Trip

I just tried pulling the previous beta off my US Cellular phone and my AT&T phone, both of which hadn't been updated, and now even that version just crashes on startup on my T-Mobile phones.  Sigh.

- Trip

Just installed the latest beta on both T-Mobile test phones (that was a bad idea), both Moto E5 Play and G4 Play and SCP is completely broken on both, reporting no data at all.  Works fine on my LG G8X (Sprint).

Diagnostics sent.  Can you send me the APK for the previous version?

- Trip

I hate articles like this.

https://www.lightreading.com/ossbss/t-mobile-outlines-sprint-network-shutdown-plans/d/d-id/765997?

Title sounds informative, article doesn't even deliver a sliver of what the title promises.

Actual webcast is here:  https://investor.t-mobile.com/news-and-events/events-and-presentations/event-details/2020/UBS-Global-TMT-Conference/default.aspx

If I find time, I'll give it a look later.  If anyone else beats me to it, then I won't bother.

- Trip

That's nearly a month old.  Nothing new there.

- Trip

1 hour ago, davidtm said:

Really? I'm afraid of losing Shentel's network quality if I switch SIMs.

There are quite a few areas where T-Mobile severely underperforms Shentel right now.  Except for actual roaming, none of the measures T-Mobile is using in Sprint territory are currently implemented in Shentel land right now, which means you don't fall back on Shentel when T-Mobile is too weak--you can only roam after everything falls off a cliff.

- Trip

I believe this $400 phone is as cheap as you're going to get right at the moment: 

https://www.t-mobile.com/cell-phone/t-mobile-revvl-5g

5G-NR is just too new, and the really cheap phones don't have it yet.

- Trip

Having made the round trip down to Farmville and back yesterday, I can say that a number of additional sites are now broadcasting 312250.  I'm pleasantly surprised how many there are.  Interestingly, two of them are actually co-located with T-Mobile sites, if I'm not mistaken about the locations of one or the other.  But the others are all not co-located and would provide T-Mobile with extra density.

I need to drive into Old Town Alexandria again and check on the three sites I had checked a few weeks ago, but if this is more representative of what T-Mobile intends to keep, I'm much happier than I was a few weeks ago.

I can also confirm that 312250 is still not present in the Shentel region.

- Trip

30 minutes ago, mdob07 said:

I'm hoping so, I've found several 312250 sites around here but some that I would have thought would be kept haven't been broadcasting it. Ill keep checking those sites. 

Could it be simpler than this?  Could 312250 be real-world testing--see which non-overlapping sites actually end up with substantial amounts of traffic on them to identify holes that these sites could fill?

The thought just occurred to me, though if that were the case, then I'd expect to see 312250 everywhere at once, I suppose, with sites gradually filtered out rather than added in.

- Trip

On 11/1/2020 at 7:30 AM, Trip said:

3) Sprint keep sites seem to be less common than I would have thought.  I've tracked the number of 312250 PLMN sites and I count 10 so far.  One was a site I did not expect, while the others were expected or, at least, unsurprising after plotting on a map.  I've specifically driven up to three more I did expect to be kept which either are not keep sites or have not yet added 312250; I don't know if they're still rolling it out or not to know the answer with certainty.  In all three cases, I suspect outdoor coverage would be okay without them, but I'm less sure about indoor coverage, and in at least one case, Sprint is the only carrier at that site.  (312250 is not yet deployed in the Shentel region, where I would expect to see many more.)

 

Follow-up to this post.  One of the sites I had driven up to suddenly showed up as a 312250 site this morning even though it did not when I first checked.  So perhaps they are rolling it out in stages after all?

- Trip

On 11/10/2020 at 4:45 PM, BlueAngel said:

Well that's what they need to do, add T-Mobile to all of their sites boom problem avoided!

And as long as they do that, then all will be well.

I'm just not putting money on it happening.

- Trip

I am, by contrast, extremely wary of T-Mobile buying Shentel.  Shentel is a known quantity, and T-Mobile is not.  I'm very concerned they will try to cover the area as cheaply as possible rather than providing the best service as Shentel does today. 

If they keep all the Shentel sites, then great.  If they decide to thin them... that's bad.

- Trip

1 hour ago, mikejeep said:

I think I'm just going to add a "Trip" menu to the app, where I put all of your special tweaks in one spot.. 😜

I rather like this idea, to be honest.

😆

- Trip

3 hours ago, mikejeep said:

Maybe I could add an option to force-log "Connecting" entries as LTE automatically.. perhaps set a flag that indicates it's questionable. I'll play with some ideas when I get close to finishing 5G logging.

I rather like this idea, to be honest. 

