Page 2 of 2 FirstFirst 1 2
Results 16 to 26 of 26

Thread: LTE (and 5G) knowledge sharing

  1. #16
    Join Date
    Jul 2009
    Posts
    1,086
    Feedback Score
    0
    I meant specifically band n71 for 5G NR. I was not talking about LTE band 71.

  2. #17
    Join Date
    Mar 2013
    Posts
    3,379
    Feedback Score
    0
    Quote Originally Posted by sheytoon View Post
    I thought I'd add this nice infographic on the different Option 3 variants that I got from this GTI white paper:
    http://www.gtigroup.org/d/file/Resou...359e11f794.pdf

    The dotted line is the control plane to MME, and the solid line is the user plane to SGW (actual user traffic):
    Attachment 159458

    ...
    Thank you for your thoughtful analysis and reply.

    One thing that we know that 5G will be used for is fixed wireless. Verizon is going live with it on mm-wave in a few major metro areas real soon now. The other carriers have announced plans for fixed wireless as well.

    I have seen a lot of wild speculation about what 5G will be able to do for fixed wireless here on HoFo. I don't think many of them have studied information theory - Shannon, Hartley, Nyquist, etc., or RF propagation and engineering. I worked in IT, but worked with a lot of EEs. Most of them knew almost nothing about RF. It is a very niche area of study.

    What I am seeing touted is that 5G will somehow magically transform the cell data landscape into Gbps everywhere, without regard to spectrum bandwidth, channel capacity, RF propagation and physics in general. I can't run the equations anymore, but I believe my gut feeling is correct that that is not possible. mm-wave may work competitively in major metro areas, but T-Mo's large holding of 600 MHz will not. There is not enough spectrum bandwidth at 600 MHz to serve large numbers of fixed wireless customers in an era when video streaming is becoming the norm.

    T-Mo's 600 MHz is large in comparison to others and in comparison to the amount of spectrum available in that frequency range, but it is small in comparison of the spectrum available at 28 GHz, and Verizon and AT&T have most of that. Of course, T-Mo is planning to get some mm-wave to enter the fray. I have heard that they may be delayed in deploying mm-wave by the lack of equipment available for other than 28 GHz.

    As I said before. 5G does not equal mm-wave. And mm-wave does not equal 5G. There is a lot more to it than that. I will add that 5G does not in any way change the theoretical maximum capacity of a comm. channel, RF propagation, or the laws of physics. 5G is not magical. It can only get closer to the theoretical maximums and it still takes a lot of capital investment to deploy any cell tech.

    I tried running the numbers for what level of fixed wireless subscribers the combined Sprint and T-Mo spectrum could support with an average half a TB of data a month at something like 20 Mbps given the peak load hours nature of video streaming. Along with reserving half of their spectrum for regular voice, text and mobile data service. I came up with about 700 subscribers per cell sector. That would only service rather small towns and rural. I don't have a lot of confidence in my estimate.

    Is there a simple estimate of what data throughput that 5G can provide per MHz bandwidth, say without high level QAM and MIMO? Then what would high level QAM and MIMO add?

  3. #18
    Join Date
    Jul 2009
    Posts
    1,086
    Feedback Score
    0
    Quote Originally Posted by bobdevnul View Post
    I have seen a lot of wild speculation about what 5G will be able to do for fixed wireless here on HoFo. I don't think many of them have studied information theory - Shannon, Hartley, Nyquist, etc., or RF propagation and engineering. I worked in IT, but worked with a lot of EEs. Most of them knew almost nothing about RF. It is a very niche area of study.

    What I am seeing touted is that 5G will somehow magically transform the cell data landscape into Gbps everywhere, without regard to spectrum bandwidth, channel capacity, RF propagation and physics in general. I can't run the equations anymore, but I believe my gut feeling is correct that that is not possible. mm-wave may work competitively in major metro areas, but T-Mo's large holding of 600 MHz will not. There is not enough spectrum bandwidth at 600 MHz to serve large numbers of fixed wireless customers in an era when video streaming is becoming the norm.

