'5G' is lame/yesterday's tech. Its all about lifi nowadays. Way faster than 5g or any wifi based data transmission schemes due to having much higher bandwidth (visible light spectrum is much larger than the wi-fi one).

Does it work in the rain?

Absolutely not. Li-Fi is extremely unreliable.

Visible light is easily blocked by trees, rain, snow, smog, fog, etcetera.

Which is why you put it in the street lights, guard rails, etc. Li-fi is far more efficient and uses way less power.

More deployment sites means a commensurate increase in site licensing and maintenance. With the number of sites you'd need to make lifi as reliable as a wired connection, these other costs will probably kill it. Especially because now you'll have to have periodic transmitter/receiver cleanup (dust, bird poop/nests, spider webs, etc), potentially non-stop if it's snowing and the snow is accumulating on transmitters or receivers. Not to mention the cost of cleaning up damage like graffiti. And while Lifi itself might be energy efficient, you still have to power it, and now have a cost to wire virtually everything for power, which makes everything a potential electric hazard.

And this is just what I could come up with in 5 minutes off the top of my head. The concept behind Lifi is decades old, and it wasn't even feasible enough to take off indoors where the challenges are nowhere near as difficult.

In rural areas there may be no street lights, the guardrail may only have reflective paint, and no one is interested in paying for the running of miles of electric power for LiFi. Might as well run fiber and be better off in the long run. LiFi is a technology for some niche indoor markets where the outside ecosystem cannot intrude much. I see it as a point to point communication system for indoor areas where for some reason you don't desire to run wires or fiber. The problem is that it only takes one random rat or measly mouse to stand in front of the light beam and your data transmission rate goes to very close to zero.

Or you could run fibre to every building and lifi for the rest. Or laser-fi to airplanes. Lifi works in any condition, it just depends on the strength of the light source and directionality to the receiver. It doesn't even need to be LOS because reflections work as well. It would be cheaper and safer to have 'light towers' on tops of buildings than 5G microwave ones.

This is ridiculous. 500 feet isn’t nearly close enough. If only 30% can receive a useful signal at that distance, then as others have said, these may need to be placed, not every 1,000 feet, but every 250 feet. In a 1,000 foot square, that could involve gridding that square into a 4 x 4 partition - 16 portions, and placing a device at every cross junction, including the outer ones. That would require 25 transmitters! In a larger area, it may not seem as bad because these outer transmitters would be within the grid, serving more people. But the outer edges of a grid would always be a problem, unless they could figure out a slightly better formation within. So maybe 20 transmitters for this 1,000 foot square area.

Still, that way above whatever could be said to be possible. It would be cheaper to run fiber. Certainly, maintaining this complex network would be costly.

And it may still not be useful.

Also, every one of those small-cell device's needs a fiber run to back-haul the traffic to the ISP backbone.

That's not true, they can work as a mesh network using each other to move data.

Maybe some of them. At the expense of not having that channel available for end users. No matter how you slice it, 5G's going to need a lot more fiber backhaul than 4G, or you'll lose a lot of the performance advantage. Doesn't matter if you get 1Gb/sec to the house, if the cell node has a bottleneck getting to the isp's backbone.

Typically a small cell will have two separate radios. One will operate on a band for local service, the other is for backhaul and will have LoS to either a cell tower or another small cell on a different frequency (public or private depending on the situation). The use of one won't interfere with the use of the other.

So what's the ratio of customer-facing bandwidth to the cell vs backhaul bandwidth?
I'm assuming that the mesh backhaul bandwidth is grossly oversubscribed compared to the bandwidth going to each of the ~27 customers on each cell?
The way you're describing the backhaul, it sounds like they're not using the newer/faster, but shorter distance 30+Ghz.

I don't work for a telco so I can't speak to their specific implementations, plus it would vary depending on the circumstances (location/geography/available licensed spectrum/etc). I am aware of it because I'm involved in planning in my city and was part of the panel that decided on rules for 5G in our city, and to reach those conclusions we got a pretty good overview of the typical equipment from several cell providers. It didn't hurt that my day job is in tech.

Ah. The overall point I was trying to make is that if they're not doing fiber backhaul to the cell's, then they're likely massively oversubscribing their backhaul bandwidth. At which point it doesn't matter if the link to your house is "5G", because you'll be limited to a lot less than that by the mesh backhaul to the cell tower.

To be honest that's a limitation of any last mile situation, even cable and fiber where you can oversubscribe the backhaul at the utility box on the block. That's really not an issue that's unique to mesh topologies or wireless itself. It's also fairly easy to remediate, just add more at whatever the nearest local tower is, with the short ranges of these access points its unlikely the local box is going to be overloaded (and if so, they are cheap to replace/upgrade).

True, but cable and fiber do their backhaul with fibre. That has a lot more bandwidth than a wireless backhaul.

