This somehow did not 'take off', it's a very interesting standard between low-bandwidth LoRA and high bandwidth Wifi, though meshed Wifi can provide more bandwidth than HaLOW at higher price.
I've been using it in a project. Previously used LoRa. For like 5 years, samples were available but FCC certification of modules (and equivalent in other countries) was coming "next year". Around 2022 some FCC-certified modules were available and finally in 2023 devices finally started appearing. Note that for commercial devices using RF you generally integrate a pre-certified module and then have to get the entire device certified in a simplified process as well. In practice companies that sell to hobbyists don't know better and don't do this.
LoRA is amazing but its max datarate is 27kbps for the sub GHz version and 254kbps for the 2.4GHz version and it's half duplex --you need to implement time multiplexing on your own or use multiple radios if you want two-way communication. It's great for something you would do over a serial cable but not great for networking.
802.11ah is however great for networking because you can buy an adapter and use TCP/IP just like ad-hoc wifi, although most products released so far use it internally and provide some other interface, like the one in the video. There is at least one access point available, and I'm kind of annoyed that's it is an unpopular way to use the technology. I'm afraid that 802.11ah won't take-off if most devices are proprietary bridges instead of following the wifi model of any device being able to connect to any access point.
Normal wifi works well globally with minor differences in power/range, and uses basically the same frequencies and bandwidths everywhere, with a channel or two excluded in some regions. And, normal wifi devices automatically conform to the region of routers. 802.11ah is restricted to use significantly different power, frequencies, bandwidths and duty cycles in at least 10 different regions. As a result, in Indonesia, it supports a few Mbps. In Europe and Japan it supports about 12Mbps. Only in North America does it support 78Mbps.
So far devices are usually only certified for use in North America, if at all. You have to go out of your way to even change settings to the correct region, if it's even possible. And since the most popular devices so far are proprietary bridges, users will buy a pair and unknowingly operate them illegally in Europe or elsewhere. The most likely outcome is import bans on devices not certified for a region. If I recall correctly using the wrong region setting can in some cases overlap with emergency services, broadcast TV and mobile networks so it will be taken seriously. I think the only realistic way for 802.1ah to "take-off" is for it to become a feature of normal wifi routers and have all the regional requirements locked on the router just like it is for the other wifi standards.
Thanks for the first hand insights, for "then have to get the entire device certified in a simplified process", is there some link for that process? I think using certified modules to do a full blown product is a common practice and I did not know there is a simplified process for that final product.
This page[1] has a pretty good explanation. "Simplified process" probably wasn't the correct term to use.
I think in a new product that integrates an RF module, you still have to at least do testing for unintentional emissions. And if you do something like change the antenna, in my understanding it's possible to update the FCCID instead of recertifying and getting a new FCCID but you'd need support from the original vendor which is probably unlikely. It is obvious that lots of small companies do not do any testing/certification and a lot of webshops practically encourage you to mix and match radios and antennas so I don't think this is well-enforced anywhere.
One funny thing is you're actually not supposed to change antenna types as an end user either. Radio module manufacturers are supposed to use proprietary connectors or integrated antennas to avoid this so the end user can't just buy a high gain antenna and connect it. Unfortunately (or fortunately) this usually means they use reverse polarity SMA connectors instead of SMA connectors. RPSMA connectors became so popular that today it's easier to buy an RPSMA antenna than an SMA antenna.
The ones shown in the video can be configured via an odd Windows EXE that lets you use AT commands to change the parameters. The manufacturers website contains a PDF that explains how to set it correctly for your region.
Yes, and I've seen ones that have drivers like normal wifi, that come with a python script and one of the lines is for changing the region but I think underneath, it also uses AT commands. And it's certified to use in US, Taiwan and Japan but nowhere else. It's barely documented except as comments in the script. In both of these cases, it's way more prone to being set to the wrong region than a normal wifi router and there is also a big incentive (increased datarate) to just use the US region.
