Official Tmobile-Sprint merger discussion thread

[Brad The Beast]

Just now, RAvirani said:

Possibly. All of this is just speculation, of course. 

Ok. Good to know. Hopefully Dr. Saw can do the AMA as he should be able to comment on how many new sites have been built, upgrades, etc.

[RAvirani]

11 minutes ago, Brad The Beast said:

Ok. Good to know. Hopefully Dr. Saw can do the AMA as he should be able to comment on how many new sites have been built, upgrades, etc.

Yes, that would be awesome. 

I'd like to ask him a few questions about future equipment, too. 

[Brad The Beast]

Just now, RAvirani said:

Yes, that would be awesome. 

I'd like to ask him a few questions about future equipment, too. 

I'm awaiting a response at the moment. The nice lady I'm working with is trying to see if he's available as he's very busy these days.

[red_dog007]

On 9/4/2019 at 2:21 AM, RAvirani said:

 

3GPP release 10.3. 

2180-2200 DL/2000-2020 UL. It's not terribly publicized because it was never used/deployed. Band 70 is a hodgepodge of Dish's spectrum holdings that was created due to Sprint's (and possibly a few other companies' although I can’t remember) complaints that the PCS H block would interfere with other PCS operations. 

Why would they want to do Band 23 though? They can use 1995-2020 as DL.  Do B70+B66.

 

[RAvirani]

28 minutes ago, red_dog007 said:

Why would they want to do Band 23 though? They can use 1995-2020 as DL.  Do B70+B66.

 

Band 70 is a hodgepodge of three different bands:

• 15 MHz of AWS-3 (1695-1710)

• 10 MHz PCS H Block (1915-1920, 1995-2000)

• 40 MHz of AWS-4 (2000-2020, 2180-2200)

The downlink is PCS H DL+AWS-4 UL and the uplink is AWS-3. This effectively wastes 25 MHz of spectrum (PCS H UL and AWS-4 DL). 

I would much prefer 20x20 B23 and TDD or UL-only AWS-3. If PCS H block NR can be deployed alongside PCS G and possibly PCS C with DSS, I'd welcome that too, although I'm not sure if that would be possible due to interference concerns (PCS H DL likely cannot sit directly next AWS-4 UL). But, even if the PCS H block was abandoned, it's better to waste 10 MHz than 25 MHz. 

[red_dog007]

On 9/4/2019 at 2:30 AM, RAvirani said:

Dish has very clearly expressed that they want to do something with their spectrum. Buildout deadlines have passed and been extended. They likely will not be extended again.

T-Mobile isn't as interested in smaller chunks of low/midband spectrum as they already have enough. What they need is large swaths of highband for NR. This is probably their biggest motivation for the merger. Dish can’t help them there.

AT&T has more spectrum than they can put on air, and isn't looking for more. They have B14 and B66 that hasn't been deployed across half, if not more of the nation.

Verizon has repeatedly stated that they are perfectly happy with their current low/midband spectrum holdings. If they wanted to bid on 600 MHz spectrum, they could and would have. 

This leaves Sprint, who I think could benefit a lot from Dish's spectrum holdings. 

Dish paid a premium for PCS H to extended the buildout deadlines of 700/AWS4.  Money talks.  Might be able to come to another agreement, especially if Dish is putting forth good faith and getting their protection site network built out.

TMobile would like more spectrum sure, but I think it is so much more than that.  If VZW is comfortable with their position and twice the customers, TMobile should be able to do well.  No other carrier than Sprit has huge chunk of midband for NR, especially spectrum that doesn't need to be refarmed.  Sprint is doing just 40MHz in places, of which Dish has. Carriers are looking at 3.5GHz, but depending who bids and who gets what, there isn't really a lot of spectrum there to be had.  Just that it will likely be cheap.  

ATT doesn't have to put all their spectrum on air across the nation.  TMobile isn't even doing it for new coverage areas.  ATT even said that B14 was going to be targeted deployments and only added where the capacity was needed.  

[red_dog007]

18 minutes ago, RAvirani said:

Band 70 is a hodgepodge of three different bands:

• 15 MHz of AWS-3 (1695-1710)

• 10 MHz PCS H Block (1915-1920, 1995-2000)

• 40 MHz of AWS-4 (2000-2020, 2180-2200)

The downlink is PCS H DL+AWS-4 UL and the uplink is AWS-3. This effectively wastes 25 MHz of spectrum (PCS H UL and AWS-4 DL). 