I think you are missing the point here. If a given AP can service around 20-30 residences and has a dedicated link (wireless or wired) of 1Gbps (same as fiber) it really does not matter. Traffic aggregation by neighborhood is a solved problem and moving the medium from wired to wireless is pretty irrelevant. The telco box in your neighborhood solves for the same issue right now. The cell tower a mesh network links up to is going to be served by fiber as well, its basically the equivalent of that grey metal telco box at the end of your street.

Except the cell mesh is an extra hop to the cell tower's fiber backhaul. And the mesh's wireless link to the cell tower does not have the same kind of bandwidth capacity that a fiber link could. Maybe it'll still be enough that it won't matter, I don't know. But the telco boxes in people's neighborhoods right now have wired backhaul connections to the isp backbone, which likely also have more bandwidth capacity than a wireless connection to the cell tower.

Like I said, I could be wrong about it mattering. To know for sure we'd have to know how fast the 5G link to each customer is, how many customers will be on one cell node, and how much bandwidth is in the cell mesh backchannel, and the mesh's wireless link to the cell tower.

It's not really any more or less of an issue than we see on cable networks, which are typically vastly overprovisioned as well. In all cases its up to the provider to provide enough backhaul, how they choose to do it really does not matter. I'm assuming some will do a good job and others will go cheap, just as has happened with LTE.

So, you think 5G is only going to use the high-band frequencies? Ever hear of carrier aggregation?

We’re specifically talking about what was written about in this article. This part of 5G, the hi band portion, is the one that’s being said to only function in small areas of heavily congested parts of large cities. It’s useless anywhere else. Some of us are saying that it isn’t useful there either. If you’re in a building 5 stories up, forget it. It will have a problem getting that far, and if it does, it can’t get through the walls, or most windows.

All buildings and trees block the service. All human bodies block the service. If you’re walking down the street, and a van goes by, so does the signal. It slows down dramatically just a couple of hundred feet away from the transmitter.

This has nothing to do with band aggregation. That’s an entirely different part of the service, and at a far lower speed.

Yes, to the best of my knowledge, Verizon is 28GHz only for fixed wireless -- at least atm.

Wow this sounds even worse than the ATSC over-the-air HDTV standard. Its near impossible to get a decent signal in populated areas because of mult-path interference -- great boon to the cable companies, leave the low density areas to broadcast, make a killing selling basic service in densely populated areas where over-the-air breaks up all the time.

Yes, ATSC needs much larger antennas with complicated elements, sometimes with amplifiers, to do what a high quality rabbit ear and loop used to do.

This 5G is headed in that direction.

The drop-off in signal strength between 300 and 400 feet suggests something sure happens at that boundary. At 201-300 feet, 90 percent of homes can still receive service, implying that the 10 percent of homes that can't are being blocked by something specific -- maybe something unusual.

From 301-400 feet, we go to basically 50/50. It has me wondering if the relevant metric is that someone's *house* is sitting in the way or something like that. To the best of my understanding, the drop-off in signal strength should be steady. So I wonder if what we're seeing is the general likelihood (or un-likelihood) of maintaining a line-of-site to the transmitter over that distance.

It is most likely the link budget on the signal that causes this. Doubling the distance means EIRP goes down by 4. If whatever modulation requires a very specific link budget and is not very tolerant then such issues happen. We see this all the time with GPS. A standard GPS needs around -135 dBm to lock. Once locked, it can go down to -155 dBm and stay locked. If this 5G won't remain locked, that mean link budget issues to me.

It isn’t just the end of service, it’s also the drop in speed. We see that with every over the air service. It’s particularly bad with this however, because as the coverage is so small, merely walking a few yards will drop the speed by a large amount.

But insofar as end of service goes, it doesn’t seem to me that a viable business case can be made if less than, oh, perhaps 95% of those theoretically within the service range can actually reliably get it. From that chart, that range is as little as 200 feet away, and likely at that range the service can be rather poor. That’s why I believe that one would need to be no more than about 125 feet from a modem.

Of course, rain is a problem, as is snow. This whole concept seems ill conceived.

I'm just wondering if you could fix some of this by using higher vantage points. The reports I link above discuss mounting 5G small cells on street lights and traffic cams. Those don't necessarily have great LoS to nearby houses -- they'll reach to the immediate vicinity, certainly, but not farther.

I know that 28GHz range isn't good but my understanding is that it was still measured in hundreds of yards, not hundreds of feet. I wouldn't expect a signal lock over two miles, but I would've thought 800 - 1500 yards was a lot more reasonable.

Like I said in another comment, 24ghz point to point setups already exist with 10-15+km ranges. I assume these jave such terrible ranges because their mounting them low and with omnidirectional antennas or something. Using them as p2p backhaul seems like the wiser choice.

The proposals here are about every 300'. The main deployment scenarios being proposed by telcos are not for home internet access, they are for heavily network congested areas such as shopping centers, stadiums and business districts.

Sprint offered something like this about 20 years ago, but with focused receiver. Horrible latency. And this wasnt satellite. The distance was just a few miles to the tower. IDK why they couldnt make it work but all I got was excuses. After a month I told them to come get their dish.

