iansltx

11 hours ago, bigsnake49 said:

I think you are forgetting about the CBRS and C-band. Maybe they will have a 2 year head start.

A two year head start (realistically 1.5) gives 'em plenty of leeway to stay in the lead for the other 3-3.5 years

Sprint B41 has been moved up so the top carrier is sitting at the very top of the band (2680 MHz center frequency). I'm sure this impacts coverage a little but in return they seem to be back to 2CA here, so I can hit ~150/10 1500' from the cell site again.

TMo really does need to add more n71 sites around here though; a couple miles south of me appeared to be LTE-only, and they have plenty of sites that they could retrofit. Of course, if they're planning on adding n41 and n71 simultaneously to a site or two, fair enough.

Looks like the 200MM coverage number is still on: https://uk.pcmag.com/networking/130839/t-mobile-4-million-sprint-customers-are-now-on-our-network

Was rather surprised how few Sprint customers have migrated over at this point. Given the spectrum thinning T-Mobile has been doing on the Sprint side, I'd think there'd be more folks on TNA or TNX, particularly with the push for new phones.

It's good to hear them quoting ~300 Mbps averages on mid-band though, indicating that sites that haven't finished optimization (or running backhaul) won't stay that way.

My bet is that if they put n41 equipment on, give or take, every site that currently has n71, they'd hit 200MM covered. They already have n41 in pretty low-density markets, going wide rather than deep. Wouldn't be surprised if merely providing citywide coverage in areas where they already have at least one live n41 site would do the trick.

106 million covered with n41 (and technically mmWave, but 99% sure unique mmWave pops are negligible), 280 million with n71: https://investor.t-mobile.com/news-and-events/t-mobile-us-press-releases/press-release-details/2021/T-Mobile-Adds-5.5-Million-Postpaid-Customers-in-2020--the-Most-in-Company-History--and-Further-Expands-5G-Network-Leadership-by-Exceeding-Ambitious-2020-5G-Goals/default.aspx

They're still saying nationwide n41 by year-end, which from previous references is ~200 million covered. I'd expect 300MM covered by n71 by then.

Curious whether we'll get another set of mid-band city announcements, or whether there was such a hard push to hit the goal that they're giving up on the announcements because they'd be too long.

Betting the next announcement is an official release of 5G home internet, including announcements on coverage for *that*, as an explicit Uncarrier event.

5G gateways are arriving for Redditors: 

Looks like at least some of them are getting 5G connectivity, including mid-band, at the same price as LTE home internet is. The gateway web UI is apparently super basic, and is missing bridge mode and port forwarding. But either way, it's happening.

On the way back to Austin, noticed that more sites are now running 15x15 B2, which nets speeds like this with 20 MHz n71:

There's definitely overlap between lightly loaded low band NR and unoptimized 40 MHz n41.

EDIT: Also saw 15+20 MHz B41 on the way back, on Sprint (above 2600 MHz). GOt 134/6 going down the highway, so TMo seems committed to providing keep sites with sufficient backhaul.

Back in Austin, B41 from the n71/n41 site seems a bit more reliable now, pulling -102 RSRP and ~130 down, though lack of HPUE on my S20 means I was only able to push ~2 Mbps up (which is I'm sure why B41 isn't prioritized, to the point that last week my brother's g fast sat on B2 even right next to the tower).

Confirmed that Kerrville's Sprint B41 has been narrowed to 15 MHz. No n41 as of yesterday, but n71 is 20x20 there, so I was able to hit ~210/50: 

Did some scouting out west of Kerrville. There were a few places on 27 that were 1x-only before dropping out entirely. VZW kept service a bit further out, but eventually fell off as well. T-Mobile 5x5 B2 was available even further west (Mountain Home area). In some cases Five Star Wireless/West Central Wireless CDMA/GSM were the only signals available from anyone.

...and then I found a 320' tower at 29.991162, -99.690123. That site only has T-Mobile on it, at basically the top of the tower, with 5x5 B12, 5x5 B2, 10x10 B71, 15x15 n71, and I believe 10x10 B66. Outdoors, I could pick up a usable signal (~12 Mbps down) ten miles away, with B2 being surprisingly durable (guessing the narrow channel width allows for more output power). B71 would pass traffic indoors at that distance but I couldn't get web pages to load. I was on a hill at the far end of that 10-mile link, so the site height definitely helped.

19 minutes ago, PedroDaGr8 said:

Considering B26 is so small as to be irrelevant, I could see them reserve the 5 MHz which differentiates B2 and B25 for legacy Sprint customers and the rest going for N2. I could even see them keeping 20 MHz or so of B41 for Sprint customers but not much more than that. 