I would much prefer 20x20 B23 and TDD or UL-only AWS-3. If PCS H block NR can be deployed alongside PCS G and possibly PCS C with DSS, I'd welcome that too, although I'm not sure if that would be possible due to interference concerns. PCS H DL cannot sit directly next AWS-4 UL. But, even if the PCS H block was abandoned, it's better to waste 10 MHz than 25 MHz. 

B70 25x15 (using all 1995-2020) and then CA that with B66 2180-2200.  Don't need an UL carrier on the second or higher carrier.  

Then you just waste 5MHz on the PCS-H UL.  

[Tengen31]

Dish paid a premium for PCS H to extended the buildout deadlines of 700/AWS4.  Money talks.  Might be able to come to another agreement, especially if Dish is putting forth good faith and getting their protection site network built out.
TMobile would like more spectrum sure, but I think it is so much more than that.  If VZW is comfortable with their position and twice the customers, TMobile should be able to do well.  No other carrier than Sprit has huge chunk of midband for NR, especially spectrum that doesn't need to be refarmed.  Sprint is doing just 40MHz in places, of which Dish has. Carriers are looking at 3.5GHz, but depending who bids and who gets what, there isn't really a lot of spectrum there to be had.  Just that it will likely be cheap.  
ATT doesn't have to put all their spectrum on air across the nation.  TMobile isn't even doing it for new coverage areas.  ATT even said that B14 was going to be targeted deployments and only added where the capacity was needed.  

Vzw is already deploying 3.5 equipment so they want it

Sent from my SM-G965U1 using Tapatalk

[RAvirani]

16 minutes ago, red_dog007 said:

B70 25x15 (using all 1995-2020) and then CA that with B66 2180-2200.  Don't need an UL carrier on the second or higher carrier.  

Then you just waste 5MHz on the PCS-H UL.  

That is also possible, but that will just exacerbate Sprint's poor upload problem. You would be adding 45x15 to Sprint's spectrum portfolio. That's a 3:1 ratio, but it will perform closer you a 6:1 or 7:1 ratio speed-wise (FDD UL tops out at about half FDD DL, and this problem will be exaggerated by higher order downlink MIMO). This is the problem with L2500 all over again. I'm not saying a 15 MHz UL carrier is anything to frown at, but with 2500 already favoring downloads, I'd much prefer a symmetrical setup.

B70+B66 also closes the door on PCS H+G+C DSS which may be possible (someone more familiar with power limit regulations/interference potential can likely chime in with a definitive answer here). 

[dkyeager]

Anyone going through this merger process (new T-Mobile, Dish, etc) will use it as an opportunity to raise more money.  Much harder to get money for organic growth.  Some of Dish's delay this summer was because they were arranging financing iirc.

Another possibility in construction slowdown is the shortage of equipment (except Samsung).  In Sprint's case they have overbuilt compared to what they have brought online. Not certain about other carriers in this case.

[red_dog007]

On 9/6/2019 at 3:16 PM, RAvirani said:

That is also possible, but that will just exacerbate Sprint's poor upload problem. You would be adding 45x15 to Sprint's spectrum portfolio. That's a 3:1 ratio, but it will perform closer you a 6:1 or 7:1 ratio speed-wise (FDD UL tops out at about half FDD DL, and this problem will be exaggerated by higher order downlink MIMO). This is the problem with L2500 all over again. I'm not saying a 15 MHz UL carrier is anything to frown at, but with 2500 already favoring downloads, I'd much prefer a symmetrical setup.

B70+B66 also closes the door on PCS H+G+C DSS which may be possible (someone more familiar with power limit regulations/interference potential can likely chime in with a definitive answer here). 

I do not see a need to be concerned about upload.  15MHz is loads of upload.  I work for an ISP and it's always between 8:1 and 10:1 ratio when every subscriber has symmetrical internet.

If upload was an issue we would see Sprint's UL slammed on B41 right now. The poor speeds come from not having the best of signal.  But even at the last superbowl, Sprint doing speedtests around the stadium with solid signal they were still pulling in 2~3Mbps.  When real world best you can get is generally just 9Mbps.  If UL is an issue you'd expect UL to be practically 0.

[RAvirani]

13 hours ago, red_dog007 said:

If upload was an issue we would see Sprint's UL slammed on B41 right now. 

Sprint's UL on L2500 is slammed right now. How else could they have the fastest download speeds and the lowest speed score in many cities across the country according to RootMetrics?

Speeds like these are unfortunately common on L2500:

Don't forget, a 9 megabit pipe can be fully consumed by a few people facetiming or live-streaming. And those kind of things are becoming more and more popular.

13 hours ago, red_dog007 said:

But even at the last superbowl, Sprint doing speedtests around the stadium with solid signal they were still pulling in 2~3Mbps.  When real world best you can get is generally just 9Mbps.  