    T-Mo's 600 MHz is large in comparison to others and in comparison to the amount of spectrum available in that frequency range, but it is small in comparison of the spectrum available at 28 GHz, and Verizon and AT&T have most of that. Of course, T-Mo is planning to get some mm-wave to enter the fray. I have heard that they may be delayed in deploying mm-wave by the lack of equipment available for other than 28 GHz.
    Agree with you there. At the end of the day, the main contributors to high throughput are: bandwidth, MIMO layers and modulation.
    LTE band 71, or NR band n71 represent 600 MHz. The entire band plan is 35 MHz paired, and T-Mobile has on average half of that spectrum nationally:

    Name:  band-plan.png
Views: 82
Size:  39.1 KB

    Name:  t-mobile-600MHzMap-800x450.jpg
Views: 82
Size:  59.9 KB

    So it's meant for basic coverage, not for high speeds.

    As I said before. 5G does not equal mm-wave. And mm-wave does not equal 5G. There is a lot more to it than that. I will add that 5G does not in any way change the theoretical maximum capacity of a comm. channel, RF propagation, or the laws of physics. 5G is not magical. It can only get closer to the theoretical maximums and it still takes a lot of capital investment to deploy any cell tech.
    mmWave is only available to 5G, it will not be deployed on LTE. So I would say it's an important piece of the puzzle for achieving the peak speeds of 5G. The obvious advantage of mmWave is the extremely abundant amount of wideband spectrum. The disadvantage is the terrible propagation characteristics and poor coverage.

    I tried running the numbers for what level of fixed wireless subscribers the combined Sprint and T-Mo spectrum could support with an average half a TB of data a month at something like 20 Mbps given the peak load hours nature of video streaming. Along with reserving half of their spectrum for regular voice, text and mobile data service. I came up with about 700 subscribers per cell sector. That would only service rather small towns and rural. I don't have a lot of confidence in my estimate.

    Is there a simple estimate of what data throughput that 5G can provide per MHz bandwidth, say without high level QAM and MIMO? Then what would high level QAM and MIMO add?
    With LTE it was very simple to predict throughput, because the Transport Block Size (TBS) tables were built into the 3GPP standards. You just have to look up the table and that will be your throughput.

    In 5G NR, it's a complicated formula, but there are some online calculators you can use, like this one: http://www.rfwireless-world.com/calc...alculator.html

    Name:  5G-NR-throughput-formula.jpg
Views: 78
Size:  26.7 KB


    BUT (and this is a big but), 5G NR allows for:
    1) Much higher channel bandwidths: 5 MHz to 100 MHz for FR1, 50 MHz to 400 MHz for FR2 (LTE is 1.4 MHz to 20 MHz).
    2) Flexible subcarrier spacing: 15, 30, 60, 120 kHz (LTE is only 15 kHz).
    3) Larger transmission bandwidth: 106 RBs in 20 MHz channel (LTE is 100 RBs in 20 MHz).
    4) A new type of coding: LDPC for traffic (LTE is Turbo Coding).

    At the end of the day, both LTE and NR use OFDMA to divide the channel for user resources. So they're going to be very similar in performance. If you want to compare apples to apples in terms of same channel bandwidth and same number of MIMO layers, I would say 5G will be approximately 10-15% more efficient than LTE. But once you allow the higher bandwidths, layers, and subcarrier spacing options, 5G peak speeds will be much faster than LTE.

    Fun fact: did you know before 256QAM was introduced for LTE, 3G was just as fast as LTE in terms of peak speeds for a single layer on a 5 MHz channel?
    Last edited by sheytoon; 09-10-2018 at 01:21 PM.
    Want to learn more about how LTE works?
    https://productioncommunity.publicmo...ls/td-p/130581

  4. #19
    Join Date
    Jan 2016
    Posts
    462
    Feedback Score
    0
    Quote Originally Posted by The_Night_Owl View Post
    Am I understanding this right? I know the Samsung Galaxy S9 in the US has three models that support B71 (Unlocked US, T-Mobile, and MetroPCS), but Canadian models don't have B71: https://www.techwalls.com/samsung-ga...r-differences/.

