iansltx

Seems like TMo has done upgrades of some sort recently on one of the sites near me. B66 is now hitting in the 60s down and 10-15M up off-peak with signal in the mid to high -90s from my place, which is maybe 5-10 db better than it was before, with much higher speeds. Pretty sure I'm not just catching things at a good time. Have to force B66-only to hit these speeds; negative n71 SNR (or trying to do useful stuff with B2) gets in the way otherwise. No idea why TMo doesn't do 66+66+n71 CA here but they don't.

Boost's new plans, announced at the beginning of this week, move downmarket from the stuff they already had: monthly fees ranging from $10 to $25, and hard-capped data (1-5 GB) rather than throttled. Oh, and streaming is limited to 480p unless you pay more, which is weird considering it's a limited data bucket. Their previously "unlimited" plans are now throttled after 35 GB. Their plan page isn't terribly clear about hotspot allotments on those plans...I see 12 GB further down the page, but not on the $50 / $60 plan cards. Big question for me is how they'll differentiate their native network from T-Mobile's service/pricing-wise, if at all. According to their FAQs, only plans at or above $35/mo get access to 5G if you have the one device that supports it. Dish could certainly build out their network where they feel like it's worthwhile, falling back to T-Mobile everywhere else, and only allowing higher-tier plans onto that network because they're the only customers where a large amount of data usage could put a cramp on Dish's profit margins. As for market launches, my bet is that Dish launches in a market where they own both band 66 and 71 spectrum. That will allow them to get phones a bit more easily. Question is how they'll handle voice. VoNR would work if they got new device support. Other approach would be a dual-SIM setup with the secondary running on T-Mobile LTE. Chipsets down to the Snapdragon 690 support that operation. Or Dish could get a phone manufacturer to throw a MediaTek Dimensity 820 in a phone for them, as that chip already supports VoNR. That should also be a cheaper chip because, well, not Qualcomm, and I doubt Dish has any illusions about targeting the high end with the initial launch either. They could get a single phone model, order a bunch of 'em, and have that single phone be their launch model for ~$500. Question is where that phone would come from; if they want a launch by year-end with an actual phone, the phone would need to be in FCC testing now if it's a new model. Right now the Redmi 10X 5G would fit the bill, albeit with different LTE and 5G bands...and the whole "being a big Chinese brand" issue. One final option would be for Dish to start their native network rollout with hotspots, sidestepping the whole voice issue. But that would require getting someone to build out a 5G SA capable hotspot for a reasonable price. Closest thing right now is Samsung's 5G capable tablets.

As I recall, B41 CA support is either 41+41 (2CA or 3CA) or 25+41. If T-Mobile wanted folks to CA, they'd have to run MFBI on their B2 sites to broadcast B25, mirroring the way Sprint would've rolled interband CA out. But, given that T-Mobile is using B41 surgically for capacity at the moment (same way international carriers tend to use B7), they'd rather phones sit solely on B41, where even on a 20 MHz channel speeds are good enough, vs. spreading load across B2, where at the moment they tend to have less capacity than Sprint did on B25. For example, Sprint has 10x10 + 5x5 of B25 here. T-Mobile has 10x10...and that's it. Contrast to double 10x10 of B4/66 (all within the B4 blocks) and I'm constantly surprised when my phone prefers B2.

I'd be really surprised if the 1x network is operating anywhere near capacity basically anywhere, even with people making more phone calls due to the pandemic. I wish I had stats to support/oppose this conjecture, but sadly no way to know without being on the internal network engineering side.

Dunno about that. BC10 1x carries a long ways, and easy enough to optimize that 1xA channel for coverage rather than capacity as the number of folks needing CDMA for voice drops. I'd be a lot more concerned about fringe data issues than voice coverage.

Yep, no real reason to add new towers in most areas. TMo's network is spaced for AWS at this point in urban/suburban areas so they have their pick of where to put 600 panels. Merely swapping 700 for 600 on the sites that have 700 would be an immense improvement. Likewise, if M-MIMO n41 is as good as folks say it is, AWS LTE spacing should be sufficient for that as well. Particularly since TMo can pick L-Sprint sites if those are better. Depending on area, those sites are spaced for either PCS or B41.

My bet is that TMo is having to pay $0 for those deals. They're intro-level promos to drive traffic to those partners, broadcast across TMo's entire customer base. Wouldn't be surprised if the money goes in the opposite direction, easily paying for whoever actually runs the program.

Question is whether SA would stick after a SIM swap back to Sprint. I have a Straight Talk data-only SIM that I could reactivate that sits on TMo, so I could test this. As for VoNR support, betting that it isn't actual VoNR, but basically the NR equivalent of CallingPlus/WiFi calling, where calls run over IP without QoS. Same idea as VoIMS on the iPhone when you don't have service from one of the SIMs.

