Glass Antenna Turns windows into 5G Base Stations
spectrum.ieee.org
spectrum.ieee.org
Let's see what they can do for a commercial product. Usually, there are tens of antennas on a single tower so they can't all look like this. Also, I'm going to assume that you have to keep anyone from getting within 3 meters just due to radiated emissions, so don't go just looking out that window!
- innocuous: not harmful
- inconspicuous: not clearly visible or attracting attention
This seems not at all unreasonably subtle to me. Even with the array of feeder lines, yeah, maybe it's not for very high end stash places but for most places this seems ay okay.
Given what the alternatives are for urban and commercial spaces, this feels like a big win.
My main concern is power level. How much power can you emit if Joe in accounting is 8 feet away from it, and how does that compare versus normal building mounted or pole mounted antennas? Also, what frequencies is this antenna designed for; it seems like 5g can run on lots of spectrum; is this mmWave gear or lower?
Apologies for soapboxing, but I want to chip in my belief that this world is driven by those who see possibility & potential.
Cynics never lose but optimists win.
My thoughts exactly. Who would like to sit that close to a 5G Base Station?
That was my first takeaway from the photo from outside. The kinds of antennas they put on top of buildings routinely run many hundreds to a thousand watts or more of power directionally out into the city. That's fine when you're putting it on equipment outside the building on a controlled access roof pointing away from the occupants in the building. Everyone actually in the beam pattern is going to be far away from the active elements.
This design doesn't seem to be incredibly directional especially outwards. You're not going to be able to run much power on that antenna, and now you're going to have it on the inside of metallized glass. A lot of that energy is going to stay in the building. I wouldn't want the desk next to this if it's going to run even 100W. Just asking to get some good RF burns.
Right. The point of these small cell sites is usually to improve coverage within the building.
Occupational RF exposure is pretty strictly regulated in most countries. I’m sure there is design/installation guidance to ensure they stay well within legal limits.
That's not what the article is stating. If that was its use, there are plenty of 5G antennas that can look like any of the other warts commonly found on office ceilings like smoke detectors and other wireless ap's and what not.
> attached to a building window inside and turn the outdoors into a service area
These aren't specifically for indoor coverage, its specifically for outdoor coverage.
The article says it's for the "sub-6" 5G bands, a.k.a. normal cellular frequencies, not mmWave.
As always, these are non-ionizing frequencies, they pose absolutely zero risk to health or safety unless you're absorbing enough power to be meaningfully heated by it.
> How much power can you emit if Joe in accounting is 8 feet away from it, and how does that compare versus normal building mounted or pole mounted antennas?
Assuming an antenna gain of 10 dBi, which seems to be "normal" for panel-style antennas in the 5G low band, just short of 30 watts in to the antenna would be safe according to the guidelines the FCC gives us amateur radio operators for "uncontrolled" environments if the antenna were aimed directly at a person eight feet away.
Obviously in the real world these antennas will be aimed outward so the energy being absorbed by anyone in the building will be significantly less than that.
These should not be installed in places someone could directly touch it or the cables feeding it, but there's no reason to believe there's any danger to someone just existing normally in the same room.
It's fairly obvious that there are thousands of different ways to camoflauge this equipment in a real-world customer deployment, just like how routers, etc., are hidden in restaurants and stores.
How long will be the useful life of these antennas be, compared to the useful life of a building that is still early enough in planning for significant integration?
I've never heard a single person complain about wall/ceiling mounted WiFi access point. Do you really think this is something that justifies the added complexit
Glass facades almost universally use Low-E glass to avoid turning the building into a huge greenhouse. Problem for 5G, is that low-e glass is remarkably good at blocking 5G frequencies[1]. Pair that with 5G smaller propagation distances, and issues of finding viable locations to mount 5G antenna becomes a real problem.
This product neatly solves that problem by allowing carriers to mount these antenna on the inside of a buildings facade, while providing coverage outside the building. Which will substantially reduce the cost and difficulty of installing 5G masts. You can place all your sensitive equipment in normal building voids, without the need for bulky and ugly weather proofing, and you need to break the buildings weather tight seals (which a landlord isn’t gonna let you do without significant assurances you’re going the cover the costs of any water that comes through) to run cables to external antenna.
To make all of this viable, someone has had to do a fair bit of work to figure out how to build an antenna that effectively incorporates the low-e window it’s attached to, into its RF design. The fact the physical antenna is made of glass and partial transparent isn’t actually the interesting part. That’s likely been done because glass is a very rigid material that will make it easy to ensure the conductive parts of the antenna are kept at a specific distance from the window it’s mounted on, to ensure the correct RF coupling occurs.