Wi-Max was using too high a frequency.

The higher frequencies do not work over the distances needed. You have to get down into the 100s of Megahertz, NOT into the multiple Gigahertz frequencies. That is the reason for the Over The Air television frequencies to be vacated. Provide signals that can penetrate trees and some light wall construction. You have to put some seriously large antennas and decent bidirectional amplifiers on the customer's houses for transmission and reception. The little pretty cute antennas will not be up to the task. At the extreme ranges you have to switch from large flat panel antennas to parabolic antennas that are precisely aimed. I used a parabolic for the 800MHz to 900MHz range coupled with a bidirectional amplifier to connect with a Verizon Tower about 5 miles away through trees to get a coworker internet service. He only stopped using that setup when AT&T built a new tower 1 mile from his house and the normal AT&T CPE worked well.

Didn't the FCC approve some spectrum transfers in the 600-700MHz range explicitly for this purpose?

Yes. T-Mobile bought large chunks of it and they are even paying for some TV stations transition costs to other channels so it will be available sooner.

This should eventually work out well for me at work. We are supposed to get a new distributed antenna system and some new phone channels in the 600 -700 range.

This isn't correct. We use as high as 47 ghz to bounce singals off the moon. Satellite comms routinely use 10ghz, and some go higher.

Sub ghz bands are not a free lunch. Not only do you have significantly less bandwidth but you also run into some unique radio propagation issues (atmospheric ducting, etc).

And yes, those "cute little antennas" *will* work and quite well. A 28ghz sigal has a wavelength of 1.07cm. A 900mhz signal has a 33cm wavelength. That makes for very differently sized antennas. This is why a 24 ghz point to point link can have cereal bowl sized dishes with huge gain but a dish for VHF TV is giant.

The real issue here seems to be height and blockage. If they can get the antennas up high and with clear line of sight to their customers locations they should easily be able to get many miles out of a link.

Rural Monroe County, State of Georgia, USA. There is no way those 5G companies are going to be able to affordably build 50 metre towers for each house so that there is no tree interference. It's a heavily forested area. You have to go to relatively low frequencies to get through the forested areas at a reasonable cost per subscriber when CPE costs are included. Yes, the data transmission rates are going to theoretically be only a tenth of what the higher frequencies might do, but that is better than getting zero transmission due to tree blockage.

There isn't enough bandwidth in those lower frequencies to support home internet or digital television services. If those home owners are buried in the forest than wireless isn't going to be an option for them. The bandwidth is too slim for the providers to offer it un-metered, the latency is much worse than microwave links, etc etc. The 600-700mhz blocks they opened up will be great for helping to extend phone coverage over large areas using fewer cell towers but its not going to be a rural high speed internet link.

Microwave links on the other hand have been used for years now to create even 10Gb links over dozens of miles using very low power outputs. But it does require line of sight.

Most of the high speed cellar networks we use now have the capacity they do and the speeds they do because they use small cell sites with narrow directional antennas. So the same bits of spectrum get to be used over and over again. The longer the reach of a signal the less capacity its going to have.

Basically, physics won't back down. If you can't be reached with a cable and can't get good line of sight to a tower than your not going to get high speed internet either.

I can get great distance with 144mhz signals but with a 12.5khz wide channel im not going to get much bandwidth. Its also only getting significant range because its using a 6 foot tall antenna and a 30-50w transmitter. The entire amateur 2 meter band (144-148mhz) is smaller than a single LTE channel on the 1900mhz band.

The 800 to 900MHz block did okay for internet service with 3G. The 600 to 700MHz block should do okay with true 5G technology at both ends of the wireless link. Five 4K streams to each subscriber? No, I admit that won't be practical. Five 480p streams? Yes, that should be possible. You still get 16x9 in a usable low bitrate way.

Satellite is lousy and way more expensive per GB transferred than 5G will be. Most of the people I know who already use 3G or 4G LTE for home internet ditched satellite. They learn to live with the monthly usage caps and budget data transfers. 5G should improve that experience IF it is in the 600-700 MHz block.

Looks like Elon will win this fight with a fleet of LEO satellites

Or T-Mobile with the 600MHz frequencies they now control, at least in the USA. Elon will get lots of oceanic business.

The constellation of new iridium satellites already take that role. He would be best to cover cities and countries of a high density to initiate a group of subscribers that could support it. Rural areas would be a close second.

Predictions of unproven tech are fairly unreliable.

And you're probably carrying a tiny computer with information about everything in the world at your fingertips. Who would have predicted this working in 1980 or 1990

Thats not the point.

How many predictions of future tech have been made and never became fact? Its easy to say, but you really have no idea whether it will work. It might, but saying it will is really unreliable.

Microprocessors were "proven technology" back then, predictions of how proven tech will evolve is equally unreliable.

Cherry picked example there. And where's my flying car?

Multirotor drone cars are being tested in China already. Maybe we're not on Marty McFly's timeline but it's coming.