B26 has just as much bandwidth as PCS-G outside the Southeast/IBEZ, though I likewise think PCS-G will be the last spectrum dedicated to Sprint. Since Sprint has B26 on effectively every site, it contributes just as much capacity as 5 MHz of B25, maybe more due to higher RSRP.

As for PCS NR, some phones already support n25 but that support is inconsistent. Either way, betting Sprint's network is long gone before there's no LTE in PCS.

As for B41, as I'm typing this the nearest B41 cell site is already down to 15 MHz, and Austin is down to a single 20 MHz carrier. Betting that Sprint B41 will be gone before B25 goes down to 5x5.

TMo is sending letters out to their business customers giving them a year notice for CDMA going away.

I want to say the shutdown is taking about as long as it took them to fully digest MetroPCS several years back. Even if Sprint LTE remains online, that'll be an impressive feat given how big Sprint's subscriber base was when the merger happened.

Thinking they'll keep the G block and B26 online a bit longer mainly because they don't have anything better to do with those networks; n25 support on their devices is rare, and there's no point in doing n26 unless Dish decides not to buy.

Looks like NYC n41 is up to 60 MHz:

Seems like a lot of sites are coming online on n41, or being widened to 60 MHz, in a push to get mid-band to 100MM pops by year-end. I fully expect T-Mobile to hit that population goal...the next 100 million will be a bit more of a push but certainly doable by the end of next year. Entertaining that there's n41 in some places that still don't (for now) have VZW NR of any type.

Was wondering last night why B41 was down to 15 MHz in Fredericksburg when I arrived yesterday. Turns out, n41 is live here too (on one site). n71 is missing for some reason, but I'm betting it'll be back at 20x20 since with the STA TMo has access to 30x30 here and they dropped B71 LTE from 15x15 to 10x10 since Thanksgiving.

If I had to guess, I'll bet Kerrville similarly has n41 live on at least one site, with Sprint B41 shaved way down. There's no TMo B41 here, and the n41 carrier is above 2.6 GHz here so I think TMo is rather spectrum constrained here.

7 minutes ago, Paynefanbro said:

It's being estimated that T-Mobile has spent $4 Billion on C-band so far. I'm most surprised to see that Dish hasn't spent more. While I'm not surprised that Comcast and Charter have spent a ton I didn't expect them to spend more than Verizon. I wonder what they plan to do with all of this spectrum that they're hoarding. 

 

The thing about Comcast/Charter is they can run backhaul basically anywhere in their footprint cheaply. They may also want to use C Band to serve areas slightly outside their existing footprint...or make internet-only installs at lower tiers self-install-able.

Likewise surprised at how much they're spending, but there are some explanations here. Curious whether they're bidding on early- or late-clesring mostly.

4 minutes ago, RAvirani said:

This is what I was thinking, yes. EDGE would be in PCS guard bands. 

Even if you did want to run 1x on PCS, you don't need to cut an entire 5 MHz of LTE. You can just blank the resource blocks that overlap with the 1x carrier. That would be less than 1 MHz of wasted LTE spectrum (provided some of the 1x carrier was in the guard band). 

Fair point. Somehow I completely spaced on the fact that VZW is doing exactly this for CDMA now.

I actually just cycled through bands + techs on my phone and T-Mobile does seem to be packing things more tightly now; the center for their 15x15 LTE channel is 1962.8 MHz, while EvDO is sitting at channel 475 (center 1953.75). GSM is at BCCH 636, or 1955 MHz. So you have a 525 KHz guard band between EvDO and GSM, and 200 KHz between GSM and the nominal edge of the LTE channel, though at 15 MHz channel width there's another 750 KHz of guard band before the channel actually starts (so TMo could stuff another GSM channel or two in there if they needed to).

My guess is that there's still 1x on PCS, below the EvDO channel, but when I forced 1x my phone hopped on channel 476 (in SMR). Also guessing that T-Mobile H+ has a center frequency around 1972.5 MHz, which fits just fine in the scheme of things. Sprint appears to be down to one 10x10 B25 channel centered at 1990 MHz (half in the G block), which is a config I've seen commonly here (yesterday I saw it at 5 MHz, solely in G, but that didn't last).

T-Mobile now has a live n41 site (at 40 MHz) within a mile or so of where I live:

I was unable to connect to B41 from home, and n41 faded in and out as I got closer to the site, but I think I actually caught them as they were optimizing the site because things seemed to get more reliable during the hour and change I spent walking around.

