iansltx

T-Mobile is leasing another bunch of spectrum from another speculator: Interesting that they were able to come to terms for San Antonio but not Austin, as the firm owns the same block in both PEAs. My bet with both this and the Dish leases is that T-Mobile won't try to renew them when they expire, or at least won't outbid Dish for them. 2.5 years from now they should have n41 online in enough places that n71 will get a fraction of the use it does now, in absolute terms, with n2/25 coming online around then as well (if I had to guess, both PCS CDMA and GSM/H+ will be refarmed directly from those techs to NR without stopping at LTE first unless there's contiguity to be had). The other side of this coin is that I'm betting Dish will try to run on 600-only as long as possible in as many markets as possible, because doing so means fewer tower leases, vs. T-Mobile who can't get away from having mid-band in all but the most rural areas. So a year after Dish builds out in an area 600 will be more valuable to them than to T-Mobile.

Android 11 + One UI 3.0 beta is out: https://www.xda-developers.com/one-ui-3-0-beta-galaxy-s20-samsung-android-11-update/ Installed it earlier today. My phone actually went through two OTA cycles to update all the way. A profile update afterward lost 5G for me, but another profile update got it back. I'm seeing some instability (system UI crashing) when the phone gets hot, but otherwise things seem fine. If you're super tied to your phone, you might want to skip this one. But since I'm at home all the time right now, the slight reliability dip doesn't bother me, so I'll keep the beta 'til the next one comes out. Network-wise, ##DATA# no longer shows CA combos other than the Sprint ones (25/26/41, interestingly including n41+B25), though they still work. ServiceMode now shows MIMO rank on the PCC. It's saying B41 near me is running 2 streams, whereas TMo B66 and 2 are at four, at least while traffic is being passed.

T-Mobile-to-Sprint connectivity is live to the point of treating both networks as one from the sound of it (vs. waiting for a service drop to switch to the other). Betting they have the kinks ironed out on that by the time the iPhone 12 gets released, so the network experience on those devices with a T-Mobile primary network will be fine.

In some of these areas (e.g. Austin) Dish is dropping to 5 MHz of band 71 until they call back the lease. This is great as far as I'm concerned, because T-Mobile can immediately put the extra 5 MHz to use, and Dish will still have some band 71 to let them hit their buildout obligation. In my market in particular (and San Antonio is the same way), the license won't be contiguous with T-Mobile's existing spectrum, but it won't matter; they have 15x15 contiguous and can move their LTE carrier from B up to F, allowing them to widen NR to 15x15. I think that the Boost reason is largely irrelevant, as effectively all Boost customers are on 4G phones, and T-Mobile is going to throw this extra spectrum at 5G. This is a low-band capacity play for T-Mobile plain and simple, a quick way to get another up to 45 Mbps of capacity on the NR band that's actually widely built out.

Checked things out and it looks like this isn't a band 71 retrofit...just the same site, configured a bit better. That particular carrier combination disappeared for a few days, but now it appears to be back, this time with more reliable uploads. Still seeing a bunch of jitter whenever I'm on 5G, but speeds are now 70+ Mbps when you were doing well to get 40 Mbps before. A god chunk of that speed has to be coming from the B66 carrier that won't normally get selected because it's 5x5 and there are two other carriers at 10 MHz in B4 (MFBI to B66) here. Feels a bit silly that 4CA (including NR) nets you all of 30 MHz of spectrum, but speeds are a little better than 4+4+2 at 10 MHz apiece so I'll take it. FWIW the anchor band is still definitely B2 here...B71 is just aggregated...but that's fine because TMo has B2 on way more sites (so no coverage benefit to SA or a low-band PCC) and the channel is double the width of the B71 channel.

Finally something NR-related! I just saw B71 aggregated with n71 on my S20. Previously I'd only see a single B2 or B66 carrier aggregated, but this time I got 10 MHz B2, 10 MHz n71, 5 MHz B66, and 5 MHz B71, netting me download speeds a bit above 80 Mbps. I can't hit those speeds on LTE alone from this location unless I switch to Sprint B41. I also saw a higher SNR than usual on both n71 and B71, so I need to check tonight on whether T-Mobile just lit a site nearer me with band 71.

Sprint-billed accounts, as a rule, tunnel through to Sprint's network before going out to the internet, no matter whose towers you're using...international (or domestic) roaming works the same way. T-Mobile's IP network handling Sprint ROAMAHOME devices (e.g. 5G phones) is being trialed.

Either T-Mobile's local distribution network or their speedtest server is better tuned...hit a new personal record for download speeds on a mobile carrier just now: https://www.speedtest.net/result/a/6473933885 That's on Sprint B41 2CA. The nearby TMo site is ~150/26 when I'm close enough...unfortunately it takes three slices of spectrum (2x B4, 1x B2) to get 30 MHz FD and it seems they aren't bothering with upload CA around here so while latency is great (I think I saw 15ms tonight) on TMo's network, peak speeds aren't phenomenal (and decay a bit after a few blocks). I wouldn't be surprised if TMo's reticence to throw n41 up here is partially due to EBS/BRS being a mess, and partially due to not having a wide anchor band to work with; 40 MHz of n41 plus a 10 MHz B66 anchor could probably get them 250 down, 20 up at a reasonable distance from the cell site, but AT&T has so much spectrum here that they can hit that on low-band NSA NR.

