The PROBLEM with 5G...

[James]

Linus heads to NYC to try out the reliability of millimeter wave 5G mobile connections. Spoiler alert, it sucks.

 

 

 

[Aprilfooldude]

2 minutes ago, James said:

Linus heads to NYC to try out the reliability of millimeter wave 5G mobile connections. Spoiler alert, it sucks.

 

Millimeter wave APs will have to get so incredibly close together, it's almost impractical to have them in singular locations, instead of a web. The mid and lower band 5G will have to do, as it IS an improvement but will substitute until Millimeter wave can get condensed enough across the globe.

[Granular]

FYI, millimeters are denoted by 2 lowercase 'm's. Two uppercase 'M's would be megamoles or something.

[Enderman]

Ah yes, literally what I predicted years ago when 5G was announced.

Useless reception that gets blocked by everything.

 

Stupid technology tbh, there is no need for hundreds of Mbps on a phone, nobody is downloading huge games or video files like on a PC.

Phones need consitent reliable connection through objects. Even 4G sometimes has dead zones in certain places like parking lots.

 

Instead of increasing frequency they should have increased power output for better reception on 4G or added multiband for higher speeds without higher frequency.

Low speed + good connection > High speed + no connection

[Lurick]

17 minutes ago, Enderman said:

Ah yes, literally what I predicted years ago when 5G was announced.

Useless reception that gets blocked by everything.

 

Stupid technology tbh, there is no need for hundreds of Mbps on a phone, nobody is downloading huge games or video files like on a PC.

Phones need consitent reliable connection through objects. Even 4G sometimes has dead zones in certain places like parking lots.

 

Instead of increasing frequency they should have increased power output for better reception on 4G or added multiband for higher speeds without higher frequency.

Low speed + good connection > High speed + no connection

There is mmWave 5G (high band) but there is low [and mid] band as well which is lower speed but better distance, penetration, etc.. The big problem is everyone is rushing to be first instead of taking the time to do it right and this is the easy path. mmWave is good for tons of clients in a condensed environment while low band and mid band are good for longer range areas or indoors where there are a lot of obstructions.

 

I think mid-band will be the nice mix, if those frequencies are approved for use, since it would be a good balance with low band good for more rural areas.

[WereCatf]

The video ends way too abruptly; Linus just says that TMobile's approach of multiple bands will allow for 5G to even work indoors and.....doesn't mention that, oh, wait, the 5G-signal indoors will not be doing those sexy hundreds of megabits per second - speeds. The way the video ends may give a less knowledgeable viewer the impression that it'll be just as fast because of somethingsomething that TMmobile is doing, which would be entirely false.

 

They really should've taken the extra 60 seconds to say this clearly.

[Enderman]

Just now, Lurick said:

There is mmWave 5G (high band) but there is low and mid band as well which is lower speed but better distance, penetration, etc.. The big problem is everyone is rushing to be first instead of taking the time to do it right and this is the easy path. mmWave is good for tons of clients in a condensed environment while low band and mid band are good for longer range areas or indoors where there are a lot of obstructions.

Low band is 600-700MHz which is good, mid band is over 2GHz which is already going to be horrible compared to 4G penetration, and high band is just stupid.

 

Either way you have to trade off between speed and reception. The 900MHz of 4G was already decent speed with good reception, they should have just added more bands to it to increase speed without losing range.

[Lurick]

Just now, Enderman said:

Low band is 600-700MHz which is good, mid band is over 2GHz which is already going to be horrible compared to 4G penetration, and high band is just stupid.

 

Either way you have to trade off between speed and reception. The 900MHz of 4G was already decent speed with good reception, they should have just added more bands to it to increase speed without losing range.

Security is a big driver for 5G since there are a ton of flaws with 4G's existing implementation.

Mid-band is being pushed into the 6GHz to 12GHz range.

[amdorintel]

21 minutes ago, Lurick said:

Security is a big driver for 5G

hence the worry about china and huweii

[mr moose]

Glad to see so many people invested in understanding the tech and the roll out before making absolute statements about it's viability.  Just because some US company rolls out a shit version  in typical half arsed way doesn't mean the entirety of the tech is a failure.  other countries are successfully rolling out usable 5G.

[ZacoAttaco]

1 hour ago, Enderman said:

Stupid technology tbh, there is no need for hundreds of Mbps on a phone, nobody is downloading huge games or video files like on a PC.

