iansltx

Did some cell site surveying this evening. Two sites within 1000'  of each other, one Sprint + AT&T (FirstNet but no B5), one TMo + VZW. By every indication, the Sprint one will go away.

IInteresting thing is, I'm pretty sure the TMo/VZW site has three different fiber providers to it now. Spectrum has been there since they were TWC; saw TWC Business branding on equipment. Then followed utility locator flags to the road nearby to find new-looking fiber vaults for both Fiberlight and Verizon. Betting T-Mobile is using Fiberlight, while VZW is using their own glass. Which makes me think this site will be one of the first to get C-Band from VZW. No CBRS on it right now though.

1 minute ago, Paynefanbro said:

The branding sucks regardless of who it is. However, at least Verizon isn’t afraid of explicitly talking about mmWave. T-Mobile pretty much has never done so, often falling back to the excuse of “It’s good for small pockets of urban areas” but never giving any update on its deployment. 

Probably because they don't feel a need to deploy it outside airports and stadiums, as they have mid-band available now and that's ideal for 99.9% of cases. The companies talking mmW are the ones that don't have 40+ MHz bid-band channels right now. AT&T mentioned mid-band, but only in the context of (parts of) airports, stadiums, and their own stores. AKA little enough not to matter. mmW isn't a priority.on AT&T even though they aren't in a great spot spectrum wise otherwise.

10 hours ago, Paynefanbro said:

I still don't like how T-Mobile is combining mmWave and mid-band together when they talk about it. Especially as Verizon said during their analyst day that their goal is to get to the point where about half of their data traffic in urban areas is on mmWave. T-Mobile will have to invest heavily to compete with that.

VZW didn't say when they would meet that mmW traffic % number though. Given that they didn't specify a date, "after 2025" seems like a pretty good guess. I don't expect them to invest much in mmW beyond this year for awhile.

Also, VZW is likewise combining mid-band and mmW into a single marketing term. TMo is calling it Ultra Capacity, VZW is calling it UW. Yes, their 60 MHz of C-Band will be branded the same as hundreds of MHz of mmW.

Per today's investor call, T-Mobile still expects 7-8 million home internet subscribers on their network eventually. They had 100k by the end of last year. Wouldn't be surprised if that number has doubled since then.

As of tomorrow, home internet is $60/mo, up from $50. Apparently $50 is so disruptive a price point that TMo can't keep routers in stock I guess, across their now-much-larger service area for the product.

Was finally successful in convincing a friend to switch from Spectrum, which has given him constant issues, to T-Mobile, as there's an n41 site <1200' from him, plus another ~3500' in a different direction. 306/36 with 22 ms ping and 6 ms jitter. If I had a better n41 signal I'd absolutely pick a unit up myself.

EDIT 2:

From a few minutes ago. That ping is downright excellent for cellular.

Looks like 80 MHz is not yet universal; the Spicewood Springs and Mesa site is still running 40 MHz, while the Spicewood Springs and 360 (up the hill a bit) site is running 80. The former overlaps basically completely with the bottom half of the latter. Both sites appear to have 2CA B41 deployed right below (as in, eating into guard bands of) the NR carrier (2538, 2558 center freqs).

The site near 360 appears not to be backhaul limited

Mom's S20 on Ting is now picking up 5G. It appears that n2 and n66 are both being used (VZW doesn't have band 5 spectrum here), but I'm not convinced that Samsung Band Selector is actually selecting NR bands so it could actually be one or the other. This is mainly notable as the first time Ting-on-VZW has been seen to use 5G, though of course it's the crappy DSS type.

I'll save the discussions on C-Band for other threads, beyond noting that VZW got 60 MHz in A here just like everywhere else that clears early, and will have 140 MHz once everything clears.

80 MHz n41 is now live in Austin!

Center frequency is 2607.75 MHz, so the carrier runs from 2567-2647. B41 2CA is below that. That could leave room at the top of the band for one or two Sprint B41 carriers, but the site nearest me on Sprint has been down since nearly two weeks ago so I'm not sure what's there.

1 hour ago, dkyeager said:

This is contrary to common wisdom of several years ago. Wider bandwidths were thought to be more resistant to interference. Wider bandwidths also use a lower percentage for buffers.

Not an interference issue, not a spectrum efficiency issue. Just a matter of throwing the signal as far as possible.