- Trip

1 hour ago, mikejeep said:

When it's showing "Connecting..." are any of the signal metrics on the screen changing? I'll have to double-check the code to confirm, but I don't think it will log anything because that displays when the device doesn't "know" what the connection type is.. that's why it doesn't say LTE. Even if the metrics appear to indicate LTE, the device isn't reporting that -- so SCP patiently waits for it to get sorted out. The logger won't know how to save any information it might be getting during that state. That label usually does not persist improperly, although sometimes when your connection is changing it may display briefly. Are these phones on active lines?

No.  I have, at present, three phones on active lines (AT&T via FreedomPop, US Cellular prepaid, and my main Sprint line) and three phones using SIMs solely to show LTE signal (Verizon E5 Play, T-Mobile E5 Play, and T-Mobile G4 Play).  The two with this issue are in the latter group (which the third, the T-Mobile E5 Play, doesn't seem to have this issue that I've noticed), but they generally say "Connected to LTE" except sometimes when they don't.

But yes, the statistics are updating.  I tend to eyeball my phones in case, for example, the GPS stops updating, which has been less of a problem lately.  If I notice something particularly interesting, I've taken to tapping it and saving an empty name, which will at least give me a row with the PLMN and GCI, and then a bunch of blank data in other places, like so:  https://imgur.com/a/lXt3DGQ

Before I locked out GSM and HSPA with Network Signal Guru so that my AT&T and T-Mobile phones only ever connect to LTE, I used to get "Connecting..." rows in my GSM log that were sometimes questionable.  (I could clearly tell some of them were actually LTE data.)

- Trip

Great!

One other question I have is about the logging functionality.  I've been having some issues with specifically my Verizon phone not logging things at times even though logging is enabled and both GCI and TAC are present.  I added back one of my old G4 Plays to try to chase down the B26 312250/B41 T-Mobile while keeping my primary T-Mobile phone available for B71, but have had a similar issue there.

I cannot be 100% sure, it seems that both of these phones will sometimes show correct and updating data but have "Connecting..." at the top right rather than "Connected to LTE."  (The notification icon will also show "--" rather than the dBm/band numbers.)  If I reset the mobile connection, it'll come back to "Connected to LTE" and then appear to log.  Does "Connecting..." when the phone is actually on LTE keep the phone from logging?

- Trip

Addendum:  I got last night's OpenCellID data and it shows additional sites across the country with 312250, but the data is very spotty.  Pretty sure I'm the only person in the DC metro area who has contributed 312250 data to OpenCellID based on what I'm seeing.

The Mozilla Location Service data doesn't have anything for 312250 at all.  I contribute there as well, so not sure where that data went.

- Trip

A few observations about T-Mobile's network over the past few weeks.

1) It looks like T-Mobile is making some type of change to the software configuration of the 600/700 panels, even if only one of 600 or 700 is actually deployed on them, such that each are getting new GCIs separate from the existing values.  It seems to be happening in batches, as I've seen sites changing gradually over the past several weeks while passing through the same areas. 

As it stands now, some sites have up to six or seven (!) GCIs, the original for PCS, a slightly newer one for AWS, then a brand new one for 600/700, and three or four additional brand new ones (one per sector) for 2.5 GHz.  It's a bit excessive, and backs up my opinion about the T-Mobile network being a trainwreck as far as tracking goes.  On top of that, none of the PCIs match across bands, except that in the DC area at least, the 2.5 GHz PCIs match the AWS ones.  (In Richmond, I've noticed, that is not the case, and the 2.5 GHz ones are a fifth, separate set of PCIs.)

More amusingly, what this means is that sites that lack any bands except 600/700 are changing GCIs entirely, with the old GCI nowhere to be found.  So all the sites in Charlotte County VA, for example, got new GCIs entirely as none of them have any bands beyond 600/700.

The good news is that the PCIs are not being changed when the GCIs change, so as long as the sites have been previously identified, it's relatively straight-forward to figure out the new GCIs.

2) Apparently it is now time for 2.5 GHz to roll out in this area (Alexandria, VA).  Two months ago, there were like four sites immediately surrounding me with 2.5 GHz on them, and very little work had been occurring.  Now there's closer to a dozen, with at least a half-dozen more that are built but not yet operational.  It took a while for them to get started here, but wow are they on a roll now.  None of the power line sites seem to have been touched yet, but that's not a surprise.

3) Sprint keep sites seem to be less common than I would have thought.  I've tracked the number of 312250 PLMN sites and I count 10 so far.  One was a site I did not expect, while the others were expected or, at least, unsurprising after plotting on a map.  I've specifically driven up to three more I did expect to be kept which either are not keep sites or have not yet added 312250; I don't know if they're still rolling it out or not to know the answer with certainty.  In all three cases, I suspect outdoor coverage would be okay without them, but I'm less sure about indoor coverage, and in at least one case, Sprint is the only carrier at that site.  (312250 is not yet deployed in the Shentel region, where I would expect to see many more.)