+1, while looking into wireless networking options for an area of 0.5 x 2.5 km with medium bandwidth requirements in Europe, 802.11ah looked like the fitting standard, but with no hardware available.
Not sure about the legality specifically of this new Wifi mode/modules, but a lot of the companies selling other RF modules bare certainly know what they are doing is in a grey area. The FCC allows hobbyists to build up to 5 devices for personal use, and these sellers usually justify it as selling development boards for prototyping. The FCC doesn't seem inclined to enforce much against hobbyists, which is how you get stuff like obviously complete products on Tindie being sold under the "development board" exception. Which I'm glad for as a lover of those hacker gadgets, but the FCC could always come in and shut it down.
It's way too soon to say it didn't "take off". Chipsets haven't been available to integrators for that long. Maybe 1-2 years for non-early access customers? Plus, if you're judging success by looking at the consumer market, you're going to be disappointed. The target market is industrial manufacturing.
Its niche to me... but it has taken off where it's useful in custom point to point wifi applications. I play with this tech here and there along side several competitors in agriculture and industrial settings.
It is a bummer that 802.11ah did not took of. This technology has a lot of potential and the implementation is pretty easy, also for ultra low power applicatons. But for some reason, no one is using this technology.
802.11ah has seemed to be a kind of weird in-between. It was made to be a compete in the LoRa type space but LoRa already provides cheaper to implement hardware that works significantly farther. It is indeed slower but for the target use case of intermittent low data IoT that's usually more than fine... or at least not solved often enough by the realistic throughputs of actually low powered 802.11ah devices to create a significantly larger market.
What I mean by actually low powered 802.11ah is most of the time people 347 Mb/s max speed, ignoring that's peak for for a 4x4 16 MHz channel. That's almost a different world, not even high performance laptop chips find 4x4 worth the power budget, let alone something built for IoT. If you go to the real IoT client hardware in a realistic use case of 1x1 8 MHz suddenly your realistic goodput of 10s of megabits per second but that doesn't really enable too many additional use cases and comes with the aforementioned loss of coverage area and efficiency (it's more efficient than normal Wi-Fi but it's still Wi-Fi based).
Take that into consideration and what you have is a bunch of people getting excited about high speed 900 MHz when the standard was actually designed around IoT use case and demand, losing out to competitors which do it better, cheaper, farther, and came first.
Related: There are bunch of other weird sub 1 GHz standards from 802, even some under 802.11. They tend to take advantage of the TV spectrum. I don't think any have been popular, partially because that's a more complicated spectrum to participate in.
I've still found the best 900 MHz IoT radios to be LoRa radios that let you operate in raw radio mode instead and are basically 900 MHz USB to wireless-serial adapters. I bought several LoStiks back in the day and that's primarily what I did with them. More for true IoT use cases than trying to cram data over the smaller unlicensed sub 1 GHz bands though.
Yes-ish. The standard was released in 2017 so it's pretty comfortably over half a decade old at this point with very few devices available. And those devices that are available are pretty pricey and locked down given they are seen as more or less specialty equipment.
This compares to other WiFi standards which tend to proliferate quite extensively within a year or two of release. And LoRA (which HaLOW competes against) had a decent number of products available for far cheaper within the first few years after release (~2-3 years before HaLOW).
No modulation is going to make up for lack of line of sight (like a small change in ground elevation). These stated distances are not intrinsic. They rely on you getting both transceivers/antennas high up above surrounding terrain.
But dropping from 2.4Ghz and 5Ghz down to 900Mhz does significantly improve the penetration from varying forms of obstructions and makes tropospheric ducting more likely to occur.
In my experience it's not the houses or trees that block you. It's always changes in ground elevation, a small rise here or there. Using 902-928 MHz (actually a 5 Mhz channel centered around 906) did not increase my range versus 2.4 or 5 GHz options at the same height above terrain.