Phones need consitent reliable connection through objects. Even 4G sometimes has dead zones in certain places like parking lots.

I agree that the 5G speeds won't be fully utilised by the general public on phones, they can probably stay on 4G for a while longer and be very satisfied but I think these companies are looking for applications beyond smartphones. The buzzword, IoT has been thrown around a lot in relation to 5G as well as communication between autonomous vehicles.

Although there still seems be much room for improvement with the technology.

 

I kind of left this video thinking, 'yeah let's wait a few years before seriously discussing adopting 5G'. @mr moose mentioned that there are some countries successfully rolling out 5G, I know China and South Korea are leading the way and that the US is lagging behind this generation. Would be interesting to see a similar experiment done in one of those countries.

[Enderman]

7 minutes ago, ZacoAttaco said:

I agree that the 5G speeds won't be fully utilised by the general public on phones, they can probably stay on 4G for a while longer and be very satisfied but I think these companies are looking for applications beyond smartphones. The buzzword, IoT has been thrown around a lot in relation to 5G as well as communication between autonomous vehicles.

Although there still seems be much room for improvement with the technology.

 

I kind of left this video thinking, 'yeah let's wait a few years before seriously discussing adopting 5G'. @mr moose mentioned that there are some countries successfully rolling out 5G, I know China and South Korea are leading the way and that the US is lagging behind this generation. Would be interesting to see a similar experiment done in one of those countries.

IOT doesn't need anywhere near that type of bandwidth. Most IOT devices are fine on kilobit connection, with many simply sending small packets at long time intervals.

And autonomous vehicles on a connection as unreliable as 5g would be a catastrophy.

 

MKBHD did a test of low band 5G and it got better speeds than 4G even though it has a lower frequency (600MHz)

So as per my previous point, medium and high 5G is completely stupid. We need lower frequency and more penetration.

[WereCatf]

1 hour ago, mr moose said:

Glad to see so many people invested in understanding the tech and the roll out before making absolute statements about it's viability.  Just because some US company rolls out a shit version  in typical half arsed way doesn't mean the entirety of the tech is a failure.  other countries are successfully rolling out usable 5G.

Usable in what way? As a replacement for 4G with similar speeds, but better handling of congestion? Sure, I have no problem with that and yes, even a small improvement is still an improvement. I take offense with the hyping of mmWave 5G, not the lower-band stuff. Hell, I'd rather they just skip the mmWave 5G entirely and focused on the lower-band stuff, because the lower-band stuff will actually benefit way, way more people in practice.

[mr moose]

42 minutes ago, WereCatf said:

Usable in what way? As a replacement for 4G with similar speeds, but better handling of congestion? Sure, I have no problem with that and yes, even a small improvement is still an improvement. I take offense with the hyping of mmWave 5G, not the lower-band stuff. Hell, I'd rather they just skip the mmWave 5G entirely and focused on the lower-band stuff, because the lower-band stuff will actually benefit way, way more people in practice.

 

Usable in many ways but most importantly in reducing congestion issues.  People never qualify their complaints and just broadly claim 5G is a stupid technology etc.  The reason people use terms like this is because they don't understand the difference between a US telco rolling it out in a cheap and half arsed speed variant for PR and serious companies investing millions in a properly organized roll out with serious advantages.     Every time a 5g thread pops up we hear the same uneducated rubbish being repeated about speed and how it will never work,  someone in this thread even tried to claim it was just as they predicted 5 years ago then went on to call it stupid while totally ignoring the reality of how it works and the breadth of the technology or why it is important. 

 

 

 

[ZacoAttaco]

2 hours ago, Enderman said:

IOT doesn't need anywhere near that type of bandwidth. Most IOT devices are fine on kilobit connection, with many simply sending small packets at long time intervals.

And autonomous vehicles on a connection as unreliable as 5g would be a catastrophy.

Yeah you make some good points, I’m not sure why they are so closely tied together. 5G and smart city concepts are associated. As with everything in technology there are sales and marketing around a technology. I’m sure there’s some great applications for it but I’m not sure of them yet.

[Enderman]

3 hours ago, gabrielcarvfer said:

It's not stupid at all, you're just seeing the first generation of something and judging as if it was already an established thing. Give it a few years. Fast downlink/uplink is not meant to be used for a single user to download games or 8k video streams, but for sporadic bursts of traffic (a.k.a. web/mpeg-dash/etc) of a ton of users. I guarantee you that the scheduling of radio resources gets much easier if the amount of available resources is far greater than the amount of requested resources.

