Hacker News Reader

Comment by AnthonyMouse | Hacker News Reader

Can somebody explain what 5G is actually good for?

With 4G you get fast enough coverage at long distances for basically everything. Maps for sure, but even streaming video as long as not everyone on the tower is trying to do it at once.

With 5G you get more bandwidth, but you're not going to get it in the middle of the woods because it also requires more towers closer together. So it's for high population density areas. But those areas are the places where it's extremely cost effective to install fiber, and cellular will never ever be as fast as fiber, and installing a million towers everywhere is just about as expensive.

If 4G is about all you're going to get in the middle of the woods and fiber (with the last hundred meters over Wi-Fi for mobiles) is the thing you actually want in the city, why is anybody even wasting money installing 5G towers anywhere?

View on HN

5G is kind of three things.

5G on existing frequencies can do more data with the same spectrum. Something something MIMO. This is broadly useful, if you have an 5G capable device and your network updated their base station equipment. 5G and 4G are supposed to be better at sharing spectrum on demand vs 4G and 2/3G.

5G on new, nearby frequency, in the US, there was a second round of taking historically TV spectrum and reallocating it for communications. These bands will most likely need new antennas on towers as well as on handsets. These tend to be lower frequencies, which could be useful for distance and structure penetration, but may not always be built out.

Then there's the millimeter wave stuff. Very high frequency, very fast, great for density, like maybe transit stations and arenas. This is the stuff that needs a tower on every light pole or whatever. I don't think this is going to get a lot of deployment, but it gets a lot of press. I think this is the part the FAA is objecting to for now.

It’ll be a boon for fixed wireless where last mile issues prevent running cable

> Then there's the millimeter wave stuff. [...] I think this is the part the FAA is objecting to for now.

mmWave 5G signals are in the 24-54 GHz range (f = c / 0.001 m ≈ 30 GHz)

This makes me nervous. The O2 peak at 60GHz. What are your opinions on it? If we get that peak of attenuation at 60GHz, what's happening to that energy?

The common wisdom would say it heats up but that's not guaranteed.

https://i.imgur.com/KUR4XmX.png

https://i.imgur.com/iUD2v4D.png

https://www.5gamericas.org/wp-content/uploads/2020/12/InDesi...

What else would happen to the energy? It’ll turn into heat + some re-emissions.

Not to feed the trolls but here is one example:

Tumor Treating Fields are an FDA-approved mechanism for treating certain cancers (mostly gliomas). The exact mechanism isn’t understood, but it seems that prolonged exposure to low power alternating electrical fields in the 100-500kHz region interferes with cell mitosis [1]. I haven’t read about it recently but at the time the speculation was that it was basically disrupting microtubule formation by physically agitating polar molecules involved with the process.

The application from Optune and others uses contact electrodes to drive a small potential gradient though the skull and into the brain.

We know radio frequency also affects polar molecules, mostly to heat up meals. Obviously at multi-GHz carrier frequencies we wouldn’t expect to see this behaivor emerge. However, most digital protocols don’t operate with continuous carriers..the break comms into frames and do things like frequency hopping at much lower periodicity than the carrier. If there was something dumping energy into the cytoplasm on a 2-10 microsecond interval, it could have similar effects.

Do i think 5g is going to cure brain cancer? Probably not, but the body is ridiculously complex and can be surprising. I think we just have to find a way to be able to talk about these concerns fruitfully rather than just try to squash them, because that’s generally where the trouble seems to start.

1 - https://pubmed.ncbi.nlm.nih.gov/33080774/

Thank you. Pretty much the answer I wanted to give, but much better than I could.

And I am not just trying to latch onto the first unknown that's brought up under my comment, these are very similar to the concerns that I had in mind and wanted to bring up.

At the wavelengths we would now be dealing with and exposing ourselves to, there may be - and with some evidence already pointing there probably being surprising and not immediately intuitive (mostly) biological interactions or disruptions to interactions our biology relies on that we are yet unaware or haven't understood leading to phenomenons we haven't considered such as the example you've given.

I tend to be a bit of a fence rider when it comes to these types of things.