    Is that what you mean?

    I have the Canadian S9. Is it possible there will be a firmware update so I will get B71 in the future (after the 600 MHz auction here)?

    Sent from my SM-G960W using HoFo mobile app
    Canadian model are very limited in bands compared to US unlocked. I wonder why. The lack of band 28 really upsets me.

    Sent from my SM-G960W using HoFo mobile app

  5. #20
    Join Date
    Jul 2009
    Posts
    1,086
    Feedback Score
    0
    Why the interest in B28? It's not used in Canada or US.

  6. #21
    Join Date
    Jan 2016
    Posts
    462
    Feedback Score
    0
    Quote Originally Posted by sheytoon View Post
    Why the interest in B28? It's not used in Canada or US.
    It's used in Australia, SE Asia, South America. Places I travel to.

    Sent from my SM-G960W using HoFo mobile app

  7. #22
    Join Date
    Jul 2017
    Posts
    232
    Feedback Score
    0
    Anyone know why Kingston is the fastest location this year for PCMAG? When I was there a couple months ago there was no 2nd B7 carrier like there is in the GTA...sheytoon you know what would make 780mbps possible over there?


    Sent from my iPhone using Tapatalk

  8. #23
    Join Date
    Nov 2004
    Location
    Vancouver
    Posts
    932
    Device(s)
    Nexus 6P, OnePlus One, Moto E 2nd gen.
    Carrier(s)
    Rogers
    Feedback Score
    0
    I would imagine low traffic would help the most.

    I could easily reach and exceed that with the X20 modem 5CA devices (S9) they used here in Vancouver, as it would have the core 2-4-7-7 (all 4x4), plus choice of B29/B30 for an extra 10MHz.

    The article is on PCMag if anybody is wondering: https://www.pcmag.com/article/363549...s-fastest-city

    Full review of Canada:https://www.pcmag.com/article/363549...ks-canada-2018

  9. #24
    Join Date
    Jul 2009
    Posts
    1,086
    Feedback Score
    0
    It's because Kingston is using 7C 4x4 instead of 7A-7A 2x2. Also, B4 in Kingston is 20 MHz wide I believe.

  10. #25
    Join Date
    Jul 2017
    Posts
    232
    Feedback Score
    0
    Quote Originally Posted by sheytoon View Post
    It's because Kingston is using 7C 4x4 instead of 7A-7A 2x2. Also, B4 in Kingston is 20 MHz wide I believe.
    B4 to 20MHz would have been recent then? Last time I checked it was only Simcoe County that had 20MHz B4.


    Sent from my iPhone using Tapatalk

  11. #26
    Join Date
    Jul 2009
    Posts
    1,086
    Feedback Score
    0
    Sorry, I made a mistake about B4. It's actually still 15 MHz.

Page 2 of 2 FirstFirst 1 2

Similar Threads

  1. New Nascar i736 - I have one and i'm sharing the pics with you people.
    By gastonrey in forum General Nextel Discussion
    Replies: 51
    Last Post: 01-15-2004, 10:04 PM
  2. Replies: 3
    Last Post: 09-17-2003, 05:05 PM
  3. Found this humorous and thought id share
    By ephyouckay in forum Nextel Lounge
    Replies: 1
    Last Post: 09-10-2003, 10:40 PM
  4. HELP! someone with alot of bluetooth and PPC knowledge!
    By trekcomp99 in forum SonyEricsson
    Replies: 3
    Last Post: 07-19-2003, 01:22 PM
  5. free and clear america share plans?
    By Kane3030 in forum Sprint
    Replies: 7
    Last Post: 06-30-2003, 01:02 PM

Bookmarks