iansltx

Will be posting various network updates from the four main Hawaiian islands over the next week or so. For the Big Island, there are a couple cell sites on Kona airport property, including B2 15x15 and n71 20x20. Want to say B66 15x15 as well. That netted me 300+ Mbps despite no mid-band, as I was only competing for service with another couple flights. At LAX they had 15x15 n71, 10x10 B71, 20x20 B66, 15x15 B2, 2CA TMo B41, 2CA Sprint B41. Spent a significant amount of time sitting on B41 while connecting there. At AUS, it appears that the DAS there doesn't support more than 10 MHz channels for TMo; the EARFCN is the same as elsewhere in Austin, but it's 10 MHz wife rather than 20.

99% sure it's something specific to Samsung gear. Haven't seen it in Ericsson markets, but did see it in geographically disparate Samsung markets.

Either 66 at 20 MHz or 2 at 15, with the other as SCC. Sometimes with B71 at 5 MHz as an SCC. I've likewise seen plenty good speeds out of even 40 MHz n41, so this was definitely a surprise. This was near the Embarcadero BART station (so, an area where I believe VZW says they have mmW), so rather dense, though traffic still isn't quite to normal levels yet. If T-Mobile can't get more than 40 MHz n41 to play with there, I wouldn't be surprised if they threw mmW, LAA, or CBRS in some areas to supplement capacity once outdoor usage picks back up.

Sitting on my flight home from SJC and hitting 400 Mbps in an aisle seat. That's on 100 MHz n41 (centered at 2560 MHz) plus 20x20 B66 + 15x15 B2. Sprint B41 also exists here; was seeing that in the terminal. Didn't lose signal between SF and here, even in the BART tunnel under the bay. In SF, had only 40 MHz n41, and it was rare to see speeds break 100 Mbps. VZW was either LTE-only or LTE+n5 but hit 150-180 Mbps on Visible. In SJC speeds were much slower with a similar config. Didn't see mmW or even CBRS anywhere. Also saw 25+41 CA (5x5 on B25 every time I looked, 1-2 CA on B41) in a few places.

Spent way too long on a Twitter Space earlier this afternoon and it seems Samsung equipment in particular can do 25+41 or even 25+26 CA. In any event, plenty of n41 around here, as well as 2CA TMo B41. Definitely a few spots where those bands are missing in Louisville, but the setup is drastically better than VZW at least, where they're working with 40x40 total as far as I can tell.

In Denver over the weekend and there's a significant amount of n41 here at either 60 or 80 MHz. Catch is, they have 5 MHz channels on B2 and B66 that get aggregsted rather often, and it seems like backhaul could be better (have seen sub 100 Mbps with good signal on 60 MHz in some cases). Also saw B25 (PCS-G) aggregated with B41 (2 CA) for the first time ever, as 312-250; indoors it was either that or 10 MHz n71 plus 2 or 66, while outdoors I could get n41. Been swapping out my Visible SIM every so often but have seen neither mmW nor CBRS yet...nor LTE wider than 10 MHz. n5 is rather easy to find, but apparently T-Mobile at 15x15 B2 is better than what VZW can do. Was quite surprised VZW didn't have CBRS or mmW even at Union Station. Will see what's available on VZW at the airport tomorrow morning.

One thought I had here is that VZW may lean on CBRS as additional capacity for their home broadband, aggregated with C-Band, particularly in 2022-2023 when they only have 60 MHz of n77 available. Unfortunately, current modems can only aggregate one B48 carrier with n77, but in urban areas VZW has at least 20 MHz of PALs so that kinda works out. If they can get someone to make CPE that aggregates multiple B48 carriers, they'll be even more competitive.

Ah, Ziply makes more sense; they're the type who'd actually have someone from their NOC on Reddit heh But yeah, the interference has gotta be C-Band related. Surprised TMo is testing in an urban-ish area like that.

C-Band testing interfering with Wave's downlink?

Found some more CBRS yesterday afternoon, including 4CA (plus B13 and maybe 2/66)? My my estimation, I was 0.8 mi from the macro broadcasting, with signal in the mid -120s, but ~10 SNR. I don't think B48 contributed much to speeds, but the fact that the noise floor was so low meant that the connection still happened. So VZW is definitely using GAA here, as they should. Just checked CA combos on my phone and it looks like they could aggregate n77 with a single B48 carrier, plus 2+13, so at the end of this year they could wind up with 25x25 FD + 80 MHz TD on X60-based devices, not far behind T-Mobile's 30x30 (2+66) FD + 80 MHz TD (n41) that they can run now.

