milliwatts and get 322Mbps with a few hundred quid's worth of equipment if you've got good clear line-of-sight, but you need to go to a way higher frequency than 5.8GHz.

Also, don't forget that there's a difference between TX power and EIRP - when I had my amateur radio Foundation licence and was limited to 10 watts, I was running 10W at the antenna port as measured on the terribly expensive Marconi test set - but well over 150W EIRP totally legally by simply using a decent aerial. This also meant that instead of blasting everything around with RF and only hearing fairly close stations, I could hear about as far as I could transmit.

Finally, I think you're grossly underestimating the cost and complexity of something that'll output 1500W on 2.4GHz, which would be incredibly expensive and difficult to operate.

Power literally does make it go further and with higher SNR (so faster).

No, it doesn't, it only does so under certain conditions.

For bidirectional communications it is *always* better to use a more efficient aerial.

It's interesting that you think "better SNR means faster". How do you think that works?

How would you stream a couple of 4K video feeds over say the 40m band? Do you think other limitations might come into play before SNR ever did?

The purpose of the ham license is to experiment with radio technology itself. If encryption was allowed, all channels would be occupied at all times by encrypted data coming from people with more powerful transmitters than you, and then how would you make any test transmissions?

Are you sure you're not thinking of the authentication carveout? Remote control of repeaters and other devices can be authenticated, but not encrypted, AFAIK?

Not one Pringles can utilized. We have failed to pass along the ancient knowledge.

Not a Senao 802.11a pcmcia card in sight

I vaguely remember some sort of stunt where they had got hold of two 3 meter dishes paired them with off the shelf wifi radios, mounted one on a pickup truck and drove out across a dry lake bed. 53 miles before signal loss is what I remember.

I will probably never be able to find the source, and my poor memory probably inflated the range like any good fish story.

https://www.techmonitor.ai/technology/wifi_and_bluetooth_dis...

Yeah, I met Radio when he was at MIT. He got a black badge from Defcon for it.

This are however just records. E.g. "I still have a connection".

These records aren't really usable in real-life, e.g. to view a movie. When it rains, or is foggy.

Just mounting a maneuvering 3 m dish to your laptop isn't everyones most desired leisure ...

> This are however just records. E.g. "I still have a connection".

This chatter about a successful 2005 Defcon shootout [0] indicates that a 125 mile link was able to pass traffic sufficiently reliably to be used for SSH, VNC, and other applications summarized as "etc" for several hours.

The one that somat appears to be thinking about was the year before and set up a ~55 mile link. The Wired article about it [1] doesn't say much about what data was shoved over the link, just that they "...shot a signal from one station to the second station with an encrypted message provided by the judges.", suggesting that at least some data transfer was required to consider the attempt valid.

> Just mounting a maneuvering 3 m dish to your laptop isn't everyones most desired leisure ...

That would have to be a very fancy dish to handle targeting a spot 1.5Km+ away while attached to a device that's being manhandled. I imagine it'd make much more sense to put the dish on a fixed mount and wire it up to the laptop, run wired ethernet from the AP to the laptop, or have a second radio in an AP to provide service to nearby clients.

[0] <https://groups.google.com/g/SOCALWUG/c/5T5wekafdv8>

[1] <https://defcon.org/html/links/dc_press/archives/12/wired_wif...>

I shot a 2.4 GHz path over 24 miles at 100 Mbps using Ubiquity bullets and their grid pack dishes. It's not that hard.

When we started wireless ISP in in early 200x, it was common to run 2.4 GHz links with 5+ km distance, having 24 dB antennas on each side. with the basic 17 dBm radio one was 20 dB over EIRP limit...

Here in Czechia, there was also another unlicensed band, 10.5 GHz, that has more sensible regulation - instead of EIRP limit, there was PA limit (power before antenna). So one can do long-range links with large enough antennas and most long range links was later converted to these. Unfortunately as this is local unlicensed band, radios for that were orders of magnitude pricier than generic wifi.

I’m pretty sure I’ve connected to a bog-standard Mikrotik Metal over a thousand feet. The surprise wasn’t that it connected, it was that the phone could transmit back.

