Quite a number of us have been looking at this issue for months, which I call Sequential GCI Sectors (SGS). For those unfamiliar it eliminates the special 8T8R market GCI and reorders all GCI band sectors first by sector then by band. Without software help this is uncomprehensible to most people.
It took a while to get a reasonable sample of data from all sectors from various sites in various tracked markets. Sprint was also making changes and initially had various models in different markets. This has only been seen in Samsung markets, although I have looked at data from ALU and Ericsson markets. It should be noted that Mini Macros and B41 small cells are not affected. It now seems to have stabilized for a while around what is shown in ingenium's chart:
Implementation is growing and is getting faster. Through the many logs I analyze I have seen it pop up in more and more markets. Still uncertain as to the reason. The latest theory is for 4x4 MIMO and 256 QAM with a long line of failed theories behind that one.
There are areas of uncertainty surrounding 1) counties with portions of various b41 radius on different frequencies and 2) the handling of 4th and 5th carriers. We are hoping that Sequential GCI Sectors (SGS) in Columbus and Cincinnati will soon spread to such areas so we have a better data sample. Currently we only have the area around Youngstown to work with in this regard. Joski1624 has been key in this along with ingenium.
I have handled this with a cutover date for GCIs and tracking expired b41 GCIs. Early on implementation stretched out over a couple of weeks, but it now seems entire sites are done at once. I process logs first to find any Sequential GCI Sectors (SGS) sites which are ranked into Yes, Maybe, and No. "Yes" is determined unique GCI endings which can be increased if you have detailed spreadsheet data on the site that is reasonably current. These become Approved SGS sites with the earliest one becoming the cutover date which is stored for future use. "Maybe" are unique GCI endings that don't have an expected EARFCN. Many of these are transition errors but a signification number are for sites where more data is needed. These become possible SGS sites and are flagged for followup.
The approved SGS site data is then processed for new carriers and missing/changed PCIs. B41 is converted back to its old GCI including GCI sectors for this purpose plus to provide data for the expired B41 sites. The expired b41 sites is also used for Mini Macro GCIs that have been replaced with 8T8R in case they appear elsewhere in the future. This effort allows old spreadsheets to be retained. Should a market fully convert (none found so far) to Sequential GCI Sectors (SGS) it will be worth revisiting this multi-step process to simplify it.
The unresolved issue of sites in counties with portions of various b41 radius on different frequencies is the most difficult for SCP. Sheets may have to track earfcns for those areas and/or store the SGS sectors for them to be properly rebuilt for SCP "national" logs.
Markets with sheets will want to let SCP know the approved SGS sites as users may not be on the right band or sector to know independently. This way regular sites could be stored in their compressed format while only the SGS sites would need more entries. SCP may need to track SGS sites for markets without sheets based on past key GCI sectors.
This may also be the right time for SCP to allow the direct importation of true site lat/longs as RSRP allows for easier detection of EARFCNs in transition.
Another issue that would help is a more resilient error handling (retries) for rooted modem earfcns. Currently this is often tripped by other apps such as LTE Discovery (used for its airplane mode like ability) or Network Signal Guru (for bandwidth, MIMO and QAM detection). Presently you must exit and restart SCP to regain EARFCN detection in such cases.
Based on past history, it is safe to assume that Sprint will not totally upgrade all sites in a market to Sequential GCI Sectors (SGS), but it will likely dominate -- it will just take a while in true Sprint fashion. Many of us S4Gru folks have retained older phones (I have 4) for assigning to different signal tracking purposes. The general public often has older phones just like these as their main phone. Without Sequential GCI Sectors (SGS) support older phones in SGS areas may soon become worthless for interactive signal detection and learning.
Here's my multi-point theory on why Sprint isn't roaming on VZW for LTE:
Fewer Sprint phones have B13 than B12, and at the point that you're roaming on another carrier you want that low-band signal, most of which is still B13 on VZW.
VZW charges more for data roaming than AT&T does, and might not even let Sprint do LTE roaming. This is bolstered by the fact that e.g. FreedomPop uses AT&T rather than Verizon as the underlying network for its stuff...they wouldn't use it if it was expensive.
Also, add the PH-1 to the list of unlocked/BYOD phones that likely have the same firmware as the Sprint version, and thus get WiFi calling (though the PH-1 doesn't have Calling Plus; fingers crossed that I get VoLTE).
I've been in a few RF dead zones (yay hotels!) lately but have seen the phone drop down to native eHRPD rather than swap to roaming on e.g. T-Mo band 12, so my guess is that I won't see aggressive roaming to LTE until the VoLTE soft launch, and maybe not even then.
I'll be heading to the UK in a few months and am trying to figure out whether it's worth it to switch off of Open World (this'll be my first time out of the country, so it's not like I'm using CA/MX data anyway) and grabbing a week pass for while I'm there.
Specifically, does anyone know which network Sprint roams on in the UK, and whether it's LTE roaming?
If the roaming situation is less than ideal, I figure I'll keep Open World and swing by an EE or Three store to pick up a local SIM, using my 3-cent-per-MB Open World data to get there. But if I can get solid coverage on LTE, I'll gladly do the week pass thing to save the hassle of finding a SIM.
Pretty sure they can slowly repurpose LTE portion of M-MIMO to 5G nr so that shouldn't be a problem in the very long run. Just like Clear equipment, and those have served their purpose quite well over the last 6 years.