T-Mobile LTE & Network Discussion V2

[RAvirani]

13 hours ago, chamb said:

How does 5G fix the band 41 upload issue???  I have to see that.  The only way I can fix it is turn off Band 41. Maybe moving to another carrier where there is no band 41 would fix the issue too.   At Sprint, 5G uses band 41 for both 4G and 5G.  So 5G on Sprint probably has the same issue. Poor upload. 

Maybe MIMO on 5G helps, but the phones just can not transmit a strong enough signal on band 41.

TDD-NR supports flex timeslots which LTE does not. In LTE, timeslots were either DL-only or UL-only, and Sprint allocated a majority of them to DL. NR allows timeslots to be dynamically allocated to DL or UL in real time. If upload is in high demand, the entire airlink can be used for uploading, and vice versa for downloading. 

[RAvirani]

57 minutes ago, dkyeager said:

5G could fix the upload issue once it is standalone, since the full channel width will be used for upload (three times faster @ 60MHz plus 5g speed boost). 

NSA ULCA can also fix this. 

[nexgencpu]

14 hours ago, chamb said:

How does 5G fix the band 41 upload issue???  I have to see that.  The only way I can fix it is turn off Band 41. Maybe moving to another carrier where there is no band 41 would fix the issue too.   At Sprint, 5G uses band 41 for both 4G and 5G.  So 5G on Sprint probably has the same issue. Poor upload. 

Maybe MIMO on 5G helps, but the phones just can not transmit a strong enough signal on band 41.

Gotta disagree with you there.. 

5G has given me pretty solid and consistent upload speeds, at the very worse Ill see between 3 and 5mbs with very little signal and 15-30mbs on a strong signal.

[Paynefanbro]

T-Mobile's 600MHz 5G network in action.

 

[dro1984]

Enjoyed hearing about 5G.    Couldn't care less about the car.    

[greenbastard]

On 12/4/2019 at 12:55 AM, nexgencpu said:

Gotta disagree with you there.. 

5G has given me pretty solid and consistent upload speeds, at the very worse Ill see between 3 and 5mbs with very little signal and 15-30mbs on a strong signal.

All those speeds are at 2:30 AM from the same date. Very anecdotal data. 

[nexgencpu]

2 hours ago, greenbastard said:

All those speeds are at 2:30 AM from the same date. Very anecdotal data. 

2:30am or 2:30pm has very little effect on uploads, most of the time only the peak speeds are affected by off peak vs peak times.

They are almost always stable.

http://imgur.com/gallery/B1LjEuW

[greenbastard]

On 12/6/2019 at 5:35 PM, nexgencpu said:

2:30am or 2:30pm has very little effect on uploads, most of the time only the peak speeds are affected by off peak vs peak times.

They are almost always stable.

http://imgur.com/gallery/B1LjEuW

My speeds during the day beg to differ. On T-Mobile, I can get 15-25 Mbps upload during the day.

Past midnight, it peaks over 50 Mbps.

[nexgencpu]

3 hours ago, greenbastard said:

My speeds during the day beg to differ. On T-Mobile, I can get 15-25 Mbps upload during the day.

Past midnight, it peaks over 50 Mbps.

Sprint 5G upload is on 2.5 which realistically is gonna cover less people than most of TMobile's bands. 

My main point is to illustrate how much more stable 5G uploads are against Sprint current TDD config on LTE.

[greenbastard]

15 hours ago, nexgencpu said:

Sprint 5G upload is on 2.5 which realistically is gonna cover less people than most of TMobile's bands. 

My main point is to illustrate how much more stable 5G uploads are against Sprint current TDD config on LTE.

Like I said, anecdotal. All of those speeds tests where done in the span of 10 minutes past midnight, in (which is safe to assume) the same location and with extremely great signal strength/LOS.

The problem with upload speeds in the 2.5 Ghz band is that the UL speeds dramatically fall off whenever you move away from the tower. Good speeds are possible, but on a few select areas near towers.