We had this ~20 years ago: We created a 3 node network (the center node was also acting as relay) with 4 yagis on rooftops. The distances between the nodes were 1.5 km and 800 m. We used this for around 3-4 years till ADSL became available at our places. Worked flawlessly in clear sky. In heavy rain or snow it was very slow though. We were electrical engineer graduates at that time and built most of it from scratch.

I am still reminded of early WiFi and ways to extend it with Cantenna or foil backed contraptions made out of cardboard.

Used a cantenna and a PCI card with modified firmware to great effect when I only had WiFi for internet at the barracks in Japan in the 00's. I was absolutely outside of the legal limits on power, which I didn't really know at the time.

Extreme contention when 250 people had 3 APs as their only connection and half were trying to play WoW at the same time.

And no passwords! That's how I got Internet as a broke college student living off campus.

There is sub-GHz WiFi HaLow now which gets you to km distance without outrageously large antennas.

Completely unusable in Europe though due to ridiculously low duty cycles.

My personal record was between Nieuwport and Zeebruge, as there was line of sight between the 2 sites.

Made some diy wifi antennas: waveguide, yagi, etc...

Back in the day used Radiomobile software to calculate the link budget between the receiver and the transmitter.

Wi-Fi with directional antennas can make long distance links easily.

Just have a look at the European Hamnet network: https://hamnetdb.net/map.cgi - There are a lot of hill-to-hill connections with 60-100km range and plenty of SNR left. Even a few links around 150km.

In professional context these super long links need a very very precise link budget planing to get enough 'nines' in their reliability rating. One of the reasons why you are not seeing them too often.

I had great success with biquad cantennas! Yagis are potentially better but so hard to tune/match.

Operate under an amateur radio license (which overlaps with the WiFi bands) and you could run up to 1,500 watts. Couldn’t be used much for an actual network but if it was just an experiment to prove art if the possible with proper station ID could be fun.

Used to work at an evtol company and we used a plain ol router connected to a huge PoE antenna to connect to the vehicle which had some plain ol usb wifi antenna.

It was TCP/IP data streaming down and commands going up.

Demi-automonous testing at some 2km range. Worked great...

If you are trying to go point to point over a long distance, I would consider something like the unifi air fiber products. They've got a 60ghz model that is basically like a direct optical link. 15km+ no problem.

As a layman, is this supposed to be impressive when any phone can connect to Starlink direct to cell service for messaging, light browsing and even audio?

Depends: long distance links with a lot of reliable bandwidth is still very demanding.

Long distance links with only a few bps or kbit/s are a solved problem (LoRa).

aren't devices in north-america also allowed 1 watt on 900mhz?

I wonder how far that can go

I think I have an ancient wireless phone on 900mhz and all my weather sensors are on 900mhz

https://www.digikey.com/en/articles/unlicensed-915-mhz-band-...

There is IEEE 802.11ah that operates on that band, you can reach up to 1 km

https://www.cnx-software.com/2026/01/13/halowlink-2-wi-fi-ha...

I wonder why the yagi fell off so suddenly. It had pretty good height.

there's wifi ah, which runs on 900mhz band and has the same 30dbm limitation

I've used it with some raspberry pis to create hi-fidelity walkie talkies it's quite pleasant.

Not about range, but speed. Wifi is surprisingly slow. I noticed downloading a model from the internet (Huggingface) to my Desktop (connected by LAN cable) is faster than copying it from said Desktop to the laptop in the same room via Wifi.

Depends on the WiFi, you can easily get 1Gbps on WiFi 7 as long as your modem + access point have the bandwidth

Yeah. I ran a hand crimped cat-5 Ethernet run to my office and speed tested it, and it turns out my wifi is faster, according to speed.cloudflare.com. Haven't had a chance to debug that yet, but there's a switch and some other wiring in the way.

And both have the 4 antenna 2 way MIMO, which not all chips / laptops support yet. And not even all routers.

It's surprisingly complex actually.

Latency seems higher, much higher sometimes, and variable. I need to complete the LAN cable runs I started long ago :-)

I was hoping for a test of the old Pringles can antenna.

Reply on news.ycombinator.com