On 1/26/2021 at 11:43 AM, red_dog007 said:

They may get it cheap, but I'd doubt that means they would deploy a lot of it. Maybe more a protection thing so they don't have huge headaches with possible interference. 

In rural and smaller towns I go to, TMobile now is far from deploying their entire pre-Sprint spectrum portfolio.  I go to places and all they do is deploy just a single band.  If they go back with 5G they just add one more band (600).  Or they have all but one band deployed. 

Unless they are going to do B41/n41 on every tower like Sprint was wanting to do, I wouldn't keep my hopes up for a ton of rural 2.5.  You'd think though with fixed wireless services that doing full builds would make sense. 

I've seen n41 in rural-ish areas already, plus some B41 on sites that were 25/26 only a year ago. They seem to be getting better about not just deploying a single band in places, though from what I've seen the more rural setup appears to be whatever B/n71 they have, plus a 5x5 B2 carrier to use as an NR anchor. Narrower bandwidths go further, hence running 5x5 when they definitely have the spectrum for more.

9 hours ago, dkyeager said:

The real question is whether they intend to use this n77 spectrum or trade it for more n41 when the EBS auctions off later.  Of course there was no WISPs or cable in this and barely any Dish, all of which could compete for EBS.

I expect them to use it; they paid above what VZW did to grab some in a ton of major markets. Dish is 100x more likely to just sell their lone license. Interestingly, TMo only got 20 MHz in Little Rock, Rochester, and SYracuse, but 60 MHz in Albany...40 literally everywhere else.

We'll find out what their strategy is on March 11, but my guess is still that n77 will provide a distinct set of freqs for a small cell build out, probably CA'd with 24 GHz, allowing for islands of extra capacity that'll interfere with neither the macro network nor, for the most part, themselves.

Confirmed that the site behind the Randall's at Spicewood Springs and Mesa in NW Austin has n41, as well as some pretty well optimized B2. Will swing by at some point later to get panel shots. 2CA B41 was also deployed but latency/jitter were unreasonably high for some reason.

4 hours ago, Destroyallcubes said:

My area got the C blocks. No C band for me until 2023 or later. Dallas got 40Mhz roughly of the A block which should be ready for use soon. Curious to see the cell sites with C band equipment

So your area is outside the top bunch of markets then?

Re: equipment, should look quite similar to 2.5 M-MIMO gear.

1 hour ago, IrwinshereAgain said:

Do we know if current Tmobile phones such as S20 Fe 5g or A71 5G would work for all of Dish Networks likely cell phone bands?

Barely. No VoNR on either so they'd have to do some weird non-QoS'd IMS thing or fall back to TMo VoLTE. Also no n26/29/70 support. But n66 and n71 are there so they *could* work, just with unimpressive performance as at launch they'd be missing 30 MHz of downlink and 15 MHz of uplink.

Looking at the results more closely, actually looks like T-Mobile just grabbed 40 MHz in the top 70-odd markets (exceptions being Rochester, NY and Little Rock, AR, where they only got 20 MHz). So this feels like less of a play to shore up contiguity issues. My new guess is that this is for small cells. By 2024 there'll be areas where added capacity above and beyond n41 will be useful, but those areas may be larger than something mmWave could handle nicely. Cool thing about using a completely different band for small cells is you're not running the risk of self-interfering with the n41 macro network, and n77 (TMo will end up with precisely zero licenses in Block A) can be set up as islands of capacity where needed to avoid self-interference there. Which should mean better performance.

One weird thing is that n41 currently can't aggregate with n77, so folks going from a 100 MHz n41 channel to a 40 MHz n77 one might get a drop in speed, but my guess is that CA combo will show up in the next three years anyway. Or maybe we'll get 24 GHz + n77 CA, similar to plans for n41 + n71 CA now, so you won't actually see a perf hit when going from n41 to n77.

One thing I doubt we'll see is n77 being used for fixed wireless on T-Mobile. It's not a huge chunk of spectrum (basically everywhere with n41 has at least that much deployed), coverage will be significantly less than n41, and the Nokia gateways don't have support for the band (not that that particularly matters nearly three years out). Could be wrong though.

So, AT&T got 80 MHz of C-Band on average, with 40 of that showing up in December. Not a horrible showing, given how much VZW spent.