[Paynefanbro]

54 minutes ago, greenbastard said:

Like I said, anecdotal. All of those speeds tests where done in the span of 10 minutes past midnight, in (which is safe to assume) the same location and with extremely great signal strength/LOS.

But he posted a second screenshot that showed that his upload speeds were consistently around 20Mbps no matter the time of day while his download speeds fluctuated a lot more.

[nexgencpu]

1 hour ago, greenbastard said:

Like I said, anecdotal. All of those speeds tests where done in the span of 10 minutes past midnight, in (which is safe to assume) the same location and with extremely great signal strength/LOS.

The problem with upload speeds in the 2.5 Ghz band is that the UL speeds dramatically fall off whenever you move away from the tower. Good speeds are possible, but on a few select areas near towers.

You obviously didn't see my second screen shot. Also M-MIMO has been great with cell edge performance especially on uploads.

[Paynefanbro]

Sascha Segan's take on T-Mobile's 5G network.

https://www.pcmag.com/news/372478/t-mobile-5g-secrets-revealed-heres-where-it-doesnt-work-w

[Brad The Beast]

So I've noticed that my Speed Test app always seems to show faster speeds than what I'm actually getting. Just now Speed Test showed I was getting 14Mbps down. OpenSignal and Root Metrics were showing 4Mbps and 5Mbps respectively. My Twitter, reddit and Safari apps are running like I'm on a 5Mbps connection. Just this past weekend in Chicago, Speed Test was saying I was getting 5Mbps. OpenSignal and RootMetrics were showing <1Mbps and 1Mbps respectively and my apps were running like they would on a 1Mbps connection. Has anyone else noticed this? I wouldn't be surprised if they were prioritizing speed test connections.  

[nexgencpu]

1 hour ago, Brad The Beast said:

So I've noticed that my Speed Test app always seems to show faster speeds than what I'm actually getting. Just now Speed Test showed I was getting 14Mbps down. OpenSignal and Root Metrics were showing 4Mbps and 5Mbps respectively. My Twitter, reddit and Safari apps are running like I'm on a 5Mbps connection. Just this past weekend in Chicago, Speed Test was saying I was getting 5Mbps. OpenSignal and RootMetrics were showing <1Mbps and 1Mbps respectively and my apps were running like they would on a 1Mbps connection. Has anyone else noticed this? I wouldn't be surprised if they were prioritizing speed test connections.  

No, its because Speedtest has considerably more servers to ping (meaning less hops and closer in proximity) from than any of those other servers you mentioned. Typically, the fastest speed would be considered the most accurate. 

[dkyeager]

When it comes to multithreaded routers with PCs operating faster than 200Mbps, Speedtest dramatically under reports.  DSLreports.com/speedtest is much more accurate in that enviroment.  In the past T-Mobile was caught by the FCC giving higher priority to speed tests iirc. Journalists have taken to random movie download times to test 5g.

[dkyeager]

5 hours ago, Brad The Beast said:

So I've noticed that my Speed Test app always seems to show faster speeds than what I'm actually getting. Just now Speed Test showed I was getting 14Mbps down. OpenSignal and Root Metrics were showing 4Mbps and 5Mbps respectively. My Twitter, reddit and Safari apps are running like I'm on a 5Mbps connection. Just this past weekend in Chicago, Speed Test was saying I was getting 5Mbps. OpenSignal and RootMetrics were showing <1Mbps and 1Mbps respectively and my apps were running like they would on a 1Mbps connection. Has anyone else noticed this? I wouldn't be surprised if they were prioritizing speed test connections.  

The ISPs and Carriers have been setting up Speedtests on their private networks which gives inaccurate internet results. Select an outside speedtest location and the speeds generally always drop. IMO their interconnects are the bottlenecks. A few tests from other ISPs have found faster speeds to the same speedtest sites. Traceroute has shown fairly direct routes in some cases, convoluted routes in other cases.