[1] https://www.ranplanwireless.com/gb/resources/low-e-glass/
“I don’t think the idea for using transparent conductive materials as an antenna existed before"
The rest of the design is just the usual microwave antenna black magic, but with a different propagation velocity plugged in.
Those electrodes are literally used as capacitive antennas to detect the position of your fingers and they range from about 300ohm/sq to 1ohm/sq. Depending on the capacitive coupling they range from GHz to 100kHz bandwidths.
Just for clarification, the reason that ITO can be transparent even though it is conducting is due to a fairly novel effect where the bandgap of the material is just wide enough to allow most visible light (red-blue) through while still allowing electron conduction (degenerate bands due to Sn doping?). It's pretty cool. Most conductors (metals) have conduction bands that reflect visible light (though they might let X-rays or IR through). Most transparent materials (eg. glass, water) are insulators which have wide band gaps, but no conduction carriers (electrons or holes).
One counter example is Ruby (chromium doped sapphire) which looks red. If you heat it up the bandgap narrows and you it turns dark/black because only IR can get through, while if you cool it in LN2 they will turn light pink as the bandgap width increases!
I don't know that a white box on the inside of the same window (which covered the wired coax connections) would be that much more conspicuous, especially from the outside. Maybe they require special exterior windows, but those don't seem to be part of the very visible "transparent" antennas. If you lowered the drop ceiling anything would be less conspicuous on the inside.
The product is for fitting into existing buildings, with minimal impact. Being inconspicuous is a secondary concern, it only needs to be inconspicuous enough to not be obvious. Drop the ceiling on an entire floor to hide would not be low impact.
As to a plastic white box, sure you could do that. But it would be a plastic box that contains some kind of antenna on a rigid substrate. At which point you might as well just use glass as the substrate, and get rid of the extra enclosure, and manufacturing fiddlyness involved in assembly.
The cost of a glass antenna vs one housed in a plastic box is going to be negligible compared to overall cost of the installed equipment. At that point you might as well just use glass, simplify the construction and install process, and get a product that’s less visually distracting as a bonus.
The idea that a plastic enclosure is difficult, expensive, or fiddly, is kinda hilarious. Maybe you'd like your monitor or laptop or microwave oven to have it's fiddly enclose removed, but I don't think it's wise.
You could easily enclose this by some architectural feature on the interior of the building or even use a window that's off the back of a maintenance closet.
Those towers you see with lots of antennas are massive MIMO installations designed for very high capacity and coverage over a wide area. But not all sites need to look like that. In this case, it’s just a small cell designed to improve coverage within a building and/or on a few local streets. Power levels are also much lower, not all that much different to a WiFi base station. People aren’t going to get cooked if they get close to it.
change my view
Mesh would be each home (or some percentage of the homes) act as nodes. These have all the homes hit a few towers around the city. Traffic isn't routed directly between (or through) the homes in this example, it is all centralized. They hit a single big tower that then does all the routing.
(It doesn't help that people in this thread are confusing "mesh" with "collection of access points each of which is individually connected to a wired ISP", which is not a mesh.)
However, creating the Wikipedia/Internet Archive of wireless ISPs would be pretty awesome.
Take a look at WiFi-dense apartment buildings. So much crowding, no centralized assignment or management of the bands. It is a wild west of people transmitting on whatever channels and whatever power levels they want (within the legal limits). It ends up with few people actually having a good experience when there's no centralized management. 5GHz/6GHz makes WiFi more usable because it naturally limits your ability to hear your neighbors. Going to 700MHz/900MHz/1.2GHz (the normal frequencies used in a lot of 5G deployments) is only going the opposite direction of where WiFi has been going to solve this problem. Expect more noisy neighbor problems as you lower the frequencies.
Then we're not only going to saturate the bands with people doing whatever they want (within legal limits), we're going to depend on mesh routing through all that noise? There goes your reliability and efficiency of sending data.
And who's to say they want to join your mesh and not Bob's super awesome mesh? Or start their own mesh? Oh, you get to decide how to operate the mesh but I can't? I guess you'll end up getting some kind of license so you can standardize how this particular mesh should operate and prevent others from running competing services on the same frequencies as your one mesh.