And the fresnel zone of 900 MHz is much larger than 5 GHz so in it requires more height above terrain than 5 GHz. Not great. Tropospheric ducting is so transient and rare it does not matter for this application. And it doesn't really help getting your signal across town.
I regret investing in a bunch of 900 MHz transceivers, bidirectional amps, filters, etc.
Would this work indoors through walls as well? It would be cool to have a wireless standard that only required one base station placed almost anywhere to guarantee a signal in every room of a house and outside in the whole yard, even if the bandwidth was limited. Seems like a lot of products could benefit from that.
Ring uses Z-Wave but needs range extenders for medium sized houses. Seems like it's not working in practice. I think those standards are for applications that need barely any bandwidth and need to run on coin cell batteries for years, I think there's room for a standard that is a little higher power but a lot more bandwidth, though still less bandwidth than regular Wifi. Something that you could actually browse the web on in a pinch.
Z-Wave has two range modes. The typical mode is designed to match the range of Wi-Fi to prevent urban areas from turning into a see of unusable noise because every suburban house and apartment is emitting kilometers of interference. "Long Range" Z-Wave devices are for the HaLOW IoT use case. 802.11ah does differ in that it has provisions to operate more as almost an actual bandwidth carrier than just an IoT network, though non-bridge type devices rarely implement the higher data rate provisions.
This is English dubbed Russian video of extending WiFi with proper antenna. There are series of these videos by Kreosan. They improve antennas over time. I believe they achieved range of a few kilometers. Wi-Fi frequency and chip are standard, antenna is the only component improved.
With a directional/high-gain antenna you're increasing the ERP (effective radiated power) which may well put you above the legal limits in your jurisdiction. So be careful.
Aside from that, a much simpler antenna is this one, which can be built from a metal can + connector + wire stub. Directionality is likely a bit worse but I've used it successfully over a distance a bit lower than a kilometer. https://web.archive.org/web/20060516103023/http://www.saunal...
Yes, be very careful to stay within lawful ERP. Some devices are illegal to use with external antennas. Of those that can connect an external antenna, the specs of acceptable antennas are provided with the equipment.
If you are in the USA, and you have the appropriate Amateur License (Tech class or higher), and you choose channels that are in the Ham Bands, Then it's perfectly fine to use these as ham radio links. https://www.arednmesh.org/
Perhaps most unusual about the pair is that in a war that has deeply divided Ukraine, where no one is indifferent, the two say they are neutral. Rather, the conflict is part of a landscape that emerges in rare moments, like the distant reports of artillery fire at the end of one video.
That neutrality is not a calculated stance to avoid offending viewers. It is one of faith.
“The lord Jesus Christ does not take sides, and we must strive to act in his image,” Mr. Kryukov said. “We study the Bible, and our members, we call them brothers and sisters, live in Russia and Ukraine. If we picked a side, we would become an enemy to someone. And we don’t want to be enemies with our brothers and sisters.”
If you can't pick a side between mass murderers and rapists building torture camps where dead kids once sat to learn in a language and culture the invaders are trying to erase that is a substantial moral error. Even jesus said love the sinner not ignore his sin.
The fact that they have themselves fled the war zone if anything makes it worse. They are hypocrites as well as cowards.
Big selling point here is that it runs on unlicensed ISM frequency bands. My question is why couldn't we use something like LTE-M on those same bands, why is LTE limited to licensed bands only? Is there something fundamentally different in these techs, or is it just cultural thing that 3GPP doesn't care about unlicensed users?
There’s LTE-U and LAA but they are more about extending normal LTE at the cost of public bands.
The basic problem is that LTE base stations assign time slots to devices when to access the spectrum - that doesn’t work well when by regulation you need to play nice with other technology on the same band.