Re: 5g we have so much accumulated knowledge and experience dealing with RF that the likelihood that we're going to be caught completely offguard and have no reasonable mitigation seems pretty low. At the same time, it's pretty obvious that our biology didn't have an opportunity to adapt to the types of energy we're putting into the environment and we'd be well served to keep an open mind about what the possible effects could be and how we would test and/or monitor for them.

There are other contemporary topics that suffer this same pathology. Reasonable questions raised in good faith get slapped down regularly and I think we pay a collective price for it. Of course some questions aren't reasonable and certainly not all are raised in good faith, so it's not exactly an easy problem to solve.

I keep wondering if any insects are going to discombobulated by thinking along those lines - but am a complete layman

Affected, absolutely. Unknown to what degree or if it would be any worse then the billions smashed on windshields or fogged out of existence every year.

Interesting paper on using EM simulations on finely detailed insect models to determine net energy absorption and associated properties: https://www.nature.com/articles/s41598-018-22271-3

Keep in mind that the EM field described in that study is quite strong and the transmitter is very close (actually in contact with the patient's skin). They talk about field intensities of 0.6V/cm. That means that if you put the two probes of a multimeter in the air separated by 1cm, it would read 0.6V. That's quite a lot. I'm not sure how to translate that to RF power, but it is many orders of magnitude more intense than what you could ever expect to receive from a mobile phone tower. Radio receivers are basically very fast and very sensitive voltmeters that work with field strengths on the order of microvolts per meter.

Totally agree there's extremely limited risk here from towers. The broader issue to me is that all of the conspiracy talk and bizarre science fiction around the topic of 5G (or vaccines or flat earth) has created a sort of conversational countermeasure where its acceptable to clap back and shut the topic down any time people appear to be wandering in those woods. As a result, some folks with legitimate questions and concerns end up frustrated and looking for answers elsewhere.

Endless patience for off the wall ideas isn't a sustainable option either, so I'm not claiming to have all the answers, I just wish we were maybe a step or two in that direction.

I asked for more details on the implication as the parent quote

> what's happening to that energy? The common wisdom would say it heats up but that's not guaranteed.

As the first and second laws of thermodynamics are well understood and tested theoretical principals that have been experimentally and theoretically verified for ~170 years, there is far more than "the common wisdom" backing the expectation that the system would heat up.

A common conversational tactic to lead the conversation into conspiracy territory is to cast unsubstantiated doubt on a settled topic - then frame the orthodox view as something pushed by an "establishment". A far more interesting discussion emerges when data is presented, or absent data a plausible mechanism for harm.

There is nothing wrong with studying the biological effects of EM fields. We certainly know that small quantities of Benzene, or Arsenic have large impacts on biological organisms - and that biological organisms react to certain portions of the EM spectrum in complex manners e.g. Visible light, and infrared. However we do have some good bounds on how much impact there can be and where it is likely to emerge based on ~120 years of radio spectrum deployments.

"The common wisdom would say it heats up but that's not guaranteed."

If it does something other than turn into heat the laws of thermodynamics are wrong which is somewhat unlikely.

Sure, 60 GHz may be a resonant frequency of O₂ molecules and signals at that frequency may transfer energy to them and thus heat them up.

However, remember that radio signals lose energy to free space proportional to the (distance and frequency) _squared_ [1]. A 60GHz signal is attenuated by 88dB at 10 meters. That means that if it is transmitted with 100 W (20dBW), the RF power that arrives at your position is -68dBW, or 0.000158 milliwatts. That isn't going to heat up much.

[1] https://en.wikipedia.org/wiki/Free-space_path_loss

Note that this math is misleading for calculating how much things will heat up, because effective antenna aperture of the thing that you'd heat up increases with the square of frequency. So the square terms of frequency cancel each other out.

It's still small, but the best simple approximation is just to take the 100W and divide by the proportion of the antenna pattern your target occupies.

A 100W source on 2GHz or 60GHz at 10 meters distance is going to dump about the same amount of energy into you. If the transmit antenna has 3dB of gain, that's going to be about 50mW.

That's interesting. I'm not quite sure how that works but I'll look into it. If you have some resources, I'd appreciate it!

See the derivation section in your link. It's assuming an isotropic antenna at the receiver, whose effective area gets smaller as the frequency increases.

But people do not get smaller as the frequency increases: they occupy the same fraction of the 10m radius sphere around the antenna.