Phones need consistently reliable connection, that's why better coverage is essencial. Best way to do that? A ton of additional low power cells.

Increasing power is the stupid option, as you reduce the opportunity of frequency reutilization, which is completely insane if you already payed a shit ton of money into spectrum licenses and the infrastructure to deliver your service.

4G/4.9G (LTE-A Pro) and 5G increased the number of component carriers that you can aggregate, but most phones can't deliver that yet, especially if the bands are not contiguous. Where can you get large amounts of cheap and contiguous spectrum? Higher frequencies.

TL;DR: air spectrum is limited and the best way to make good use of it is by utilizing it as much as possible, transmitting with less power and better antennas. You want to blame someone? Blame cabled ISPs that have the luxury of doubling their bandwidth by installing double the cables and refuse to do that.

1) first generation? You think that somehow future generations can break the laws of physics? Higher frequency = lower penetration, period.

 

2) The only solution to this is to have every place covered in towers at line of sight, which would require millions of them just for a small city, and wouldn't solve the problem if you went inside a building or were in a crowd of people. Lower power would even make that worse, as you could see the signal couldn't even go through an umbrella in the video.

 

3) power has nothing to do with what frequency is being used, or reutilization. You might want to reread what you wrote or try explain yourself better.

4G isn't going to magically be removed in favour of 5G, if that's what you meant by reutilization.

 

4) Cheap and contiguous spectrum which you won't be able to use if you're standing under a tree? Excellent.

[Enderman]

17 minutes ago, gabrielcarvfer said:

1) Yes, first generation. Higher frequency = more bandwidth. Requires better coverage to guarantee line-of-sight (LOS) or quasi-LOS.

 

2) That's the point. There should be a small cell in every lamp, indoor and outdoor. Low power is key to efficient use of the spectrum. Don't like it? Convince your carrier to let you pay for the wasted spectrum.

 

3) Yes, it does. Transmission power and carrier (not the telco, but the wave) frequency are the major components that determine how far a signal can propagate. If you transmit with lower power, your signal doesn't travel as far, reducing noise to other signals being transmitted on the same carrier, allowing that frequency to be reused in a nearby cell.

 

Didn't get the 4G/5G point. Not what I meant at all. 5G will probably coexist with 4G in frequencies supported by 4G, just like 2G/3G are still alive and kicking depending on how your country spectrum agency licensed the bands to the mobile carriers. Some agencies require a specific technology for each licensed band blocks; others licenses are technologically neutral, meaning the mobile carrier can repurpose the band if it deems appropriate (e.g. TIM kicked everyone who had 4G capable phones from the 3G infrastructure to repurpose some of the bands to 4G).

 

4) Until your wait for the mobile carriers to fully deploy their new infrastructure, that should happen in the next 5 years.

1) If you go look at MKBHD's test, lower frequency will actually give you better bandwidth in the real world because it penetrates objects better.

 

2) It is stupid to expect a city to put a cell tower in every lamp, that's absurd, they can barely manage to replace old lamps with LED lamps and you think they will want to run optical cables to every streetlamp? Do you know how difficult it is to even get fiber to people's HOMES in most parts of the world?

 

3) The problem with low power is again you would need something line of sight AND super close to you. Even with high power you can see high band 5G was blocked by an umbrella. Again, you can't magically ignore how physics works in the world. Increasing frequency is not the solution to getting good speeds and reliable coverage.

 

4) btw at the rate that technology progresses it is impossible for something like having a 5G access point at every block in a city. The successor to 5G would be out long before full 5G coverage was in place. Then whatever comes after 6G will be out before 6G is in place. Now imagine if the company had to swap out a few million low power access points instead of a few hundred high power cell towers.

[WereCatf]

29 minutes ago, gabrielcarvfer said:

There should be a small cell in every lamp, indoor and outdoor.

Hm. For some reason, I don't see running thousands of fiber-optic cables quite literally everywhere as any sort of a reasonable plan.

[Enderman]

6 minutes ago, gabrielcarvfer said:

1) Not properly/rigorously tested. Real-world results depends on a ton of additional variables (background noise from equipments, temperature, LOS/NLOS, etc). The ability to better penetrate buildings is what justifies the standard to use those lower frequencies as mobility anchors (your mobile carrier track you, your calls, manage connections, etc with lower frequencies and provide quick data bursts with mmWaves).