Tried that. Phone still picked up B2 as an anchor. I'm convinced that SBS can't actually control NR band selection at this point.

Got the August security update this morning. No mention of NR-SA. Betting that's because connectivity is still routed through the Sprint backbone, and Sprint never even started building a 5G core (no huge need to when you're guaranteed to have LTE everywhere you broadcast 5G). T-Mobile's certainly capable of rerouting Sprint customers over to their own backbone, and they've already done so for a few Sprint customers, so maybe we'll get SA at some point. My guess is that this update enables it, but it wasn't mentioned in the patch notes to avoid confusion.

The thing with "connections" is, if your network is historically wider-ranging coverage wise, you'll get more IoT customers, paying you a few dollars per line per month on average. Those customers are rather cheap to service from a network load perspective, but they're also low-revenue. So there's at least some method to the madness of cherry-picking that particular number. They've definitely shut some sites down, per Reddit posts. If you have a dense network for capacity reasons and you've migrated enough customers over to TMo in that area, might as well drop the site before the lease needs to be renewed. TMo has definitely dropped B41 in some areas, either entirely or by dropping one of the carriers. We were 3CA here and are now 2CA everywhere I've tested except UT campus, where they have a third carrier at 10 MHz With all that said, for the sites that remain, I expect non-contiguous B25 not to get touched until TMo can move >= 10 MHz at once. In areas where TMo has less on its block than Sprint's they could do that by swapping the entire block. e.g. if TMo has 10 MHz and Sprint has 15, reband the 15 MHz over to T-Mobile. But that'll come pretty late in the game, as it would mean dropping Sprint to one or two 5x5 LTE carriers in the entire band, vs. the typical current config of 10x10 + 5x5. Through all this, B26 will be the absolute last thing touched. I fully expect Sprint will end up as a single 1xA channel on SMR + B26 LTE + PCS-G LTE (nothing in PCS A-F, nothing in B41) before the network finally gets turned off.

Rural sites would be the last to get decommissioned. Too many areas where one of the two networks has service and the other doesn't. Two hours west of here, Sprint has the advantage. Three hours north of that, T-Mobile does. I'd expect urban sites to be the first ones to get the ax, in areas where a single carrier is dense enough to maintain mid-band coverage. There's certainly the case where I'm sitting...T-Mobile could drop Sprint's network entirely here and folks would still be able to run on B2/66 LTE. I'd be annoyed, since the two sites that serve me on TMo are a bit further away than the Sprint one, but I understand why they'd drop the (rooftop) Sprint site. Makes perfect sense; no need to shove everyone onto 5-10 MHz of n71 when you can aggregate, even though latency will suffer. And for n41, if you can get it you'll always be within range of B2/66 LTE from the same cell site, so SA doesn't confer a range benefit.

They could do a merger rather than an acquisition. Use the AW brand since Sprint will get retired. Tricky part would be coordinating networks/spectrum.

Dish is making unconventional decisions to save money. It will succeed in doing so. As someone who uses Ting, their billing systems are fine. Nothing super special, but fine. As someone who runs plenty of stuff on cloud services, they're fine, but I'm skeptical about putting them at the core of your mobile network. Using a smaller company for NR gear may be fine, but Dish will be the first one Fujitsu builds low band NR equipment for. If one of these links in the chain goes askew, Dish will wind up competing against CricKet/Metro/Visible rather than the postpaid equivalents. They may have enough of a cost advantage to make a go of it, just like CricKet/MetroPCS did back in the day. But it's definitely a tougher road when you can't get anyone to pay more than $45/mo for your service.

I want both. Looks like you don't have a 5G phone anyway?

Any tower with n71 also has B71 LTE, though the reverse isn't true. B71 LTE is the longest range tech T-Mobile has, so from a coverage perspective a 4G phone with B71 won't be missing anything. The catch is that in many places the B71 channel is 5 MHz wide, and TMo's B71 deployment is way thinner than Sprint's B26 deployment, so having only 5 MHz hurts more. Though not as bad as having only 5 MHz of PCS. Now, there are two more factors at play here. First, outside urban areas where Dish etc. grabbed spectrum, T-Mobile has more.to work with, so.performsnce on B71 will be better (unless some rural carrier scooped band 71 up, and some did). Additionally, T-Mobile is touting range improvements for standalone 5G, which launches this quarter, so there is a chance they'll start adding NR-SA n71 cell sites in some rural locations, similar to how they added B71 in some areas they hadn't touched before. In that case your phone wouldn't work.