The power levels are low to begin with, so heating up the air is hard to even measure, let alone notice.

It's why 60GHz is license free and used for WiGig, wireless HDMI and short haul microwave links like the ones from Mikrotik.

The heavy oxygen attenuation makes it commercially worthless.

Same reason why there's a license free band on 24GHz too. 24GHz suffers from water vapor absorption and has heavier attenuation than adjecent bands thanks to that. For example Ubiquiti AirFiber uses license free 24GHz.

Stop spreading FUD. ~60ghz radar has been used by cars for for cruise control and parking for many years now at much higher power levels than what phones can transmit. Some phones have mmWave radar for gestures too. We’d use even higher frequencies if it wasn’t for the terahertz gap.

Spreading FUD isn't my intention. If anything I am hoping to alleviate the FUD that open-mindedly reading these documents and plans induce in my mind.

I appreciate and read the honest and detailed technical takes on the topic here.

Especially acknowledging uncertainty and a healthy amount of doubt are pretty essential to the epistemological pursuit. And fear may arise when the uncertainty and doubt are not dispelled with reason.

There's also the issue that with 5G phased array antennae, tx power limits are not restricted as tightly. With a standard single tx antenna, it's easy to calculate the amount of EM energy it can create in the 3D space around the antenna. With phased array beamforming, you're using using multiple small antennae and amplifying the signal beyond the sum of the individual outputs. The safety ratings for EM radiation around phased arrays tend not to be calculated on the theoretical maximum TX power, but on the typical usage. The result is that with low enough level access to the firmware, someone could theoretically weaponize a 5G antenna to produce unsafe levels of RF energy, as well as aim the antenna without physically moving the hardware.

This sounds far-fetched, and it is, but I wouldn't be surprised if we see it done in a lab, like some of the other practically infeasible but bleeding edge cool hacks.

> With phased array beamforming, you're using using multiple small antennae and amplifying the signal beyond the sum of the individual outputs.

No, that's impossible. The RF power imparted onto the EM field by any antenna array is equal to or less than the sum of the output power of each antennae.

I think the argument is that while the total power over the full 4π is equal to the sum of the input power, the constructive/destructive interference cause there to be higher power output within some solid angle. But my understanding is that the limit are for avoiding interference from other devices, not for human safety, and that any reasonable amount of non-ionizing radiation is incredibly safe for humans. Congestion wouldn't be impacted by the beamforming, because on average any increase in intensity at one angle would be countered by a decrease at some other angle.

Higher power density, not higher power. You aren't going to get more watts anywhere than went in, but you will potentially get more watts per cm^2 than at the antenna.

The FAA is objecting to the use of the 3.5GHz band, as it's supposedly too close to the 4400-4700MHz radar altimeter band. This band is the primary one for 5G deployment world wide and interestingly no planes have been dropping or crashing due to it's use. Strange how the same planes landing in European capitals with widespread 3.5GHz 5G deployments seem to do fine there, but will experience dangerous problems on USA soil.

3.5GHz is also exlusively 5G and not used for anything cellular before. Outside of some old wimax deployments here and there, but those ware usually for fixed users not mobiles.

3.5GHz is also firmly in the centimeter bands. :)

Europe is only using 3.4-3.8 GHz. The US is going up to 4.2 GHz.

Ah! That certainly explains some of it. I even checked and sure enough, the highest allocated block here is 3670 - 3800 MHz.

AIUI the European usage of this band only goes up to 3.8GHz (see https://www.everythingrf.com/community/5g-frequency-bands-in... for example), but the FCC auctioned off band space up to 3.98GHz. Aircraft altimeters use the 4.2-4.4GHz band, so the new US allocation nearly halves the band gap. That is what the FAA is worried about.

>halves the band gap. That is what the FAA is worried about.

It's worth noting that a band gap of 200mhz is ridiculously wide, at least according to the top comment from the last thread https://news.ycombinator.com/item?id=29642777

The UK makes available shared access licences in the 3.8 to 4.2 GHz band. These are power limited and location constrained, to manage interference with satellite earth stations.

You can get a license in this band and run your own private 5G network in it. There doesn't appear to be any concern from the UK aviation regulator (the CAA).