 

2) Let the carriers pay to use the lamps with their maintainance service. Installing fiber is expensive, but licensing spectrum is far more. Pretty sure they would prefer to lay down fiber under sidewalks and connect a few lamps than pay for more spectrum just to cover a few densily populated areas that require that infrastructure.

 

3) For mmWave, yup, LOS or quasi-LOS. Good for speed, not that good for reliability, but realistically speaking, those are different markets that pay different prices (eMBB vs URLLC). Just wait for the infrastructure to be deployed before complaining it doesn't works. Even LTE-A Pro haven't been fully deployed in most markets.

 

4) I wouldn't be so sure of that. 6G is already in the works, but the focus is on the backbone and transmission power (seems like OFDMA will be switched to GFDMA or something like that). Now imagine if those million lower power cells are actually built on top of a reconfigurable FPGA, requiring a firmware update to work with the latest standards? AFAIK, that already is, at least partially, the case for Ericsson and Nokia deployments.

1) It was a long distance test, the way 99% of people use cell phones, not standing beside a tower.

 

2) carriers don't even want to upgrade from copper to fiber half the time. You are severely underestimating how greedy they are and how little they care about actually improving technology for their users.

 

3) Linus and many other people are already showing how badly it works.

 

4) Amplifiers and antennas are physical, they need to be replaced, even if you can reprogram your FPGA or change the firmware.

There is nothing that will 'fix' high frequencies that suck at penetration.

[Enderman]

11 minutes ago, gabrielcarvfer said:

1) Because all previous mobile standards focused on range.

 

2) Lower margin markets. Mobile market is far more lucrative.

 

3) By not properly testing the technology and focusing on spotty coverage of the initial deployment. People forget that there are laws and regulations that prevent the installation of new cells, that people freak out and say they will surely get cancer and make politicians ban stuff, etc.

 

4) Easier and cheaper to replace than the entire radio. They suck at penetration, that is a fact. You can work around it or keep using whatever old standard the mobile carrier still supports.

1) yeah because that's how cellphones are supposed to work, at long range, if we wanted short range we would use wifi

Oh wait, some places already tried putting wifi accesspoints everywhere for full coverage, and even that fell short.

And wifi still has far better range and penetration than high band 5G!

 

2) that's funny because like half the world doesn't even have proper 4G coverage yet. A cell company doesn't make extra money by upgrading the speed for its users anyway. It's just an additional infrastructure expense.

 

3) great, you just made another point for why it would be nearly impossible to put a cell tower in every streetlamp.

 

4) It's not cheap to replace an antenna on a million different devices scattered around a city. It's much cheaper to replace a few hundred big antennas on main cell towers.

 

So basically the argument is spend a ton of money to put millions of access points line of sight everywhere for higher speeds (which nobody really needs anyway) at no profit to the provider, or spend less money and only have a weak connection that drops out when you open an umprella. Yup, sounds about right, medium/high band 5G is useless.

 

The low band in marques' test actually seemed promising, that would be nice to have over 4G.

[realpetertdm]

Just a heads up: This video talks about 5G mmwave, which uses very high frequency bands (usually 24+ GHz). mmwave is a PART of 5G, not 5G itself. 5G implementation as of right now is more focused on lower frequency 5G (usually sub-6 GHz). 

 

So don't go around saying "5G's shite" or something, that's misleading. 5G's still alive and expanding.

[captain_to_fire]

10 hours ago, Enderman said:

So as per my previous point, medium and high 5G is completely stupid. We need lower frequency and more penetration.

Any wireless carrier cannot always rely on sub 1GHz spectrum because it doesn't have enough bandwidth even though it has good building penetration and coverage. That's why in late 2013, Verizon admitted that its 700 MHz 4G LTE has congestion problems. To solve their congestion problems, T-Mobile has to give AWS (1700 MHz) and PCS spectrum (1900 MHz) to Verizon after T-Mo bought unused 700 MHz spectrum for $2.4B.

Basically, a wireless carrier needs to have good coverage and enough bandwidth for their users and sticking to sub-1 GHz waves will only result to congestion.

