No confusion. mmWavelength is where the huge gains in speed will be seen. The rest (using legacy bands) is mainly an incremental upgrade that isn't going to mean jack to most people's use cases. Agree that 5G will not be any more useful than LTE in a Tesla for the foreseeable future.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Does the Model 3 have 3G connectivity or just LTE?
- Thread starter Kabillyhop
- Start date
SammichLover
Banned
Curious, I've not read much on so what kind of peak (and normally seen) bandwidth is expected for devices operating in those bands? At in those frequency bands they're going to be at or below 4G bandwidth rates, right? Or does 5G also have improved efficiency in the spec's design that makes up for the lower theoretical maximum that using a lower radio frequency implies?
<edit> I ask because my understanding has been that the high frequency stuff (above 20GHz) is the payoff for going 5G, and the lower frequencies are just fallbacks there to allow very wide coverage within the same spec umbrella that lets the devices slide back and forth between those trade-offs very quickly and seamlessly. So without any of the above 20GHz frequencies (AKA "mmwave") available for use I didn't think it had much of a point?
<edit2> .....aaaand I've been ninja'd.
Last edited:
I used to have this happen all the time when I drove through the Thorold tunnel where the cell signal is very weak. I would lose LTE and when I came out the other side of the tunnel the signal would never come back.
However after receiving a software update, I think 2019.32.2.2 it started falling back to 3G and has stopped losing the signal through the tunnel.
Don't know if this helps you but I definitely see 3G.
Agree on (coverage constrained) speed gains with mmWave-length, but how much speed does one really need?
Seriously, I get 200Meg up/down on my home internet, and probably use 10% of that during busy hour.
You can get 100-200Megs with MIMO LTE.
Today.
Would you pay more for that 100-200 Megs (vs. 80 Megs down / 38 Megs up I just observed on my LTE speed check) tomorrow?
More yet for 1+Gig?
I'm not sure I will.
Well, maybe a little, just for bragging rights!
I am already paying for way more bandwidth than what I can realistically consume at home, and that 200Megs router supports 15-25 devices, some of them concurrently streaming video.
1+ Gigs down over mmWave (24+GHz) makes awesome commercials, but comes at the price of massive attenuation, as signals is blocked by atmospheric gases, humidity, glass, trees, buildings, etc.
Even with clear line of sight, signal strength degrades fast, and practical range is ~500-1,000 feet.
And you can materially degrade that further by literally farting in the general direction of the cell tower
+30+50% over LTE on the same spectrum.
Re-farming (or dynamically "sharing") 4G spectrum would reuse the existing licensed spectrum that all US carriers own. It's up to each carrier to decide how much of what band to dynamically reallocate to 5G.
5G is slightly more efficient with RAN resource block management, so add ~15% spectrum utilization gains (aka cost savings to all the carriers).
A little cheaper for the carrier, and a bit faster for the end user: so a win-win.
Mid-band (3-6 GHz) might be the sweat spot of "better enough" speed with still "wide enough" coverage.
5G is not restricted to mmWave (24+ GHz) bandwidth.
Traditional spectrum bands are also covered, and benefit from 5G improvements. As does the packet core, which benefits from a lot of new capabilities.
It's just that those mmWave frequencies support "stupid fast" speeds that make for great commercials, so all carriers jumped on deploying those to maximize publicity from the 5G CapEx.
So 99+% of the world now thinks 5G == mmWave.
Which has it's place, but with hard limitations that will prevent mmWave from going nationwide. Ever.
HTH,
a
scottysize
Member
The changes to packet core in 5G are useful because they're supposed to reduce latency significantly. We've got these real-time applications on Teslas like Summon where you have primitive remote control of the car. This is a good application of 5G because you could send "heartbeat" packets to the car with a much tighter tolerance instead of a delay of hundreds of milliseconds.
True, but it's got to work first. The software-defined packet core is still evolving rapidly. From what I've seen, going to be a little while yet ... the fundamental idea is sound, but the practical deployment needs more time to fully bake.
CharleyBC
Active Member
It definitely falls back to 3G when that's all it can get, and shows you on the screen that it has done so.
One thing I've noticed (and it's not surprising) is that even with a moderately good 3G signal strength, it can struggle to keep up with satellite map images. Then when LTE comes back, it speeds up again.
One thing I've noticed (and it's not surprising) is that even with a moderately good 3G signal strength, it can struggle to keep up with satellite map images. Then when LTE comes back, it speeds up again.
Probably not, 5G is pretty horrible at any conditions other than line of site. Even when 5G is wide spread, we probably won't be seeing 500mbps+ when moving or downtown, it's more likely they'll still be in the 100mbps range, which is what LTE can do with the right upgrades. 5G should improve access time so things are closer to real time and on the backend may allow carriers to move more data with the same spectrum and amount of hardware/power demands which of course will improve things, but I doubt you'll see it in parking garages or anything.
hridge2020
Member
Looks like the internet will be up 2020 approx.
SpaceX launches 60 more internet satellites - CNN
Video: SpaceX successfully launched 60 Starlink satellites into orbit as part of Elon Musk's high-speed internet plan
SpaceX launches 60 more internet satellites - CNN
Video: SpaceX successfully launched 60 Starlink satellites into orbit as part of Elon Musk's high-speed internet plan
When Starlink gets up a running, there will be no need for 5G. It's will be "6G".
640k
Member
if the vehicle had a built-in hotspot like most of the legacy makers that have in-vehicle cellular, then yes, i might consider "upgrading" my bandwidth....Would you pay more for that 100-200 Megs (vs. 80 Megs down / 38 Megs up I just observed on my LTE speed check) tomorrow?
More yet for 1+Gig?
I'm not sure I will.
Well, maybe a little, just for bragging rights!
I am already paying for way more bandwidth than what I can realistically consume at home, and that 200Megs router supports 15-25 devices, some of them concurrently streaming video....
if the vehicle had the ability to share its connectivity with a cellular provider (think, bring your own SIM, or use an eSIM), that may be an even better solution.
why does Tesla have to have a proprietary network?
SammichLover
Banned
Starlink is going to be, for the foreseeable future, a rather fixed-location solution. The first antennas are going be "pizza box" sized, so roughly 1/3 the size of the Model 3's roof, and it'll be far more susceptible to [overhead] LOS interference in comparison to cellphones. Not even sure if it's operating software is set up to handle mobility rather than relying on assumption of a fixed location?
It is a very different type of solution.
That's not true.
Again, folks are confusing mmWave frequency bands with 5G. The former is a subset of the latter.
Nothing prevents carriers from deploying 5G over the existing sub-3 Gig registered spectrum, over faster (but shorter range) 3-6 Gig spectrum, or the stupid fast but very short-propagation distance 25+ Gig frequencies.
Over time, all of the above will be utilized, with seamless handover between them.
Agreed.
Quite the opposite.
Device to tower distances with any satellite system are guaranteed to be significantly (10+x) greater than those with any terrestrial cell network.
That's not to say that it wont be usable. It can, and will work.
But the latency will be greater, and speeds slower, than over 5G.
The costs (equipment and monthly recurring) are entirely unknown.
Similar threads