Yep. Traceroutes time.out so can't really tell that way. Latency is also a good bit lower.

...and back on T-Mobile's backbone. Seems like they got the speed issues sorted out; forcing B41 just netted me 130 Mbps down. Will have to check closer to the cell site tonight to see what top speeds are.

(Sorry for the quote...on the web on my phone) Back to Sprint link today. I guess the past two days was a test, or a rollout that didn't go as TMo would've liked. Speeds are solid, at least on the download side.

As of early yesterday morning, data on my phone is now routed through T-Mobile's backbone network rather than Sprint's. There are a few odd caveats to this: Hotspot data still runs through Sprint, with what appears to be a worse-than-usual latency penalty. I can simultaneously have on-phone data from T-Mobile and hotspot data from Sprint, with active data sessions on both. Latency on-phone is better across the board, even on Sprint towers; I'm seeing as low as 29ms on B41, and a few ms lower at times on 25/26. On the T-Mobile network I'm seeing as low as 16ms on LTE; thinking that latency there is 5-15ms lower than it was when routing through Sprint. T-Mobile's network seems to be having issues keeping up with this transition, particularly when using their towers. Mobile hotspot appears to be less affected, as does connectivity on Sprint towers, though for the latter I'm seeing speeds in the 70s on B41 when I was seeing 100+ Mbps before. Depending on time of day, 5G appears to be more unusable than usual. 80 miles west of here on Ting/Sprint, another S20 is still getting a CGNAT'd Sprint IP on-phone. Wouldn't be surprised if this transition only happened with ROAMAHOME lines, which MVNOs don't have. TMo seems to be performing this switchover area by area; saw a post a week or so ago on Reddit where someone saw the same behavior...though in their case the phone was Sprint-primary rather than T-Mobile-primary as I recall. If I had to guess, it'll take the better part of the month to iron out connectivity such that data speeds on the T-Mobile backbone are back up to what they were on Sprint, though latency is already better as mentioned above. During that time, T-Mobile will push other areas over to their backbone. Guessing this switch is necessary (but not sufficient) for getting access to T-Mobile's 5G core, and thus standalone 5G. It will be interesting to see when mobile hotspot switches over to the T-Mobile backbone (at least for folks who don't have a static IP), as well as what happens to the SprintLink backbone in general. Maybe CenturyLink picks it up since they seem to like buying up Tier 1 network providers and running them into the gr...er...integrating them into a diverse portfolio of international connectivity? I was going to mention Dish, but have a feeling they wouldn't want the overhead of existing SprintLink customers.

Saw some B25 neighbors when checking SCP this evening. Judging by signal levels, they were Sprint sites. My phone was still preferring T-Mobile sites, and switching to B25 still showed as Sprint, but *something's* happening.

Got a CricKet data SIM a week and a half ago for my Nighthawk M1. AT&T is 15x15 B4 and B2, 10x10 B12 and B14 (though my M1 doesn't have B14), 5x5 (!) B30...and 20x20 (!!!) B5 around here. Looks like they're also licensed for some 700 SDL, as well as another 10x10 of B66. Maybe some more PCS too? Anyway, AT&T has an absurd amount of spectrum here. In any case, I have to think they're running DSS on band 5, with a single H+ carrier taking up the remainder. Otherwise they would be limited to 5x5 of NR, which wouldn't make a whole lot of sense. I suppose you *could* just aggregate with a bunch of other bands and call it a day, but...well, I wish I had an AT&T 5G device to test out my theory. Seems like CricKet's download speeds are pretty heavily deprioritized to "whatever's left over on the network at the time", but either they aren't deprioritizing uploads or there's *just that much* upload capacity available; I've seen upload speeds of nearly 40 Mbps near cell sites...on plain old LTE. AT&T definitely got their money's worth when climbing the bunch of sites they needed to add B14 to.

The point of "I can't switch to T-Mobile on these couple of lines" is that, for the area in question, T-Mobile has one cell site with B66 + B71 + n71...and the rest of the sites in the area are on 5x5 B2. So T-Mobile's is the network that only supports calls and texts unless you force your phone to ignore B2 (which reduces coverage drastically). T-Mobile owns no B12 in that area. By contrast, Sprint has their standard setup: 10x10 + 5x5 B25 and 5x5 B26 on four sites in the area, plus a couple carriers of B41 on the single site connected to fiber. If T-Mobile took over the existing Sprint sites and swapped their 5x5 with Sprint's 10x10, that would likely be enough to make TMo viable in that area...and, incidentally, not enough to trash Sprint in the area as it would only drop Sprint LTE capacity by 25%. So at that point I could swap to T-Mobile based service, either on Ting (assuming they get better data pricing) or Mint. Either way, I'm not signing a contract with anyone, and all affected phones (S20, Pixel, Moto X4, Moto E4 Plus) were bought outright unlocked so swapping carriers is not a big deal. One catch is that only the S20 has B71, but again T-Mobile has one site in the area with band 71 on it so B71 coverage is actually less than B2...and the current owner of the Pixel lives where T-Mobile is definitely the superior option, with both B71 and B12 deployed. tl;dr: there's a reason that, for now, I want to stick with a Sprint MVNO.