Dish bought Ting today, will continue to contract with Tucows for MVNO billing purposes, and will switch Boost over to Tucows' billing system in the second half of next year. So Tucows has a bit under a year to get their system set up to handle 30x the customer count that it does now...more if Dish gets super aggressive on customer acquisition. As an aside, there's every indication that Dish is building their network out as cheaply as they possibly can, using cloud providers for their core network rather than their own data centers, buying base stations from Fujitsu rather than Nokia/Ericsson/Samsung, and now farming out their billing software to a company that they fully expect will sell that software to other customers. Which...fair enough. This feels a bit like Clearwire's strategy, among others', and if that's what it takes to build out a nationwide network in three years without getting bought by Comcast/Spectrum/AT&T/VZW in the process I can live with it. This'll also make it way easier for me to test out Dish's network, since I'll just be able to pick up a Ting SIM for it when the network becomes a thing.

You have to show up in Incognito to get the redirect. As for network stuff, T-Mobile does indeed not have enough capacity on their network as it stands right now to pull everyone from Sprint over. Sprint has more capacity per customer basically anywhere B41 is deployed. The most TMo can do is allow roaming, and encourage Sprint customers to buy 5G phones, which will use T-Mobile by default. They pushed 1MM customers over several weeks ago and my guess is that there's a reason they haven't done so again. They need another 6 months to get n71 where it needs to be deployment-wise...and get n41 out to more markets...to drop the load enough on TMo LTE to move over another chunk of 4G-only customers from Sprint. It's a delicate dance because you have to refarm PCS 5 MHz at a time, and do so only after there are few enough Sprint customers that the post-refarm experience isn't horrible.

Contrary to reports, I'm still able to access Sprint's website just fine, but they're now selling the LinkZone 2 there rather than the previous set of hotspots. Based on TMo's site, it doesn't support B41. On Alcatel's site, it states that it does actually support that band, albeit without full HPUE: https://us.alcatelmobile.com/alcatel-linkzone-2/ It's not a high-end device by any stretch...at most 20 MHz of download CA, and only 1x1 WiFi...but it's less than a hundred bucks, and you can get it with Sprint's 50GB for $50 plan, so it's not the worst deal ever.

Another thing they could work on is remapping B5 to be 10x10+15x15 in areas where AT&T or VZW don't already control the whole thing. Whoever "wins" the extra 5x5 should be happy to give up 5x5 of PCS or AWS to the owner of the other half of the band. Right now, there's a sizable advantage to owning both sides of B5, but that's harmful for low-band competition.

Interesting. I'll swing by the lone 600 site near me and see if I can catch AWS-3 there.

The nearest 5G tower to me seems to be bandwidth constrained. Couldn't hit more than 25 Mbps on it with an NR signal in the mid -60s. No way is that site that congested at midnight-ish. On the LTE side, just went through various band combos and realized that T-Mobile doesn't own any AWS-3 here I guess. Wish there was an easy way to check but the FCC Spectrum Dashboard doesn't include AWS-3. What TMo calls B66 here via MFBI is actually two 10x10 carriers in AWS-1...and if you aggregate with NR you only get one of 'em. Or B2 if you're unlocky...at least around here B2 is quite the congested 10x10 slice.

Add Austin to the list of areas where B41 hasn't had serious coverage issues for awhile. Yes, you can find places where you'll drop to B25, but if T-Mobile put n41 only on sites that have B41 now, the fact that their n71 network isn't nearly dense enough won't matter anymore because no one would be on it. Speaking of which, walked over to near the closest n71 site to me last night, around a mile away as the crow flies. With a signal in the -60s and an SNR around 15 I wasn't pulling more than about 30M down. Band-selecting away from 5G got me 50+ Mbps. Something about that site is pretty broken. Also noticed while I was there that TMo appears to be the only tenant. The bottom rack is vacant, vs. the other sites nearby that have both them and AT&T.

It's still limited because B71 isn't being used as the anchor AFAIK. I definitely want the lower latency, but the higher upload speeds are a direct quote from Neville so either he's wrong or there's something else there.

Broadcast linear TV is dead though, whether OTA or over cable. People largely want to watch what they watch when they want. Hence "cloud DVR" systems, among other things. Running broadcast has limited utility, and you can use the spectrum for better purposes. You can also assume at this point that folks will have an internet connection capable of video streaming. Certainly TMo can once they roll out n41 and the like. Home broadband is an excuse for T-Mobile to roll out n41 in areas where they normally wouldn't do so. T-Vision is another revenue source/bandwidth sink that can make a case for a denser n41 build, or even mmWave (ugly VZW small cells notwithstanding).