Perhaps they're able to achieve sufficient spatial separation from airports and their final approach paths, or perhaps they've found there aren't issues with the reduced transmit power available on these shared access licences? Nonetheless, it will be interesting to see if this 200 MHz guard really is insufficient in the US - at a certain point you need to consider if existing equipment needs to be replaced (like with the swapping of 800 and 700 MHz from digital TV to mobile, and the need for some people to put a band-stop filter onto their TV antenna feed to avoid interference from mobile networks).

Using this band is only an aviation problem if you're in the middle of an approach (approximately 20 miles from a runway, but not a big circle). Hence the location and power limitations.

3.5 is fine, that's CBRS. This is about C-Band, which is 3.7ghz to 4.2ghz in the US.

NTT Docomo operates n79 (4.5GHz) for 5G in Japan. Maybe it's the only network use n79. (It's hard to find supported phones in open market rather than docomo model, except iPhone) https://en.wikipedia.org/wiki/List_of_5G_NR_networks

According to assignment map, it uses 4500-4600MHz. https://www.tele.soumu.go.jp/resource/search/myuse/use/10000...

I want to think it also handled congestion better. I wanna think I've experienced aread (airports? Downtown) where apparent thruput is better than I was used to when places were full of people.

5G means so many things to people it gets muddy.

If you are Verizon or ATT, you have lots of mmwave spectrum, so your interest is in dense urban environments and multi-dwelling units where your speed let's you compete against wifi and cable providers. You spent a bazillion on spectrum and you need to give your customers reasons to pay extra for their service.

If you are t-mobile, you have a ton of mid range spectrum. After all, that was the only reason you bought sprint. You will want to go after consumers with a coverage play, since that is better than mmwave from T and VZ. You might also dip your toe in business services like IoT for farmers and factories.

If you are a hyperscaler, you need CBRS (3.5GHz) to take off. You'll be pitching IoT and edge compute like crazy because the radio coverage is better than wifi, the speeds are good and best of all the spectrum is available at essentially nominal cost.

What you won't get just yet, but will hear about a lot, are the toys like vr robots. The ugly truth is that no one really knows what the killer app for 5g is yet, but every operator is terrified that just like 4G theyll miss the value chain and leave apple or whoever to ride their network and slurp up the profit.

Yet, those legacy operators haven’t invested a bit in an engineering and software organization capable of competing with the apple’s that ride their network.

Id argue that the telco industry tried, and failed, as opposed to not trying at all. ATT for example sent a boatload of people on training from Georgia tech, invested massively in automation... and couldn't pull it off.

I also blame the industry for missing the point on the whole pivot to virtualization. For those of you not in telco, telcos got together in 2012 and said, "how can we be more nimble like the cool kids?" and concluded that virtualizing everything, was the way to go, as if all it took was a vm and a k8s pod. Really it was about culture and supply chain, but that's a whole other story.

They like to outsource everything. Vendor, vendor, vendor. No rich culture and own fundamental IP generation.

5G covers what 4G does as well IIRC. The really high frequency stuff is only what makes the news.

Also the advantage of 5G as far as I'm concerned is being able to connect things (possibly in a mesh?) At very high throughput without cables. If you have an oscilloscope and a signal generator, you should be able to synthesize a (say) network analyzer out of the two of them by linking them via 5G.

> 5G covers what 4G does as well IIRC. The really high frequency stuff is only what makes the news.

Right, right, the really high frequency stuff is what we're talking about. It's confusing that they stick so many things under the same label. The parts of "5G" that are just 4G with better radios and algorithms are useful but also not so contentious. But for that you wouldn't need more towers at closer intervals.

> Also the advantage of 5G as far as I'm concerned is being able to connect things (possibly in a mesh?)

And now we're talking about a third thing.

It could actually be quite useful to open up more unlicensed spectrum for things like this -- if nothing else so we could get better Wi-Fi to hook up to the fiber.

But that's a different thing than having Verizon install a cellular tower on the end of every street corner for high frequency cellular 5G and then not having them come the rest of the way with the fiber into your house, which still seems like some kind of a bad joke.

I think they aren’t wasting money installing towers. You’ll pay for those increased speeds as they build many towers.