> The basic problem is that LTE base stations assign time slots to devices when to access the spectrum - that doesn’t work well when by regulation you need to play nice with other technology on the same band
Is there good source to read more on this? Based on quick Wikipedia reading, Wifi traditionally (and including this 802.11ah) uses OFDM in sort of TDMA arrangement, which also includes AP assigning time-slots to stations if I understand it correctly (which I very much might not). LTE is based on OFDMA/SCFDMA which (again afaik) are just further refinements on the concept, but similarly rely on allocating resources (time-slots on sub-carriers). And later Wifi standards have also started to use OFDMA. So on surface-level its not super clear what here makes Wifi more suitable for unlicensed use... something maybe on how those time-slots are allocated, but its difficult to compare the two especially when they use somewhat confusingly (for an outsider) different terminology.
Sorry, I have been imprecise and my knowledge somewhat outdated. The basic issue is above TDMA/OFDMA; unlicensed bands require that you have some mechanism to detect if the channel is in use before you begin transmission, to ensure you do not hog the entire spectrum and can coexist with incompatible/foreign technologies. This is the best summary I could find:
„Supporting LTE over unlicensed bands is not
trivial. The key challenge is how to achieve har-
monious coexistence between LTE and other sys-
tems that are already operating in these bands.
Conventional LTE cannot operate in unlicensed
spectrum as it has no concern for cross-technol-
ogy coexistence. For example, transmissions in an
LTE radio access network (RAN) are continuous
in time, and subject to centralized scheduling at
the eNodeB (eNB). Even in the absence of data
traffic, control and reference signals are transmit-
ted over the air (at the OFDM symbol level) and
are ubiquitous over time and its channel band-
width.“
Bluetooth LE has a similar long-range mode since Bluetooth 5, and it even seems to have been supported by some iOS beta versions (which makes sense given that it’s only a new coding scheme and not a completely different frequency).
It’s a shame it isn’t more popular. There are many situations where I’d love to be able to chat with people a few hundred meters away without a network connection: Friends sitting in a different row on an airplane/train, supermarkets with bad cell signal, when hiking…
As an aside… I am been amazed lately at the Bluetooth range I do get. I remember when the technology was in its infancy and it was extremely short range… like same room no walls if you are lucky. These days I can leave my laptop at the docking station and walk nearly across the building before having issues.
There are three Bluetooth transmission power classes, and traditionally, most devices (including headphones and headsets) used to be Class 2, which has a range of roughly 10 meters, but Airpods are Class 1, which has 10 times that.
There's no real signal processing magic happening other than (significantly, due to the inverse square law involved) bumping up the maximum transmit power from 2.5 to 100 mW.
Both ends of a connection need to support Class 1 for that to work, but at least Apple's devices have consistently done so for many years now as far as I know.
What's the point of having different classes? Was it harder to make the higher power chips when Bt was invented?
Now almost everything can do at least 25mW or so, and with dynamic TX power it's not like having too many mW is a problem, it will just turn it down if not needed.
> There are many situations where I’d love to be able to chat with people a few hundred meters away without a network connection
It's a shame because the necessary technologies and protocols have all existed for a decade or more: XMPP over mDNS-announced endpoints and wifi in ad-hoc mode was supported by Pidgin many years ago.
70 comments
[ 4.3 ms ] story [ 76.9 ms ] threadLoRA is amazing but its max datarate is 27kbps for the sub GHz version and 254kbps for the 2.4GHz version and it's half duplex --you need to implement time multiplexing on your own or use multiple radios if you want two-way communication. It's great for something you would do over a serial cable but not great for networking.
802.11ah is however great for networking because you can buy an adapter and use TCP/IP just like ad-hoc wifi, although most products released so far use it internally and provide some other interface, like the one in the video. There is at least one access point available, and I'm kind of annoyed that's it is an unpopular way to use the technology. I'm afraid that 802.11ah won't take-off if most devices are proprietary bridges instead of following the wifi model of any device being able to connect to any access point.