The 600 they're talking about is already out there. Channel 51 loaned it for COVID reasons, and TMo wants to convert that loan to a lease. WIth how broad TMo's n71 is, they can turn up additional n71 spectrum with the flip of a switch. Though, again, *not* having access to the spectrum just means n71 speeds will get slower in the affected areas as bandwidth decreases. I saw that here when TMo went from 30x30 of band 71 to 15x15. Oh, and if Channel 51 can't lease to T-Mobile, that basically lets Dish take their sweet time building out their own network, asking Channel 51 for cheap leases on the spectrum at a later date, while TMo customers get crappy speeds due to congestion. Someone on Twitter brought up that what Verizon really wants here is TMo to divest PCS or AWS in order to get the leases. This is an inconvenient time to divest PCS because not enough Sprint customers have been migrated, but TMo could probably sell some AWS-3 and not feel it too badly. I'd expect TMo would want to divest AWS rather than PCS not only because they have more of it deployed, but also because there would be at least three bidders in the divestiture (AT&T, VZW, Dish) rather than just two (DIsh isn't going to touch PCS).

Welcome to RAMAHOME-land.

If they aren't right out of the box then they will be with a profile update. Easy enough to force B25/26/41 if needed though.

It doesn't have 5G. I wouldn't spend more than $250 on a 4G phone on TMo right now.

Fingers crossed for ya on TMo converting the site and adding n41. As far as I'm converned at this point, there isn't a compelling reason to get a 5G device at all for AT&T or VZW, as for non-mmW you'll have access to the same spectrum via DSS anyway. So for AT&T/VZW it wouldn't be a horrible idea to buy a Pixel 4a right now, where for T-Mobile doing that would limit you to 5, maybe 10, MHz of low-band spectrum in most places...and you'd miss n41.

mmWave components are expensive. Just look at VZW's phone pricing (and, for the S20, spec drops). 5G phones are already rather expensive as it is. Additionally, with 60 MHz of n41 and a decent amount of LTE (potentially with some LAA mixed in), the only benefit of mmWave is turning your phone into a pocket warmer and being at the top fo speedtest charts. Plus, you basically *have* to use small cells to do a mmWave deployment, and T-Mobile has already said they prefer the more consistent footprint that is a nearly macro-only network. When you can throw 60 MHz of n41 a mile or two, the number of places you *need* that extra mmWave capacity is way lower than if your only mid-band options are CBRS or LAA LTE, with your only actual NR running below 1 GHz. Put another way, I fully expect T-Mobile to have a smaller mmWave footprint than AT&T or Verizon, because they don't have to choose between negligible coverage plus high throughput and reasonable coverage that runs the risk of congestion. Because you're missing way less on T-Mobile by not having mmWave than either AT&T or Verizon, T-Mobile isn't forcing phone manufacturers to include expensive mmWave components in cheaper phones like VZW is (note that AT&T sells non-mmWave 5G phones as well).

5 MHz of AWS-3 is now live near me, in addition to the two 10 MHz carriers in AWS-1. Unfortunately it looks like my S20 can't aggregate three B66 carriers, so the benefit I get is more network breathing room in general rather than higher peak speeds. Thinking it would absolutely be worth T-Mobile's while to swap spectrum to get some contiguity there, ven if that means swapping 10 MHz of AWS-1 for 10 MHz of AWS-3. In the mean time, AT&T has 15 MHz each of B2 and B4, and appears to be running DSS on most of B5, allowing for a whopping 20 mHz channel there, shared with NR. My guess is that anyone not on FirstNet with B5 on their phone will use B5 as the PCC here as a result by default, so other bands may have a decent amount of upload capacity unused for folks who band-select to them (I believe AT&T's UL-CA usage is still hit-or-miss).

Oh hey, it's mmWave speeds, but with all of the used spectrum sub-6. Which means that, while not all phone can simultaneously lock on to these channels, the capacity is there across the cell. Now if TMo would just put B66 (and ideally 71) on the cell sites near my parents' place that already have Sprint, I could switch them over.

Ting just announced they're dropping the 20 GB for $20 Sprint based custom plan they had. Family has gotten used to not having to watch call or data usage, and T-Mobile service is poor where some of them are (so no Mint SIM). Thinking of going with Tello or Twigby next month when things roll over. What do y'all think? Maybe Ting will have better plans by then or something, but I doubt it, and I'm betting those won't be for Sprint based service. One person actually has better TMo service in his area so planning on putting him on Mint.

Re: guesses on H+, guarantee H+ will be gone by the time n25/66 start getting used. Re: VZW LTE going fast, betting they're running LAA, CBRS, or both. As we know from Sprint B41, throwing a bunch of TD-LTE carriers at things is a pretty good way to have high speeds. Particularly if you have to run a dense network anyway (concrete jungle) and are the local telephone company so running fiber to sites is no big deal.