Although I agree, what we will end up with is going to be broad area WiFi effectively. Seems like the goal is to replace fiber and then you can move around the city always connected at high speeds.

Comcast is already doing this, their routers will let other Comcast users connect over WiFi.

Personally, I believe that is the end objective. I personally am just getting satellite via starlink and probably will get fiber if it’s available.

> but you're not going to get it in the middle of the woods because it also requires more towers closer together

This is a misconception. You're only thinking about high-band (aka mmWave) 5G. The media has largely focused on this because of the higher speeds and its low range which makes it easy to dunk on with clickbaity articles. But this isn't particularly common outside of sports stadiums or music venues, large train stations, airports (but note that the interference in the article is not in the mmWave spectrum), university campuses, etc.

5G has much better range than 4G at the low-band, and at the mid-band they're much faster at a similar distance. If you're in the US, the only carrier that's been rolling out low-band 5G at scale thus far is T-Mobile.

5G is under 3GPP release 17 now [1] which includes many new features than 4G as release 8 [2]. 5G is much more than just changes in speed.

[1] https://www.3gpp.org/release-17

[2] https://www.3gpp.org/technologies/keywords-acronyms/98-lte

Semi conspiracy alert: 5G is mainly motivated by surveillance through IoT. 4G is still mostly ok as a regular "cellphone" network but isn't cheap enough to be embedded in every device. Many consumer goods now have a wifi option (like my new dishwasher) but it still need to be configured after installation, which many people won't care to do. 5G will make it cheap enough for manufacturers to bundle always-on cellular modems in every device. The cost of networking plan will be assumed by the manufacturer and be offset by the value of telemetry data collected and subscription services made possible by the always-on network. Special dishwasher cycles for an extra 3$ and other shenanigans. Obsolence no longer has to be preprogrammed but can be declared unilaterally at the time most convenient by and for marketing departments.

I'm not seeing it.

I mean yes, corporations are definitely going to attempt that and they can all die in a fire, but telemetry like that requires minimal bandwidth. It doesn't need "5G" for anything.

And beamforming is going to be useful for having precise location as a part of the protocol itself.

Doesn’t the number of antennas make location more precise without beamforming? Doesn’t seem like it has anything to do with location precision.

> 5G will make it cheap enough for manufacturers to bundle always-on cellular modems in every device

How do you figure that? Is there something special about the 5G electronics and chipset manufacturing that make would make it much cheaper than 4G electronics? I don't think this is true.

Maybe scale? Not that long ago shields for Arduino with WiFi were crazy expensive, now esp8622 devices cost so little I don't know how many I have lying around.

Why would 5G be cheaper to implement in IoT devices vs 4G?

> Can somebody explain what 5G is actually good for?

Surveilling your irl location 24/7 with centimeter-precision

https://www.fastcompany.com/90314058/5g-means-youll-have-to-...

https://venturebeat.com/2019/04/29/sk-telecom-will-use-5g-to...

https://www.ericsson.com/en/blog/2020/12/5g-positioning--wha...

Lots of stuff covered. Here’s a few more.

In theory you should be able to take your private network anywhere there’s 5G.

Industrial Ethernet. Latency, jitter, and guaranteed delivery are required in many factories. Meaning they’re still relying on cable, wiring, and proprietary PLC hardware close to the sensors.

Private network slices.

> but even streaming video as long as not everyone on the tower is trying to do it at once

5G Broadcast solves that. Radio, live TV, popular Netflix TV shows etc. could be broadcast and save a lot of bandwidth.

How is that supposed to work when everything is on demand and Netflix alone has thousands of shows?

Not to mention vastly higher power consumption.

In an age of increased awareness around climate change, it's fairly mind-boggling that 5G is being pushed so hard, given the questionable value of the benefits it brings.

How does it increase power consumption?

Source: https://www.ericsson.com/495d5c/assets/local/about-ericsson/...

This is a brochure for Ericsson's lower power innovations, but look at the baseline without their system. Graph on page 2.

>In an age of increased awareness around climate change, it's fairly mind-boggling that 5G is being pushed so hard, given the questionable value of the benefits it brings.

How much % of global carbon emissions is due to cell equipment? Is it a big enough portion to worry about?

Reply on news.ycombinator.com