[Brad The Beast]

10 hours ago, dkyeager said:

The ISPs and Carriers have been setting up Speedtests on their private networks which gives inaccurate internet results. Select an outside speedtest location and the speeds generally always drop. IMO their interconnects are the bottlenecks. A few tests from other ISPs have found faster speeds to the same speedtest sites. Traceroute has shown fairly direct routes in some cases, convoluted routes in other cases.

I always make sure to pick a non-carrier provided server. I usually use the same one every time that provides the fastest results. 

[dkyeager]

3 hours ago, Brad The Beast said:

I always make sure to pick a non-carrier provided server. I usually use the same one every time that provides the fastest results. 

If you are testing the site (or the modem etc for a home/office ISP), then a carrier provided speed test might be best. Multithreaded routers being the exception in many cases.

[RAvirani]

I've been noticing lowband coverage loss in the greater Seattle area as T-Mobile has been swapping out 6-port 700/midband antennas for 8-port 600/700/midband antennas. 

Judging from the antenna gain figures, this isn't surprising, but I was wondering if anyone else has been noticing the same? There are quite a few areas I frequent that have gone from reliable weak L700 to EDGE or no service after 600 antenna upgrades...

Antenna comparison for anyone interested:

• Commscope SBNHH-1D65C (almost all 700 setups use these antennas)
  • 2x2 700 (16.2)
  • 4x4 1700 (17.7) & 1900 (17.9) & 2100 (18.5)
  • 65°
  • 96.6 x 11.9 x 7.1
  • 49.6 lbs
• Commscope FFHH-65B-R3 (most new 600 setups use these antennas)
  • 4x4 600 (14.1) & 700 (14.3)
  • 4x4 1700 (17.6) & 1900 (18.4) & 2100 (19.0)
  • 65°
  •  72 x 25.2 x 9.3
  • 101.4 lbs
• Commscope FFHH-65C-R3 (some 600 setups on large structures use these antennas)
  • 4x4 600 (15.4) & 700 (15.8)
  • 4x4 1700 (17.9) & 1900 (18.4) & 2100 (18.8)
  • 65°
  • 95.9 x 25.2 x 9.3
  • 127.6 lbs

[JThorson]

10 hours ago, RAvirani said:

I've been noticing lowband coverage loss in the greater Seattle area as T-Mobile has been swapping out 6-port 700/midband antennas for 8-port 600/700/midband antennas. 

Judging from the antenna gain figures, this isn't surprising, but I was wondering if anyone else has been noticing the same? There are quite a few areas I frequent that have gone from reliable weak L700 to EDGE or no service after 600 antenna upgrades...

Antenna comparison for anyone interested:

• Commscope SBNHH-1D65C (almost all 700 setups use these antennas)
  • 2x2 700 (16.2)
  • 4x4 1700 (17.7) & 1900 (17.9) & 2100 (18.5)
  • 65°
  • 96.6 x 11.9 x 7.1
  • 49.6 lbs
• Commscope FFHH-65B-R3 (most new 600 setups use these antennas)
  • 4x4 600 (14.1) & 700 (14.3)
  • 4x4 1700 (17.6) & 1900 (18.4) & 2100 (19.0)
  • 65°
  •  72 x 25.2 x 9.3
  • 101.4 lbs
• Commscope FFHH-65C-R3 (some 600 setups on large structures use these antennas)
  • 4x4 600 (15.4) & 700 (15.8)
  • 4x4 1700 (17.9) & 1900 (18.4) & 2100 (18.8)
  • 65°
  • 95.9 x 25.2 x 9.3
  • 127.6 lbs

I have noticed some changes as well around the Seattle region. I am up North near Kenmore and Bothell and been seeing more drops.