The Super Bowl is probably the worst example you could have picked to make this point. Sprint had a DAS and several multibeam antennas deployed across the stadium and surrounding areas. This type of high-capacity setup is most definitely the exception, not the norm. 

How often do you speedtest lower download than upload on Sprint? I'd guess almost never. But how often is the downlink stable while the uplink is unusable? At least in my experience traveling across the country, its's quite often. Most large studies/online reviews tend to agree as well.

If you already have excess downlink capacity when in mid/high band coverage areas while the uplink is struggling, why ignore the problem when you have a chance to solve it?

[RedSpark]

On 9/7/2019 at 2:10 PM, red_dog007 said:

I do not see a need to be concerned about upload.  15MHz is loads of upload.  I work for an ISP and it's always between 8:1 and 10:1 ratio when every subscriber has symmetrical internet.

If upload was an issue we would see Sprint's UL slammed on B41 right now. The poor speeds come from not having the best of signal.  But even at the last superbowl, Sprint doing speedtests around the stadium with solid signal they were still pulling in 2~3Mbps.  When real world best you can get is generally just 9Mbps.  If UL is an issue you'd expect UL to be practically 0.

Slow upload speeds can render a device unusable, whether for web browsing or even sending emails with attachments.

Slow upload speeds are just as bothersome, if not more so than slow download speeds, in my opinion.

[RedSpark]

T-Mobile workers seek assurances from DT CEO about merger-related concerns

Everybody's on their toes here....

[RedSpark]

NY lawsuit alleges T-Mobile sold used phones, claiming they were new

Quote

 

A New York City agency has filed a lawsuit against T-Mobile and some of its authorized Metro by T-Mobile dealers, alleging that their stores sold used phones, claiming that they were new, among more than 2,200 violations.

The complaint filed in the state supreme court in Manhattan said T-Mobile’s "pervasive" illegal activity spanned 56 Metro stores in all five boroughs, including authorized dealers and stores run by its MetroPCS NY unit, according to Reuters.

 

The lawsuit alleges that "abusive sales tactics are rampant at Metro stores"..... Yikes!

Sounds like a housecleaning is coming by Corporate. I imagine they don't want anything interfering with this merger.

[belusnecropolis]

Honk Honk 🤡🗺️

35 minutes ago, RedSpark said:

NY lawsuit alleges T-Mobile sold used phones, claiming they were new

The lawsuit alleges that "abusive sales tactics are rampant at Metro stores"..... Yikes!

Sounds like a housecleaning is coming by Corporate. I imagine they don't want anything interfering with this merger.

 

38 minutes ago, RedSpark said:

T-Mobile workers seek assurances from DT CEO about merger-related concerns

Everybody's on their toes here....

 

39 minutes ago, RedSpark said:

Slow upload speeds can render a device unusable, whether for web browsing or even sending emails with attachments.

Slow upload speeds are just as bothersome, if not more so than slow download speeds, in my opinion.

 

[red_dog007]

On 9/8/2019 at 7:05 AM, RAvirani said:

Sprint's UL on L2500 is slammed right now. How else could they have the fastest download speeds and the lowest speed score in many cities across the country according to RootMetrics?

Speeds like these are unfortunately common on L2500:

Don't forget, a 9 megabit pipe can be fully consumed by a few people facetiming or live-streaming. And those kind of things are becoming more and more popular.

The Super Bowl is probably the worst example you could have picked to make this point. Sprint had a DAS and several multibeam antennas deployed across the stadium and surrounding areas. This type of high-capacity setup is most definitely the exception, not the norm. 

How often do you speedtest lower download than upload on Sprint? I'd guess almost never. But how often is the downlink stable while the uplink is unusable? At least in my experience traveling across the country, its's quite often. Most large studies/online reviews tend to agree as well.

If you already have excess downlink capacity when in mid/high band coverage areas while the uplink is struggling, why ignore the problem when you have a chance to solve it?

Superbowl is a great example.  No matter how much technology they deploy, they are at a huge UL disadvantage reguardless. If just a few people can max out a 9Mbps UL, then 65,000 people should easily be able to kill Sprint's UL capacity.  This did not happen. 

What I find in my testing is B41 really needs a very good signal to pull in at least ok uplink speeds.  I might be able to pull in 30~70Mbps DL but my signal isn't really that great so the UL suffers hardcore.  A lot of times I'll just get 0.01Mbps on Speedtest.  But if I can get close to the tower and get a nice clear signal I will be at 4~9Mbps.  So my original 0.01Mbps is not a result of congestion. 