The two TMo B41 carriers are at 2525 and 2545 MHz center frequencies, with n41 right above that. Sprint B41 is at 2660...I'm not seeing B41 CA anymore on Sprint and I want to say Sprint B41 was at 2640 and 2620 before so they're definitely moving stuff around now. At home the TMo B41 comes in at ~-110 but I can't connect...closer to the site I'm able to connect.

The above speedtest is using 45 MHz of spectrum; the channel is anchored by 5 MHz B66, hence the low upload speed and unimpressive downlink contribution.

My guess is that TMo will get better speeds/coverage as they dial this site in, but as it stands the site should already take a fair amount of network load off in a relatively dense area with a lot of vehicle traffic, so that's great for everyone else stuck on n71 (or LTE for that matter). Then eventually when TMo decides they don't need 3-4 B41 carriers active they can run 80 MHz of n41 here without issue. Aggregate that with 15 MHz n71 and 15 MHz B2...both of which are live now...and I could totally see gigabit down, 100 Mbps up near the cell site.

UPDATE: They seem to have done a little more optimization. I can now pull ~120/3 over 2CA B41 a mile away indoors. No n41, and I have to force B41 because apparently TMo would rather put me somewhere with better upload performance. But that's long enough range that LTE offloading will cover enough territory to matter...and if the B2/66 network is running slowly I know where to go

2 hours ago, RAvirani said:

1x/EDGE/LTE/NR would be ideal in my opinion. 

You're thinking 1xA on SMR + EDGE in guard bands? If you run 1x in PCS you might as well run EvDO because there aren't three 1x carriers worth of voice traffic anymore and you don't want to bother with <5 MHz on LTE or NR

37 minutes ago, ingenium said:

On the Pixel 4 at least, the MVNO profile that's loaded for the SIM disables the ability to turn on VoLTE. Same for red Pocket T-Mobile. I used root to set a prop that forces the toggle to always be present in settings and VoLTE works fine on GSMA and GSMT. AT&T is now at least allowing VoLTE on MVNOs on their network. But devices need updated to allow it to be turned on.

Yeah, for the short period family was on Pure Talk, VoLTE worked on the g fast and S20, but not on the Moto X4 and E4 Plus. All four devices support VoLTE on VZW. 99.99% sure all would support VoLTE on TMo but lack of B71 on the older Motos mean I haven't checked.

Looks like TMo is beginning to send out Nokia 5G-capable gateways for home internet: 

Later in the thread someone mentions late January as the go-live date for home internet over 5G. Will be interesting to see if TMo lets people onto n41 with no speed cap (or, say, 200 Mbps) at the same $50/mo they charge for LTE home internet.

If that happens, once I get n41 covering my area I'll happily drop my cable connection in favor of T-Mobile until Google Fiber shows up here, since 20+ Mbps uploads are achievable without n41 here, and if I can get better upload speeds while still getting decent download speeds I'll absolutely switch ISPs.

Saw TMo B41 at the Chipotle on Anderson Ln in Austin earlier today. Couldn't stay connected, and didn't get any n41, but was definitely using the part of the band I saw used for B41 on n41 sites when I investigated the two n41 sites north of here a few months back (~2540 MHz; Sprint B41 carriers are ~2630 MHz). Signal was rather low at that location so we're talking about neither the Sprint site on top of the Chase building nor the TMo one at Anderson and Burnet. May investigate tomorrow if weather is reasonable.

On 12/19/2020 at 2:16 PM, PedroDaGr8 said:

I'm on ROAMAHOME and saw 312-250 in the engineering screen. 

Same. 312-250 sites on ROAMAHOME behave like native TMo in that you can bounce between them and standard (B2/4/66/12/71) TMo seamlessly, versus having to lose a TMo signal first.

2 hours ago, RedSpark said:

Sprint also had huge sponsorship deals with NASCAR and the NBA. Of course, if you look at the customer metrics at the time, that's when the network began to get creaky and Sprint began to lose a substantial number of customers over time. Sprint would have been better off spending the money on the network instead and none on sponsorships. I'd love to see a customer survey of how many people stayed with Sprint because of its sponsorship agreements vs how many left because of a poor network experience. In fact, I'd love to see it broken down to a per-person cost.

The bean-counters came in and shut-down the NASCAR sponsorship and NBA sponsorship.

The NASCAR sponsorship was there from the Nextel days (hence "Nextel Cup"). Nextel's brand meshed well with NASCAR. Sprint's? Not so much...but that's just as much a testament to the Sprint-Nextel merger as anything else (Sprint should've merged with Alltel, though had that happened the 2.5 GHz band would've looked very different ownership-wise).