You'll put out standards on what kinds of devices are certified to work on it and ensure certain settings so tx/rx errors are reduced to ensure good usage. You'll start encouraging people to not put up more nodes in a certain area because it's just getting too crowded here, but hey we need to incentivize someone to set up a node on the other side of town.
Snap now it seems like we're running a regular carrier.
I participate and use city-sized WiFi mesh networks in the amateur radio world. They're not anywhere near a replacement for what normal people think of as internet connectivity. I can't imagine swapping WiFi for 5G cellular stacks would end up making a radical difference. The issues are largely with having to make multiple wireless hops, mesh routing inefficiencies/problems, and having everyone actually play nice all the time.
That's not really a "mesh" then. It is just a bunch of infrastructure AP's everywhere.
If it's a part of every modem/gateway router, why would you bother routing it through a bunch of mesh hops just to eventually get out instead of just routing it through the far more reliable wired networking available at every modem/gateway router?
Those regular WiFi networks only have tons of available bandwidth because they're not trying to repeat a bunch of wireless traffic. Even the current mesh WiFi networks only really work when you're using frequencies that aren't trying to compete with neighbors. Start getting actual density and it'll all fall apart.
Also your idea of "standard QoS can still apply" isn't exactly true. That QoS is only going to work if people play along with it. In the end its a shared medium. Get some clients to not play along with your configurations, you'll start getting collisions regardless of what you configure your QoS settings.
The question is can devices connect to other devices that route to router with internet. Is it possible to have router for house without internet connection that routes to the neighbors that do?
Why would it be part of every modem/gateway? Since there's no monetary incentive to participate, in all likelihood all nodes would be run by volunteers who are shelling out extra for a compatible modem/router.
Actually come to think of it, you can run a volunteer network providing internet connectivity with off the shelf equipment right now. It's called setting your wifi network to "open". Why don't people do that? How would your mesh network fix those issues?
The bandwidth you care about in this scenario, where you don't have backhaul, is in the air. And very limited.
Not used everywhere, but seen as something that would be rolled out for critical communications, natural disasters, etc.
My neighbor still has the Helium antenna and radio on his balcony but it's offline due to costs/profits disproportion. It's the LoRaWAN, pre-5G hardware though, and I don't know anyone running the 5G version, if it's even a real thing. I liked the idea from technical perspective but the project itself was off-putting for me due to being built around a crypto token and having overall web3 smell.
So, if you want to run legally, you're going to need spectrum licenses and transmitter licenses and all that. That will make you a mobile network operator, regardless of how you arrange labor and sites.
The biggest hurdle is that reliably running high performance transmitters is not easy for amateurs, and the payoff for any one transmitter is not that much. I'm going to use the example of a residential ISP but this applies to cell networks as well. The "meshier" the network is, the more people revenue needs to be split between, exacerbating the problem.
Another issue is that reliability is extremely important for internet access. Given the fact that amateurs are not going to be able to maintain high uptime, for a decentralized mesh network to succeed at actually providing internet service, you need to have a lot of redundancy in any given area, further reducing income from any one node.
The solution to this is to have a team of technicians that can go around and fix and optimize nodes as soon as there is any problem. This is basically what an ISP or cell carrier does. An added difference is that in a mesh network, the idea is generally that the property owner owns the node, while with a conventional ISP, the property owner leases to the ISP who owns the node. Property owners generally prefer the latter, since this is the model they are used to operating under as landlords.
You'd just need one crooked technician who can recruit some number of node owners. (If the tech did it via remote-controlled power interruption and only did it on scattered nodes in areas with lots of redundancy, they wouldn't even need to recruit node owners and split the cash. But they'd also be leaving evidence that could easily get them in trouble once someone started getting wise to it.)
An end user wants to pay a steady monthly fee for internet that never goes down. A property owner wants to get a steady monthly check for leasing a site. A technician wants to get a steady salary for fixing nodes.
A decentralized repair network is likely to reduce reliability for end users, predictability for property owners, and job security for technicians. All three of these parties may find it more optimal to have an ISP business which can finance and coordinate things, even if the business is taking profit which could have gone to the other participants.
One, what frequency are you going to use? If you use 2.4GHz or 5GHz, your Wifi-using neighbors will hate you. There is the 6GHz spectrum but has problems with long ranges. The 3.5GHz CBRS is probably the best bet but that requires spectrum allocation and organization to run it. The mobile operators have all the good low frequency, long range spectrum.