Normal wifi works well globally with minor differences in power/range, and uses basically the same frequencies and bandwidths everywhere, with a channel or two excluded in some regions. And, normal wifi devices automatically conform to the region of routers. 802.11ah is restricted to use significantly different power, frequencies, bandwidths and duty cycles in at least 10 different regions. As a result, in Indonesia, it supports a few Mbps. In Europe and Japan it supports about 12Mbps. Only in North America does it support 78Mbps.
So far devices are usually only certified for use in North America, if at all. You have to go out of your way to even change settings to the correct region, if it's even possible. And since the most popular devices so far are proprietary bridges, users will buy a pair and unknowingly operate them illegally in Europe or elsewhere. The most likely outcome is import bans on devices not certified for a region. If I recall correctly using the wrong region setting can in some cases overlap with emergency services, broadcast TV and mobile networks so it will be taken seriously. I think the only realistic way for 802.1ah to "take-off" is for it to become a feature of normal wifi routers and have all the regional requirements locked on the router just like it is for the other wifi standards.
I think in a new product that integrates an RF module, you still have to at least do testing for unintentional emissions. And if you do something like change the antenna, in my understanding it's possible to update the FCCID instead of recertifying and getting a new FCCID but you'd need support from the original vendor which is probably unlikely. It is obvious that lots of small companies do not do any testing/certification and a lot of webshops practically encourage you to mix and match radios and antennas so I don't think this is well-enforced anywhere.
One funny thing is you're actually not supposed to change antenna types as an end user either. Radio module manufacturers are supposed to use proprietary connectors or integrated antennas to avoid this so the end user can't just buy a high gain antenna and connect it. Unfortunately (or fortunately) this usually means they use reverse polarity SMA connectors instead of SMA connectors. RPSMA connectors became so popular that today it's easier to buy an RPSMA antenna than an SMA antenna.
SparkFun for example definitely is aware: https://www.sparkfun.com/tutorials/398
What I mean by actually low powered 802.11ah is most of the time people 347 Mb/s max speed, ignoring that's peak for for a 4x4 16 MHz channel. That's almost a different world, not even high performance laptop chips find 4x4 worth the power budget, let alone something built for IoT. If you go to the real IoT client hardware in a realistic use case of 1x1 8 MHz suddenly your realistic goodput of 10s of megabits per second but that doesn't really enable too many additional use cases and comes with the aforementioned loss of coverage area and efficiency (it's more efficient than normal Wi-Fi but it's still Wi-Fi based).
Take that into consideration and what you have is a bunch of people getting excited about high speed 900 MHz when the standard was actually designed around IoT use case and demand, losing out to competitors which do it better, cheaper, farther, and came first.
Related: There are bunch of other weird sub 1 GHz standards from 802, even some under 802.11. They tend to take advantage of the TV spectrum. I don't think any have been popular, partially because that's a more complicated spectrum to participate in.
This compares to other WiFi standards which tend to proliferate quite extensively within a year or two of release. And LoRA (which HaLOW competes against) had a decent number of products available for far cheaper within the first few years after release (~2-3 years before HaLOW).
And the fresnel zone of 900 MHz is much larger than 5 GHz so in it requires more height above terrain than 5 GHz. Not great. Tropospheric ducting is so transient and rare it does not matter for this application. And it doesn't really help getting your signal across town.
I regret investing in a bunch of 900 MHz transceivers, bidirectional amps, filters, etc.
Watch your Fresnel Zone!
This is English dubbed Russian video of extending WiFi with proper antenna. There are series of these videos by Kreosan. They improve antennas over time. I believe they achieved range of a few kilometers. Wi-Fi frequency and chip are standard, antenna is the only component improved.
Aside from that, a much simpler antenna is this one, which can be built from a metal can + connector + wire stub. Directionality is likely a bit worse but I've used it successfully over a distance a bit lower than a kilometer. https://web.archive.org/web/20060516103023/http://www.saunal...
Made a GIF of the transit of the milky way using this software on a raspberry pi with some SDR/RF equipment. https://github.com/byggemandboesen/H-line-software
Unfortunately I only have some images on my phone right now, a git page is on my todo.