It's the same effect of say a 5MHz DL channel.  Close to the tower you get 22Mbps DL, back up some and now you get 2~5Mbps.  Same thing is happening on B41 but in the other direction and more pronounced due to lack of UL air time. 

 

Has Sprint even enabled 2xCA and 3xCA UL yet?

Edited September 9, 2019 by red_dog007

[Tengen31]

Superbowl is a great example.  No matter how much technology they deploy, they are at a huge UL disadvantage reguardless. If just a few people can max out a 9Mbps UL, then 65,000 people should easily be able to kill Sprint's UL capacity.  This did not happen. 
What I find in my testing is B41 really needs a very good signal to pull in at least ok uplink speeds.  I might be able to pull in 30~70Mbps DL but my signal isn't really that great so the UL suffers hardcore.  A lot of times I'll just get 0.01Mbps on Speedtest.  But if I can get close to the tower and get a nice clear signal I will be at 4~9Mbps.  So my original 0.01Mbps is not a result of congestion. 
It's the same effect of say a 5MHz DL channel.  Close to the tower you get 22MBps DL, back up some and now you get 2~5Mbps.  Same thing is happening on B41 but in the other direction and more pronounced due to lack of UL air time. 
 
Has Sprint even enabled 2xCA and 3xCA UL yet?

2x has been a thing for a while now. 3x up no

Sent from my SM-G965U1 using Tapatalk

[RAvirani]

28 minutes ago, red_dog007 said:

Superbowl is a great example.  No matter how much technology they deploy, they are at a huge UL disadvantage reguardless. If just a few people can max out a 9Mbps UL, then 65,000 people should easily be able to kill Sprint's UL capacity.  This did not happen. 

If my memory serves me correctly, Sprint deployed seven 9-beam fast-roll-off antennas at the stadium. This is 126 (9 beams * 2 tx/rx rows per antenna * 7 antennas) times the capacity of an average site and comes without the interference/SNR concerns we have in the real world. Don't forget that capacity was supplemented by other basic COWs and macros which covered the parking lots and other areas around the stadium. The super bowl is absolutely NOT a model for Sprint's performance across the rest of the network. 

28 minutes ago, red_dog007 said:

What I find in my testing is B41 really needs a very good signal to pull in at least ok uplink speeds.  I might be able to pull in 30~70Mbps DL but my signal isn't really that great so the UL suffers hardcore.  A lot of times I'll just get 0.01Mbps on Speedtest.  

This is exactly the point I'm making. Sprint needs to address upload speed problems, not download speed problems. 45x15 doesn't do that. 20x20+15 MHz TDD or even 20x35 would be a much better addition to Sprint's spectrum portfolio. 

Imagine Sprint's current DSS setup (40-60 MHz N2500 + 3x20 MHz L2500) aggregated with an additional 35 MHz of NR uplink. Now that would be a hell of a network. 

28 minutes ago, red_dog007 said:

Has Sprint even enabled 2xCA and 3xCA UL yet?

2x is live nationwide. 3x is not deployed, but there aren't any devices that support it, so it doesn't really matter. 

[red_dog007]

19 hours ago, RAvirani said:

If my memory serves me correctly, Sprint deployed seven 9-beam fast-roll-off antennas at the stadium. This is 126 (9 beams * 2 tx/rx rows per antenna * 7 antennas) times the capacity of an average site and comes without the interference/SNR concerns we have in the real world. Don't forget that capacity was supplemented by other basic COWs and macros which covered the parking lots and other areas around the stadium. The super bowl is absolutely NOT a model for Sprint's performance across the rest of the network. 

This is exactly the point I'm making. Sprint needs to address upload speed problems, not download speed problems. 45x15 doesn't do that. 20x20+15 MHz TDD or even 20x35 would be a much better addition to Sprint's spectrum portfolio. 

Imagine Sprint's current DSS setup (40-60 MHz N2500 + 3x20 MHz L2500) aggregated with an additional 35 MHz of NR uplink. Now that would be a hell of a network. 

2x is live nationwide. 3x is not deployed, but there aren't any devices that support it, so it doesn't really matter. 

The Snapdragon 855 has the X24 modem which does support 3xCA UL.  Are the phones not coming out of the FCC with full support?

The speed problems are solved just by having a larger dedicated channel at lower frequency for supplemental UL.  Having 15MHz at 1695-1710 solves that. Maybe could even use it to increase the range of B41/N41?