[dkyeager]

600 repack is reported as finished.

https://twitter.com/AjitPaiFCC/status/1282720582447124482?s=19

[bigsnake49]

On 7/6/2020 at 12:07 PM, RAvirani said:

I've been noticing lowband coverage loss in the greater Seattle area as T-Mobile has been swapping out 6-port 700/midband antennas for 8-port 600/700/midband antennas. 

Judging from the antenna gain figures, this isn't surprising, but I was wondering if anyone else has been noticing the same? There are quite a few areas I frequent that have gone from reliable weak L700 to EDGE or no service after 600 antenna upgrades...

Antenna comparison for anyone interested:

• Commscope SBNHH-1D65C (almost all 700 setups use these antennas)
  • 2x2 700 (16.2)
  • 4x4 1700 (17.7) & 1900 (17.9) & 2100 (18.5)
  • 65°
  • 96.6 x 11.9 x 7.1
  • 49.6 lbs
• Commscope FFHH-65B-R3 (most new 600 setups use these antennas)
  • 4x4 600 (14.1) & 700 (14.3)
  • 4x4 1700 (17.6) & 1900 (18.4) & 2100 (19.0)
  • 65°
  •  72 x 25.2 x 9.3
  • 101.4 lbs
• Commscope FFHH-65C-R3 (some 600 setups on large structures use these antennas)
  • 4x4 600 (15.4) & 700 (15.8)
  • 4x4 1700 (17.9) & 1900 (18.4) & 2100 (18.8)
  • 65°
  • 95.9 x 25.2 x 9.3
  • 127.6 lbs

Is it because it went from 2x2 to 4x4 on 600 and 700MHz?

[RAvirani]

9 minutes ago, bigsnake49 said:

Is it because it went from 2x2 to 4x4 on 600 and 700MHz?

Not exactly...600 MHz antennas haven't been in production for very long and manufacturers are still refining their designs. Since T-Mobile wanted to deploy their spectrum as quickly as possible, they ordered what was available when they started the rollout. Now we're feeling the pain that comes with that early tech. And unfortunately, the shared transmit path means 700 performance takes a hit as well. 

There are a lot of much higher performing non-600 4x lowband antennas. For example, Verizon has deployed Amphenol HT4C6318x000 antennas (https://amphenol-antennas.com/product/ht4c6318x000/) throughout the greater Seattle area. These antennas get ~18 dBi gain from 696-900 MHz. As a result, there is a tremendous coverage difference between Verizon and T-Mobile setups from the same site - I'll often drop T-Mobile LTE altogether while Verizon is hovering at -110ish or better.

Newer 600 MHz antennas have significantly more robust coverage patterns than the antennas T-Mobile is deploying. For example the Amphenol TWIN6510LU000G-T/TWIN658LU000G-T (https://amphenol-antennas.com/product/twin6510lu000g-t/ and https://amphenol-antennas.com/product/twin658lu000g-t/) get ~17.1dBi and 16.4 dBi gain respectively in the 617-906 MHz range and both support 4xLowband/4xMidband. That being said, I doubt T-Mobile will climb towers to swap out 600 MHz antennas in the near future. As a result, where T-Mobile's density is similar to AT&T and Verizon, as far as raw coverage goes, AT&T and Verizon will have a definite and noticeable edge. 

[red_dog007]

TMobile appears to be shutting down 3G in a few months.  I wonder if this was more an emergency decision.  Possibly seeing speeds tank as they bring in more Sprint customers. It's the fastest way to get more spectrum for LTE.

In my market, HSPA is on PCS.  TMobile only had 13.75MHz on PCS.  Right now they have 5MHz on LTE.  Post merger they now own D and the B block in full.  So they could easily do 15x15MHz now with HSPA gone and fully owning B block.  Could go 20x20 possibly as well.  

In my brothers market, new coverage, they only deployed AWS or PCS which they only had 20MHz total and used 10MHz of it for HSPA for whatever reason.  Now they have 60MHz in just PCS.  To bad they did only deploy sites with just AWS.