Yesterday and today, I got reports from folks on VZW that they have 5G here in Austin now. VZW doesn't have B5 here, so they're either doing n2 or n66. Speeds are 105-140 down, 27-30 up. VZW has 15 MHz of contiguous PCS here as I recall, and I believe they have at least one 15 MHz AWS channel as well; will have to check one of my now-VZW-based family members' phones when they're here next to confirm channel bandwidths.

If I had to guess, I'd say VZW is running n2 here because, to my knowledge, site spacing isn't particularly dense so the extra propagation loss of using AWS would present a problem, compounded with NR being more fragile than LTE. Kinda wonder how much further n2 reaches vs. M-MIMO n41 (or if VZW's n2/66 also uses Massive MIMO).

Either way, in order to get consistent 5G coverage, VZW will need to deploy a *much* denser network in n2/66 markets than in n5 markets (which in turn is denser than n71 requires). This probably means they'll have more NR capacity here than T-Mobile or AT&T as a baseline; AT&T is 10 MHz n5, TMo is 15 MHz n71, both dedicated. Of course, that changes as soon as n41 shows up, and TMo *can* exceed VZW's capacity even on low-band by adding site density on n71.

Further down the road, I expect VZW will light up a ton of markets simultaneously with C-band NR as soon as the lower part of the band is cleared (zero doubt in my mind that they'll get a significant amount of the early-clearing spectrum that also sites in the n78 band), rather like they flipped the switch on Nationwide 5G overnight. With that said, I'm pretty sure VZW's site density in my neck of the woods isn't enough to provide consistent C-band coverage. On the other hand, I expect that anywhere with CBRS LTE now...in an early-clearing area...will get C-Band NR coverage in a year.

1 minute ago, Paynefanbro said:

Interesting choice by T-Mobile to unite mmWave and mid-band 5G under a single name. It seems like they're doing so specifically to counteract Verizon's marketing message around mmWave. Verizon emphasizes UWB as high capacity and super high speed and now it seems like T-Mobile is saying the same for n41 with the added bonus of greater coverage. 

The only thing that worries me about this is that it seems like T-Mobile may be using this new name to de-emphasize their own mmWave network. It's not like T-Mobile doesn't have the spectrum, they just don't want to deploy it.

 

I wish they had a higher-resolution version of the Ultra Capacity map. Looks like they've added more n41 around here but I don't want to go hunting for it unless I know where it generally is.

As for mmWave, for the time being there's no point to doing more mmWave deployment. TMo can get orders of magnitude more mileage out of getting 2.5 built out, MU-MIMO added, and 100 MHz bandwidth turned up. For TMo, mmWave should only be used when they've run out of capacity on 2.5...which will take awhile...and at that point they can add in surgical deployments where needed.

As an added bonus, by the time TMo actually *needs* to launch mmWave, they can buy radio gear from vendors who've put an extra year or two of effort toward fine-tuning with VZW...so they'll wind up spending less on better gear. Compare that to right now, where T-Mobile's existing mmWave gear is hit-or-miss, to the point that you can't distinguish performance from n41.

I fully expect that, unlike AT&T and VZW, TMo will have about as many mmWave sites 3-4 years from now as they have mid-or-lower-band sites. Maybe fewer. Whereas by the time VZ and AT&T can start on their own mid-band rollouts, they'll have way more mmWave sites (certainly VZ, maybe AT&T) than they have macros.

Of course, I'm biased here because I don't live in NYC, which is about the only place dense enough to maybe need a mmW overlay in addition to mid-band.

5G hotspot is finally out, and plans aren't bad: https://www.t-mobile.com/news/network/5g-just-got-better

They now brand their mid/mmWave as Ultra Capacity. The hotspot device doesn't support mmWave, which is fine given how much of a cost premium that would entail.

Data plans start at $5/mo for 500MB and go up to $50/mo for 100GB. By comparison, CricKet's 100GB plan costs $90/mo.

https://www.t-mobile.com/news/network/5g-just-got-better

Low- and mid-band support (no mmWave), plans start at $5/mo and go up to $50/mo for 100GB (so a little better than the old HTC Hub plan, which as I recall was $60), and if you buy the device with a plan, bill credits drop it down to $7/mo ($268 total). So basically finance the device and get 500 MB of data for free.

Curious which chipset the device uses; if it runs an MTK one that would allow for NR carrier aggregation, making it a better device for 5G than most phones right now.

EDIT: It's an X55 chipset, so no NR CA. Not the end of the world, but if you have an X55 based phone, performance is likely not any better with the dedicated hotspot device.