Im not saying 35MHz UL would be bad or anything or not for 35MHz UL.  But people don't UL as much as they think they do.  Facetime, Duo, Skype, etc don't use as much data as people think they do.  Sprint's UL is not at capacity. This is why UL on the other carriers FDD has plenty of speed behind it, even with the DL is slammed.  The low speeds are a connection/time issue which on Sprint gives huge inconsistencies.  15MHz easily solves the low speeds, gives plenty of additional capacity and fixes consistency issues within B41/N41 coverage area. 

But considering the DL capacity that is now available on these new massive mimo sites... what is another 20MHz. 

But if Dish keeps this stuff to build out their network, they totally want all of AWS4 for DL. 

Im working here with bands that already exist.  AWS3 UL only exists in B70.  If you went B23, 15MHz AWS3 gets orphaned off as B66 does not cover that range.  Making it TDD would not be a good idea either because then the FCC and those adjacent will want guard bands, especially next to AWS1 UL.    It also won't solve the issue that Sprint has with TDD unless a configuration is picked that makes most the airtime for UL. But then you are still wasting a lot of airtime for DL that isn't needed.

[Tengen31]

The Snapdragon 855 has the X24 modem which does support 3xCA UL.  Are the phones not coming out of the FCC with full support?

The speed problems are solved just by having a larger dedicated channel at lower frequency for supplemental UL.  Having 15MHz at 1695-1710 solves that. Maybe could even use it to increase the range of B41/N41?

Im not saying 35MHz UL would be bad or anything or not for 35MHz UL.  But people don't UL as much as they think they do.  Facetime, Duo, Skype, etc don't use as much data as people think they do.  Sprint's UL is not at capacity. This is why UL on the other carriers FDD has plenty of speed behind it, even with the DL is slammed.  The low speeds are a connection/time issue which on Sprint gives huge inconsistencies.  15MHz easily solves the low speeds, gives plenty of additional capacity and fixes consistency issues within B41/N41 coverage area. 

But considering the DL capacity that is now available on these new massive mimo sites... what is another 20MHz. 

But if Dish keeps this stuff to build out their network, they totally want all of AWS4 for DL. 

Im working here with bands that already exist.  AWS3 UL only exists in B70.  If you went B23, 15MHz AWS3 gets orphaned off as B66 does not cover that range.  Making it TDD would not be a good idea either because then the FCC and those adjacent will want guard bands, especially next to AWS1 UL.    It also won't solve the issue that Sprint has with TDD unless a configuration is picked that makes most the airtime for UL. But then you are still wasting a lot of airtime for DL that isn't needed.

If the merger fails Sprint needs to do something to host Dish's spectrum. For sure the 600 mhz and maybe even the aws. How would they fund that you ask? Dish pay for it then a MVNO deal on the Sprint network.

 

Sent from my SM-G965U1 using Tapatalk

 

 

 

[Brad The Beast]

On 9/8/2019 at 2:05 AM, RAvirani said:

Sprint's UL on L2500 is slammed right now. How else could they have the fastest download speeds and the lowest speed score in many cities across the country according to RootMetrics?

Speeds like these are unfortunately common on L2500:

That's how band 41 always is. 

On 9/8/2019 at 2:05 AM, RAvirani said:

How often do you speedtest lower download than upload on Sprint?

100% of the time on band 26. Sometimes on band 25. Just ran a speed test now on 5x5 band 25 and got 3.23Mbps down and 12.1Mbps up.

[dkyeager]

5g properly implemented should fi x b41 upload issues. Multiband CA would also help.

[Flompholph]

11 hours ago, dkyeager said:

5g properly implemented should fi x b41 upload issues. Multiband CA would also help.

Not if qam is the reason no matter how much B41 thrown at upload it won't matter. It would only go from .1 to .2 Mbps. As long as downloads are working then there is enough uploads. Or you would not be able to request data packets. 

If someone needs upload that bad they should be forced off the band(load balancing) like they do for downloads. If someone is skyping that is only .1 to .5Mbps upload. If live streaming that is only ~1Mbps. There is no big demand for upload unless you run p2p server on phone.

[nexgencpu]

44 minutes ago, Flompholph said:

Not if qam is the reason no matter how much B41 thrown at upload it won't matter. It would only go from .1 to .2 Mbps. As long as downloads are working then there is enough uploads. Or you would not be able to request data packets. 

If someone needs upload that bad they should be forced off the band(load balancing) like they do for downloads. If someone is skyping that is only .1 to .5Mbps upload. If live streaming that is only ~1Mbps. There is no big demand for upload unless you run p2p server on phone.

I can tell you from experience, 5G uploads are exponentially more stable than 3xCA LTE. Where I see 1-3mb on B41 ill see 10-30mbs on 5G coming from the same site.