2 hours ago, Paynefanbro said:

FCC Announces Winning Bidders in C-band Auction

Verizon spent $45.4 Billion for 3,511 licenses, AT&T spent $23.4 Billion for 1,621 licenses, and T-Mobile spent $9.3 Billion on 142 licenses.

Seems like T-Mobile spent a lot for the little amount that they where they purchased.

Results here: https://auctiondata.fcc.gov/public/projects/auction107/reports/results

Guessing they spent in areas where they have contiguity issues on 2.5. They basically grabbed two licenses per PEA, so 40 MHz. Not a ton, but if they can find 60 MHz contiguous in 2.5 that gives them 100 total, which is plenty.

Dish is apparently behaving like they didn't know T-Mobile was going to drop CDMA relatively quickly post-merger:

https://www.fiercewireless.com/financial/dish-sheds-363k-wireless-subs-warns-t-mobile-3g-shutdown

This is the sorta thing that you price into your acquisition of Boost et al.; they're just posturing here to get money/devices out of T-Mobile. Or an extension to the CDMA EOL. They're weirdly silent about capacity being allocated away from Sprint on the LTE side.

On 2/17/2021 at 11:07 AM, bigsnake49 said:

The best thing to do with respect to PCS-G is to swap spectrum with the other two so you can get the adjacent C block. That's what they tried to do the last couple of rounds of spectrum swaps.

Yep. Only reason AT&T is doing B25 MFBI is for Sprint roaming as far as I can tell. They do the same thing on B5, which they also broadcast as B26. No need for CA on either because you're looking at ~256 kbps capped speeds for roaming anyway, and reasonable CA combos don't really exist for 25/26 vs. anything else on AT&T.

1 hour ago, RedSpark said:

Wow.

Yep. The $70 plan still isn't great...not a ton of hotspot and still SD-only video...but the 100GB deprioritization limit is nice and of course 5 GB of hotspot is better than 3.

Bumping the top-tier plan from 720p to UHD for video is significant though, as is doubling the hotspot cap...and removing the deprioritization limit on the entire network is a testament to how string TMo feels their network is. Guessing they figure they can add mid-band capacity quickly enough to deal with network congestion. They also probably know that VZW and AT&T can attempt to match this once C-Band comes online, so they might as well give themselves a 9+ month head start. THe other two can compete with this now, but VZW will have to use CBRS to do it, and AT&T will have to get creative via other means, so it'll cost those two more vs. just dropping n41 on a site.

12 hours ago, floorguy said:

how many lines do you have??

Single line. So my guess is that the cost of the shared data pool would be excessive for me (and it would be nice not to guess at usage as it varies for me month to month).

11 hours ago, greenbastard said:

Same in Houston. Areas with blackouts have a lot of T-Mobile sites down even though they have permanent gas generators on site. My guess is backhaul is an issue.

Sprint network is useless in these areas as well as they take on T-Mobile users who are roaming.

Stuff's somewhat recovered here cell network wise, though the Sprint site I was using before is now offline. TMo finally figured out how to run B41 again from the NR-enabled site, which is nice, as that covers a mile or so radius, including a fair number of folks without power.

So, all of my nearby T-Mobile sites are offline right now due to blackouts. Closest signal is B12 in the mid -120s...no B71 etc.

Fortunately the Sprint rooftop site near me still has power somehow, and TMo very recently (last hour or two) added back a B41 carrier centered at 2680 MHz, next to the 2660 one, so connectivity is fine as long as you're close enough to the site to get B41. B25 is a mess though...probably 26 as well.

Apparently ROAMAHOME is no more for new LTE-only device swaps; MOCN is the new hotness: 

MOCN looks to be an upgrade, as ROAMAHOME only allows seamless bouncing between both networks for keep sites; for everything else you have to lose signal first (or force band selection). With MOCN there's a band priority list that doesn't put all Sprint bands at the bottom. This *sounds* equivalent to making every Sprint site a keep site, though from the Reddit post B66 will get priority over B41 LTE (fine except in areas that have B41 CA on either carrier), and B66 will get priority over B2 (which is annoying here as B2 is 5 MHz wider unless you're doing non-contiguous CA).

ROAMAHOME is still in effect for 5G devices apparently, which tells me that keep sites are slightly better integrated than the rest of Sprint's sites, as keep sites don't serve as anchors for NR either way. They probably *could* be, but it would require hosting spectrum entirely within PCS A-F and MFBI'ing that spectrum to B2...so probably not worth the effort.