So, VZW has 60 MHz of CBRS live on the site 0.5 mi north of me. Can't reach inside at this distance, but getting a bit closer to the site turns in some pretty solid speeds, given that LTE is being used: Nothing you can't do with B41 3CA of course, and there are a ton more phones that can do B41 3CA than can do B48, but I'll take it.

n261, actually. n257 is overseas.

Wait, someone's seen B12 as an anchor? Looks like it'd have to anchor n41 based on the S21's CA combos. Doesn't seem like much of a point to do that as 2 or 66 will always have at least as much bandwidth, if not more.

Not correct unless they're counting 20x20 as 40 MHz.

Do you have a source in the field indicating n66 is in use? I don't believe it is on T-Mobile. Might be for VZW or AT&T. VZW uses n2 here. TMo is n71/n41 exclusively.

The spectrum Tetris continues! T-Mobile now has 20x20 B2 LTE here (centered at 1955 MHz), using their 25x25 contiguous block from 1945-1970. CDMA is down to 1x-only, one channel wide (plus 1x in 800 I assume). GSM is at the top and bottom of the band, plus maybe a channel that overlaps 1x somehow. I assume H+ is in between the LTE channel and the 1x channel; spacing seems right, and TMo doesn't have H+ in AWS here. I believe this also subsumes Sprint's secondary B25 carrier here, which was just 5x5 anyway; I'm seeing 10x10 on Sprint, including PCS-G. B41 is still live, but when I connected it wouldn't give me CA. Will be interesting to see when stuff gets reconfigured again. If I had to guess, the next tweak will be when CDMA goes away, at which point they could start running 2/25 MFBI/MOCN on the big B2 block, and switch the old Sprint B25 carrier to n25 entirely. Depends on what % of their customer base are on 5G phones at that point I guess...and whether any sites can even do n25 right now (guessing the new L1900 equipment can).

Swung over to Fredericksburg to get vax'd yesterday. Looks like there are now multiple n41 sites SE of town, including the one that went live in December or so. 40 MHz centered at 2653.35. speeds are 250-280/15-20 because they're usually anchored by 5 MHz B2 or B66. CA with 20x20 n71 there should let things break 300/30 easily. Interestingly, they figured out how to cram in a 20 MHz Sprint B41 channel centered at 2626 MHz. This is a bump from the 15 MHz I saw previously. Assuming 5% guard bands on each side exist on TDD like they do on FDD, that means they have a 350 KHz guard band between NR and LTE. Maybe they have to do some PRB blanking as well, but NR and LTE aren't broadcast from the same site so maybe that helps. 312-250 is rather widespread there, with B25 in the G block and either 5 MHz at 1953 or 10 MHz at 1955 depending on which tower you're on. CDMA is in another non-contiguous block (1940-1945, looks like), while TMo B2 is 1970-1975. In short, PCS is a mess for TMo. AWS isn't really any better in terms of what they've deployed: 10 MHz in AWS-1, plus another 5 MHz in AWS-3 I'm pretty sure. Not sure if they own any other AWS there. By contrast, VZW has the same 4/66+66+13+2 setup as in Austin, at 10+10+10+15, with 4CA, but without n2 DSS.

Got a Visible SIM in today and...yeesh, deprioritization is a thing. Speeds are all over the map. With that said, ServiceMode isn't blocked like with normal VZW SIMs, and I was able to confirm that they will now aggregste n2 with 66+66+13, for 45 MHz total, 30 of which is LTE. Saw 164/30 on a speed test (albeit with horrid ping) so that was cool. I may test again tonight when the network is unloaded, to see if I hit the 200 Mbps speed cap. I also confirmed that mmW, if the panels I saw are really mmW, isn't het active on the site. Though I had to get within 1/3 mi of the site to even get NR...outdoors! No CBRS either. I'll keep checking of course.