Two, the range with home routers is going to be pretty short, maybe 1mi. That means lots of node to cover a city. Also, 5G routers are not that cheap. It also means that there will be no reception away from the city. Most routers are meant to be used inside, and good coverage, requires mounting them outside on a pole.
Three, I'm not sure there is 5G device-to-device. There was LTE-Direct but it never got implemented. There D2D in 5G spec but I can't find any implementations.
Many slightly older cars (2000-2020) had antennas embedded in glass. The idea is solid. Antennas in glass are protected, so they can be very thin and almost invisible, more aesthetically pleasing than a shark fin or a rod on the roof.
I would consider this an aesthetic choice, not so much engineering. A small antenna sticking out on the roof solves the engineering problem adequately.
Though you are probably right and many cars with just FM/AM and DAB still put them in windshields.
I’m not sure about that 3+ antenna claim for FM. Do a fact-check there. ;) I think most diversity antenna systems that you describe use 2. And they are still considered “premium”. 3 is a bit overkill for FM, FM is very resilient against obstacles.
If I’m wrong, would be interesting to see an example of a car that uses 3 or more. Probably far outside of my pay grade to say the least.
They really didn't want you losing signal!
This isn't even mentioning the festoon of other antennas for DAB, mobile (which has a backup too and phones home in a crash: in case the shark fin is broken in a crash)
It’s cool they did that but it’s almost like they are trying to prove some point — maybe to be the best in the market for FM by a hair. :) One antenna is more common than 3 to my knowledge, by far.
This is probably so the car can act as a Wifi hotspot, with the Wifi antennas located in the interior rather than in the sharkfin.
It also can act as a hotspot if I bothered paying for its data plan. I'd rather if I could just load an eSIM for my existing service to have it be a hotspot for me. Since its antennas are on the outside it should get way better GPS and cellular connectivity than my phone in my pocket deep in a metal box.
I've seen what you are talking about, too. My VW RNS had settings to use the car's 4G for passenger WiFi.
Cars really want to phone home these days
It had a magic box that used feedback from the stock radio's IF section to (try to) erect the best pattern for good reception at any given time, and which presented one unified RF output to that radio's singular input.
Cadillac has used the same, or very similar, system as well.
Which is neat and all, I guess, but we seem to be getting off into the weeds: These antenna elements were not invisible at all except for the fact that they blended in with the heated defroster elements that were also printed onto the glass.
https://press.siemens.com/global/en/pressrelease/obb-puts-fi...
Most modern windows use an aluminuzed coating for UV reflection (usually called low-e glass), which surprise surprise, is absolutely great at attenuating mm-wave frequencies, making windows pretty much the worst possible place on a building to place an antenna.
But, turns out most people also dont need gigabit wifi for their phones and other devices, so true mm-wave 5g seems to mostly be reserved for wireless home internet at the moment.
"A self-healing multispectral transparent adhesive peptide glass" https://www.nature.com/articles/s41586-024-07408-x :
> Moreover, the supramolecular glass is an extremely strong adhesive yet it is transparent in a wide spectral range from visible to mid-infrared. This exceptional set of characteristics is observed in a simple bioorganic peptide glass composed of natural amino acids, presenting a multi-functional material that could be highly advantageous for various applications in science and engineering.
Is there a phononic reason for why antenna + window?
Transducer: https://en.wikipedia.org/wiki/Transducer
https://news.ycombinator.com/item?id=41442489 :
> FWIU rotating the lingams causes vibrations which scare birds away.
But it is clearly an AI-generated image. Look at the reflection of the lamp in the window, it overlaps the window frame.
Is this now the norm, AI-generated images pawned off as begin real? Good lord...
The shit you see at the top are the wired connectors for the antenna that is inside of the window.
Are bands of that width not typically delivered with mm waves?
(There's more definitions of the word "bandwidth" than counting bits per second. It has, at least, uses in both RF and in data networking -- and the former use is predates the latter.)
i don't know why you're bringing up bits per second. neither i nor the clueless loser who wrote the article were talking about bits per second
the quote doesn't say 'designs that deliver 10 to 50 gigahertz of bandwidth typically use millimeter waves'. it says the opposite: 'millimeter waves can deliver typically between 10 and 50 GHz of bandwidth'. that's clueless nonsense. it's not even wrong. bandwidth is delivered by a medium, not by a signal in it; the signal is closer to being what the bandwidth is delivered to. the signal occupies or consumes or has bandwidth, which is close to the opposite of delivering it. and a millimeter-wave signal can be of any bandwidth at all up to about 300 gigahertz, including—obviously, one would hope—bandwidths of well under a megahertz. (in theory you could transmit or detect a millihertz-bandwidth millimeter-wave signal, but that probably requires exotic instruments like an atomic clock.)