If you are in the USA, and you have the appropriate Amateur License (Tech class or higher), and you choose channels that are in the Ham Bands, Then it's perfectly fine to use these as ham radio links. https://www.arednmesh.org/
https://www.nytimes.com/2015/07/18/world/europe/in-bleak-ukr...
Perhaps most unusual about the pair is that in a war that has deeply divided Ukraine, where no one is indifferent, the two say they are neutral. Rather, the conflict is part of a landscape that emerges in rare moments, like the distant reports of artillery fire at the end of one video.
That neutrality is not a calculated stance to avoid offending viewers. It is one of faith.
“The lord Jesus Christ does not take sides, and we must strive to act in his image,” Mr. Kryukov said. “We study the Bible, and our members, we call them brothers and sisters, live in Russia and Ukraine. If we picked a side, we would become an enemy to someone. And we don’t want to be enemies with our brothers and sisters.”
They apparently has left Ukraine since the full-scale war broke out, and travelling around the world: Thailand, India, Kazakhstan, etc.
The fact that they have themselves fled the war zone if anything makes it worse. They are hypocrites as well as cowards.
Please.
The basic problem is that LTE base stations assign time slots to devices when to access the spectrum - that doesn’t work well when by regulation you need to play nice with other technology on the same band.
Is there good source to read more on this? Based on quick Wikipedia reading, Wifi traditionally (and including this 802.11ah) uses OFDM in sort of TDMA arrangement, which also includes AP assigning time-slots to stations if I understand it correctly (which I very much might not). LTE is based on OFDMA/SCFDMA which (again afaik) are just further refinements on the concept, but similarly rely on allocating resources (time-slots on sub-carriers). And later Wifi standards have also started to use OFDMA. So on surface-level its not super clear what here makes Wifi more suitable for unlicensed use... something maybe on how those time-slots are allocated, but its difficult to compare the two especially when they use somewhat confusingly (for an outsider) different terminology.
„Supporting LTE over unlicensed bands is not trivial. The key challenge is how to achieve har- monious coexistence between LTE and other sys- tems that are already operating in these bands. Conventional LTE cannot operate in unlicensed spectrum as it has no concern for cross-technol- ogy coexistence. For example, transmissions in an LTE radio access network (RAN) are continuous in time, and subject to centralized scheduling at the eNodeB (eNB). Even in the absence of data traffic, control and reference signals are transmit- ted over the air (at the OFDM symbol level) and are ubiquitous over time and its channel band- width.“
From Huang, Yan, et al. "Recent advances of LTE/WiFi coexistence in unlicensed spectrum.", use https://doi.org/10.1109/MNET.2017.1700124 in sci-hub.
https://github.com/Ysurac/openmptcprouter https://www.openmptcprouter.com/
I mentored the port of MPTCP to OpenWRT years ago, and OpenMPTCPRouter took some of this work in their port.
It’s a shame it isn’t more popular. There are many situations where I’d love to be able to chat with people a few hundred meters away without a network connection: Friends sitting in a different row on an airplane/train, supermarkets with bad cell signal, when hiking…
There are three Bluetooth transmission power classes, and traditionally, most devices (including headphones and headsets) used to be Class 2, which has a range of roughly 10 meters, but Airpods are Class 1, which has 10 times that.
There's no real signal processing magic happening other than (significantly, due to the inverse square law involved) bumping up the maximum transmit power from 2.5 to 100 mW.
Both ends of a connection need to support Class 1 for that to work, but at least Apple's devices have consistently done so for many years now as far as I know.
Now almost everything can do at least 25mW or so, and with dynamic TX power it's not like having too many mW is a problem, it will just turn it down if not needed.
It's a shame because the necessary technologies and protocols have all existed for a decade or more: XMPP over mDNS-announced endpoints and wifi in ad-hoc mode was supported by Pidgin many years ago.