So, VZW is continuing to insist that its C-Band propagates comparably to, or better than, 2.5: https://www.fiercewireless.com/operators/verizon-defends-c-band-plans The latest argument is that they can get 4db of extra gain from same-size antennas on C-Band vs. 2.5, and their power output can be much higher over a large amount of bandwidth vs. the 2.5 GHz band's restrictions. My immediate questions here are: Is VZW actually deploying the same size antennas a T-Mobile, so they actually get the gain improvement? Is 2.5 even bumping up against power restrictions to start with (from what I can tell, the answer is a solid "no", as sectors are limited to 640W EIRP total by hardware)? If C-Band is really better for coverage, why did T-Mobile completely ignore the band outside large markets where they have tighter cell spacing anyway? How does this solve transmit power limitations on the subscriber side, for mobile devices in particular? Now, maybe VZW's radio vendors have some super high power C-Band equipment in the pipeline for late 2023 that *would* run afoul of 2.5 EIRP limitations, but there's a significant jump required to actually get there. Or maybe they care about coverage with a much higher end-user transmit power (NR SA with high-gain outdoor antennas). But that's not really comparing apples to apples. In other news, I've confirmed that the VZW site 0.4 mi south of me has n2 at 15 MHz, with no n66, as VZW's B66 channels are the same non-contiguous 10 MHz setup that T-Mobile has in B4...except one of VZW's is actually in B66 (T-Mobile has another 5 MHz in B66). Pretty sure the lower speeds I was seeing on 5G vs. LTE were because 5G uses 25 MHz of spectrum (n2 15 MHz + B13 10 MHz), while LTE uses 45 MHz (B2 15 MHz + B66 10 + 10 MHz + B13 10 MHz). What's interesting is that there appears to be mmWave gear on the site. I have a Visible SIM coming in tomorrow that should help me better confirm what's going on (assuming mmW is live yet). VZW has their own fiber to the site, while T-Mobile has Fiberlight to the same site. Which brings up its own set of questions: Will this site get C-Band as well? When will mmW actually turn on, if not already? Will mmW reach my place (probably not)? One eventual issue is that I'll have an apartment complex between me and the site, so if I get mmW reception now from there, I won't in a few years when the complex goes up. But by that point there'll be 140 MHz of C_Band to play with I suppose. Or I won't care because I'll have FTTH

More upgrades! My nearest n41 site can now hit 500+ Mbps down over its 80 MHz channel. ...and in even more recent news, n71 is live on the site to the south of me (Burnet and Justin are the nearest cross streets). Hit ~200 Mbps down on what appeared to be LTE alone, interestingly preferring B12 as the PCC. When taking advantage of NR, I was able to hit as much as 90 Mbps up! On the same site, it appears VZW has mmW, though I'm not sure if it's active yet. Ordered a Visible SIM to check further. No CBRS from what I can tell. VZW definitely has a sweet spot where you can hit 200 Mbps from them as well. Also confirmed that they're running n2 at 15 MHz DSS there. No n66, which makes sense as apparently both they and TMo only have 2 10 MHz channels to play with.

That doesn't seem to match with what I'm seeing. If we're talking about a hard throttle here, I wouldn't be able to hit >100 Mbps over TMo B41. I'm also still getting my 8 Mbps on Fast.com. I can't prove deprioritization, but this doesn't look like throttling to me. Now, I'd be fine if they sent me a TNX SIM, but my plan apparently doesn't support that.

The question here is, are they doing either for this particular site...because 2640 completely overlaps the n41 channel. You might as well either run n41 at 60 MHz or not run the 2640 B41 carrier. Unless there's little enough traffic on both that the towers are silent most of the time, and the TDD pattern is set to avoid self-interference. Which...maybe?

This is all 20 MHz channels. Sprint is 2640, 2660, 2680. TMo is 2538, 2558. TMo NR is 80 MHz centered at 2607.75.

Just confirmed that I *can* connect to 312-530 Sprint sites still...got near one, hand-picked, and was good to go.l, including N41, which is still deployed at 2CA here at 2640, 2660, and 2680 MHz, in addition to TMo's 2538 and 2558, in addition to their n41 carrier (80 MHz centered at 2607.75), though I've only been able to get 2CA B41 anywhere. The bizarre thing is that this means the lowest Sprint B41 carrier basically completely overlaps the top of the TMo n41 carrier. But...I guess that's why I can't actually get n41 at my place anymore. VZW deploys DSS and T-Mobile goes "hold my beer, we're going to cram 180 MHz of band 41 into 160 MHz without bothering to do DSS".

Dumb question time: on B41, is the frequency listed the center of the carrier (like it is for FD?) or the bottom of the carrier? I'm convinced I'm seeing TMo overlapping n41 with Sprint B41 here but am not sure by how much.