[greenhorn]

I just want to add/comment on a few points menioned in this discussion:

It is true that 5 not only focusses on "high" frequencies in the mmWave spectrum but also on frequency ranges that are already in use. As many have already mentioned, higher frequency signals experience higher pathlosses (signal attenuation over distance) but also allow for higher bandwidths (meaning there are more transmission ressources, allowing fpr higher speeds and/or more users to be served simulataneously). With that in mind, high-frequency deployment makes more sense in densely populated, urban areas while sparsely populated, rural areas are better served with low-frequency systems.
While the received signal's strength is important for coverage, this is not the only important factor. Typically, one looks at the SINR (Signal to Interference and Noise Ratio). The signal can only be used if not overwhelmed by noise and/or interference. The high propagation losses of high-frequency signals also reduce the strength of the interfering signals. Of course signal strength still is a problem, that's why so much work is dedicated to beamforming. This process involves antenna arrays that are operated in such a way that they generate a beam towards the intended recipient, thus making sure he receives a stronger signal while also reducing the interference this transmission causes to other users.
As to the frequency band that are being looked at for 5G, this depends on the region. Frequency band allocation is not the same everywhere, available band are being auctioned off with operators willing to pay billions for that. I have noticed that the 28 GHz band for example has received a lot of attention.

For those of you who want to have a little bit of further reading, especially on propagation losses of mmWave signals as well as on penetration losses (losses coming from the outdoor to indoor transition of a signal depending on wall materials), I can recommend the relevant sections in the 3GPP TR38.901 (which can be downloaded for free). Most formulas are quite simple actually and good enough to get an idea.

[Timmyfox]

If I'm not mistaken, this should get better with time as adoption increases.

 

See, to my understanding, the mmW 5G band is not so much designed to be very ideal like this, with only a single (or a select few) devices actually directly connected to a hotspot, but rather intends to use MU-MIMO technology and have each 5G device within proximity create a type of mesh network with each other to increase the range and signal strength. Each compatible device will essentially function as its own 5G antenna and later on, in a couple years when adoption is much wider, there should be enough 5G devices out and about in most crowded areas to better distribute the signals instead of relying exclusively on the limited number of hotspots (and having direct line of sight to them).

 

Of course, I don't know much about this myself so I might be completely wrong here but that's at least how I understood it during a conference thingy I attended last year at Ericsson's office here in Linköping, Sweden.

[Enderman]

1 hour ago, gabrielcarvfer said:

1) Phones are supposed to work at long range, that's why you have the mobility anchor/control traffic at lower frequencies, but you want to offload most data transmissions to higher frequency bands. It's basically the same principle from LTE-LAA, but LAA uses unlicensed bands and competes with WiFi/Bluetooth/etc, while on mmWave you have exclusivity. The major ISPs (Oi, Claro, Vivo) have a fairly decent WiFi coverage here, but they're not placed correctly for mmWave applications.

2) What exactly is your definition of "half the world"? Half of the surface area? Economically unviable. Half of the population? Viable at reasonably dense urban centers (not necessarily cities). They don't want to improve your service for nothing, they do it for money. Better plans, higher efficiency, diverting expensive spectrum to clients that are willing to pay more, etc.

3) Just saying that you can't complain that the coverage is spotty ignoring it's bad because other people refused to let carriers improve the coverage. We have a huge antenna deficit in Brazil (estimated to be around 100k) because each and every city can legislate on where, when and how many antennas can be installed...

4) Depends on how you plan to do it, how cheap it is, how it was deployed, etc. The underground one seems to be the most promising, and your already need to give maintainance to sidewalks from time to time and it was basically a huge cable with antennas every few meters. Baseband processing can also be done in the cloud, so the cells can be as cheap as possible. Few hundred? Few hundred thousands. That requires specialized technicians to install and maintain, when the underground dumb cells can be deployed similarly to a fiber cable.

The argument is that there are better use cases for low frequency bands and you will be ejected from them to make space for other applications that are more profitable or that actually require them to work reliably. That's the trend 3GPP, telcos and researchers (including me) are following.

 

1) well idk where you live, but even major cities here in canada aren't covered in wifi, and will never be covered in 5G either.

 

2) half the population. Also when ISPs upgrade their speeds they don't just add more expensive plans, because people already are paying more than they want to. For example here telus replaced their previous copper cable for fiber and kept the pricing almost the same even though the speeds increased. Why? Because people paying $100+ for an internet plan aren't going to want to pay $200+ just because better speeds are available.

 

3) It's spotty because high frequency has bad penetration, and putting millions of antennas so that every corner of a city is LOS is impractical.

 

4) Dude an umbrella blocks the signal, what do you think happens if it's underground, lol