- Trip

312-250 appears to be for sites that T-Mobile plans to keep that are not co-located with existing T-Mobile sites.  I've seen something like 8 or 9 of them.  I'm not sure how it should be displayed though; it currently is Sprint** which is not bad.  I feel like it should be Sprint since it's not a site that's fully integrated yet... maybe it should be as simple as "Sprint Keep"?  The one thing I would like is for the notes to follow from the associated 310120/312530 site onto 312250 and do the 0x1450 offset like 310120/312530 does.  I've merged my Sprint and T-Mobile databases and so the Sprint data is available, it's just not following from the 310120/312530 rows when 312250 appears.

I can't really comment on the others.

- Trip

If it's anything other than making the two networks appear native to each other, I'm not interested.

- Trip

5 hours ago, Trip said:

The obvious way to find them would be to find a way to lock my phone to the 312250 PLMN, but do you happen to know if it's being applied to Band 26 on those sites as well as Band 41?  Because I could just lock my phone on Band 26 to achieve the same goal in that case.

- Trip

Answered my own question.

https://imgur.com/a/e1o4UlR

Yes, by putting my T-Mobile phone on Band 26, I was able to force it onto 312250.  When it should have handed off to the neighbor cell, it instead dropped to No Service.  So I'm set to try to identify some keep sites.

- Trip

The obvious way to find them would be to find a way to lock my phone to the 312250 PLMN, but do you happen to know if it's being applied to Band 26 on those sites as well as Band 41?  Because I could just lock my phone on Band 26 to achieve the same goal in that case.

- Trip

3 minutes ago, iansltx said:

Someone else on the TMo Subreddit mentioned they were seeing 310-250 as well in the Seattle area. Maybe "Sprint B41 that we want to make available to T-Mobile customers because the sites are gonna stick around"?

Amusingly, that would fit the bill.  The site I connected to was definitely one that represents a gap in T-Mobile service on any band.  I have no idea why T-Mobile wasn't on that building in the first place. 

If that is in fact the case, I'd be thoroughly interested to find out which sites fit are broadcasting 312250.  Hm, I wonder if I can lock my phone to a specific PLMN...

- Trip

I think this is new.  I have my T-Mobile phone locked on Band 41 to pick up sites with the new Massive MIMO gear.  Until today, moving outside an area with T-Mobile Band 41 has always resulted in "No Service," but this afternoon:

https://imgur.com/a/TCo2cs8

I'm not sure I've seen the 312250 PLMN before, and considering the phone has no service (just a never-activated SIM), it doesn't roam, so I'm guessing the phone thinks it's native.  Any ideas what I'm seeing here?

- Trip

Looking at the map in this area (Fairfax County, VA), I see sites that I know do not have Band 5 LTE that are showing as having 5G NR.  I need to check if PCS is now running on some of them, as I've not done that in a while.  If it is, then PCS.  If not, AWS.

EDIT:  In Richmond, they hold both the A- and B-blocks in 850 MHz.  Pretty sure they're using 850 MHz there based on the far less defined coverage area, but I wonder if they're doing DSS at all and are, instead, splitting the spectrum.  Maybe next time I'm down that way I'll take my spectrum analyzer and see what I can see.

- Trip

On 10/11/2020 at 9:20 PM, mikejeep said:

Your database issues befuddle me.. I did make a minor SQL syntax adjustment in the latest beta in the hopes that it would eliminate that weird issue where the app crashes and reports that it can't create a table that already exists. That is the only relevant error I've seen from you--and I think I've only seen it from you. The API says my previous code shouldn't generate that error, but maybe this new way will completely resolve it. Please keep using the database functionality how you normally do and see if it's better on this latest version. You mess with it more than the average user so if it works for you, it'll work for anyone!

 

I would like you especially to use the database stuff like you normally do.. export, modify externally, re-import, etc. That's one of the changes. Besides that it is generic instability. Changing a lot of deeply integrated routines and want to make sure I don't miss something. So far the testing has been surprisingly stable as far as I can tell. The only issues seem to stem from seemingly coincidental OS changes (see next post shortly).

It befuddled me too.  I must be such a pest to work with!  I'm sorry!

But I've not had a problem since I rebuilt my database for the second time.  Over the long weekend, I merged my T-Mobile and Sprint databases together at last.  I'll let you know if it misbehaves at all.  I haven't seen any issues with it thus far.

https://imgur.com/a/P7RjuRK

That's what it looks like now, with the subscripts indicating which carrier the note is for.

EDIT:  One question I had is what the sector ID in brackets at the end of the note is supposed to indicate.  I've been wondering that for a while.  I'd love to have the ability to either turn it off or have it give the sector ID in decimal rather than hex in that spot.

- Trip