it's complete nonsense to try to describe the 'typical' bandwidth of a millimeter-wave signal. it's like talking about the typical mass of objects made out of atomic matter, or the typical distance traveled by photons. is jupiter or a tardigrade a more typical-sized atomic-matter object? it depends entirely on context. there is an upper limit (objects much larger than jupiter will stop being made of atomic matter pretty soon) and a lower limit (probably you need at least a dozen or so atoms of lithium before you have an 'object') but there are many orders of magnitude of slop within that limit
if we were talking about bits per second, it might make sense to talk about a signal delivering something, but even in that case the information capacity of the signal depends on many more considerations than just the wavelength. the bandwidth, for one, but also the relevant sources of noise, the transmit power, the path loss, and the efficiency of the coding scheme used. so you'd still have a lot of orders of magnitude of slop, and plenty of mmwave signals aren't even used for communication, so trying to characterize their shannon capacity is a somewhat questionable enterprise
Here's another generalization: A car typically weighs between 1 and 3 tons, and typically travels on the highway between at speeds between 50 and 80MPH.
This generalization misses (many!) rather common outliers, and that's OK since it not meant to be particularly precise. That's the way of generalizations: They're generalized.
Meanwhile, please take a moment to read this: https://news.ycombinator.com/newsguidelines.html
but, as i explained, that's not the only dimension along which the statement is nonsense; millimeter waves also don't deliver any hertz of bandwidth. it's a really multifaceted gem of cluelessness
the implication of the last line of your comment seems to be that you think i am not familiar with the site's guidelines, but it is unclear why. perhaps you think criticism of the ieee's editorial and human resources choices runs counter to the guidelines? perhaps you think there is nothing to be learned from my criticism? you'll have to be more specific if you want to have a substantive conversation
I have nothing further to discuss.
(Full disclosure, I'm a previous employee)
Can imagine both good and horrible points to that happening.
More seriously, the terrible nature of start up/project names meant I actually read this as being
“Glass Antenna[software name] turns windows [operating System] in base stations [an actual base station or network hub]”
And it was not until I saw Ieee.org that I realized it actually meant real antennas, and even then my incorrect interpretation of the headline still seems plausible to me
> Because 5G networks include spectrum comprising higher frequencies than 4G, base stations for 5G networks serve a smaller coverage footprint.
Sure
> It [the window antenna] is compatible with frequencies in the 5G Sub6 band—meaning signals that are less than 6 gigahertz (GHz). Sub6 antennas represent critical portions of a 5G deployment, as their lower frequency ranges penetrate barriers like walls and buildings better than the substantially higher-bandwidth millimeter-wave portions of the 5G spectrum.
But 4G seems to go into at least the 3-ish GHz range just fine. At least my layman understanding is that sub-6Ghz doesn’t have the range problem, the whole point of adding that (IMO, less-than-intuitive-to-consumers) frequency band was that (while it didn’t fulfill the bandwidth promises of 5G) it also didn’t have the range problem. So it is there to fill the gaps.
The ability to deploy sub-6GHz antennas everywhere seems like it misses the whole point of that band.
I’m sure there’s some advantage to the 5G tech in general, because it is newer. But that’s a different pitch, right?
5G is still better than 4G on sub6 for lots of reasons, but yeah, it doesn't have the oodles of bandwidth that 5G promised.
Being able to put more sub6 base stations in more places is still good though. There's plenty of areas with poor coverage, and sometimes the reason there's no coverage is the aesthetics of the base stations are poor. Blending in helps. Which is why some antennas look like weird saguaro or palm trees and pine trees. It might be nice to have antennas in windows in office buildings instead hanging on the side, and it might be easier to install as well.
Here’s an interesting paper on how to make it work efficiently:
https://www.mdpi.com/2079-9292/12/24/4983#:~:text=Silicon%20....
Silicon is not a conductive metal. Its a semi-conductor, it needs doping to become a good conductor. That's why its used in IC's. Naturally not very conductive but react a little with something else and suddenly it becomes a pretty good conductor. Make a mask of the channels where you want that conductivity, and suddenly you can draw little wires.
>NTT Docomo reports that it uses transparent conductive materials as the basis for its antenna, sandwiching the conductive material along with a transparent resin, the kind used in laminated windshields, in between two sheets of glass.