JWMaloney

Ericsson did a few of those in the New Orleans area back in 2014. Short poles with omni "cans" at the top. They only ever lit them up as SISO 3G though, and they used the old RRUS 11 as opposed to the 31. I'm hoping they canned that project in favor of Nokia small cells.

Anybody notice this? Sprint seems to have made a significant change in the way they are calculating their map projections. The result is that a lot of coverage has disappeared from the map, but what remains is much more accurate. I don't have access to the old map tiles, but I do have the ability to overlay the Project Fi tiles (which at this time have not yet been updated). The reduction in depicted coverage is pretty significant in some areas (many of which contained the elusive 12-mile-radius circular LTE cells). Take a look: East Texas / Shreveport Southwest Louisiana Northeast Louisiana (Duck Dynasty territory) -- those ridiculous circles are gone! Baton Rouge seems to have experienced the opposite (the exaggeration grew significantly). Note that I hid the LTE Plus layer in this one to make the difference more visible.

I'm assuming the latter is more likely. In my region, they've been busily completing a full rip/replace on their band 4 equipment and installing new multi-mode band 2/4 BTSs and RRHs with dual-sector antennas (for six-sector sites). The band 13 equipment is the only thing they haven't touched at this point, so it would make a lot of sense that they're waiting to tackle it until the FirstNet deployment.

I've been bitten by this as well. Walked into a corporate store to purchase a phone at full cost without activation, and the rep literally told me to go to Best Buy down the road because he wouldn't sell it to me. The same store has also accessed my account without authorization when a family member needed to swap a device. I don't mind that they needed to access my account to swap the device, but I do mind that they put TEP on a 5-year-old slider phone without authorization, especially given that the family member did not know the account PIN, and that this is my secondary security answer: Sprint did make changes recently to help curtail this by requiring account holders to delegate explicit permission for others to make these types of changes in-store, but I've been burned enough times to know that I'm better off ordering from Amazon at this point.

Please understand that I am referring specifically to the device's ability to consistently maintain a usable LTE connection to a single cell site. Anything more than that would not be a fair comparison, as even though Sprint and AT&T are (were before the RRUS 11/31 swaps) using the same Ericsson equipment, there is clearly a disparity in coverage between the two: That said, there are plenty of opportunities pictured for one to drive miles away from any cell site on either carrier; and as I have done so on many occasions, I can tell you, without a doubt, that it is on those occasions that the Nexus 5 will maintain a usable LTE connection much farther out and to much lower signal levels on other carriers. Now, as I have said, this could easily be due to network-side configuration parameters which direct the device to switch to 3G. But if that is the case, then you would think Sprint would be able to correct such a thing after all this time. The modem firmware shipped at launch would report much higher signal levels and would hold onto an LTE signal under circumstances where the later firmware would not. But under those circumstances, the device would easily miss incoming calls, so they corrected this in later firmware. That said, I have not seen a significant overall difference in throughput between earlier or later firmware -- only time connected to LTE. What I find happens much too often with the Nexus 5 on Sprint is that it will fall directly from band 25 to 3G even under circumstances where a usable band 26 signal is present. This is again related to the changes made to to the modem: Earlier firmware had a much longer timeout period for handover, while the shorter timeout in later firmware causes a much greater number of handover failures. This is done on purpose to limit the amount of time the user is without a data connection during the transition. This is 100% reproducible in areas with low site density. Flash the factory modem firmware, and drive between two cell sites spaced decently apart. If you begin a speed test shortly before handover begins, you can see the test pause visibly for several seconds while the phone is transitioning. If you do the same test with the latest firmware, the phone will quickly switch to 3G instead, and the speed test will fail. The bottom line is that Sprint took serious action to reduce the number of dropped/blocked calls on their network; and when your modem is tied up constantly waiting for handovers to complete (or even just holding on to a weak signal), you miss more calls. The modem firmware has preset configurations for voice-centric and data-centric usage, and Sprint chose to go with the former (and no, this is not something that the end user can change easily short of disassembling the firmware).

As someone who regularly uses a Nexus 5 on all 3 compatible major carriers, I'm going to have to disagree. I'm in Nokia/Ericsson land (Ericsson for Sprint and AT&T, and Nokia for Sprint and T-Mobile), and Sprint has absolutely been the worst performer when it comes to LTE on the Nexus 5.

RSRQ and SNR are the numbers you want to be comparing here, not RSRP.

How about indoors?

With only a 5 MHz LTE carrier, CDMA is sometimes more "usable" in those conditions, so your device will move over at the threshold dictated by the eNodeB. This is configurable just like the power levels and electric downtilt.

Xfinity Wi-Fi actually does work like that. Since it's often *open* but not necessarily *free*, your device's MAX is added to your Comcast account when you sign in. This actually answers the other concern as well, as you can manage / remove devices from your account page. Their app will even use EAP automatically for security with compatible access points. I believe AT&T may have a similar automatic connection feature on certain branded devices.

Realistically though, are there any areas where AT&T doesn't own both? They've been doing 10x10 in my market for a while. I can't see them going to the trouble for anything less.

Same experience in Louisiana. I suspect much of it is possible due to the improved performance of the RRUS 31 in Ericsson land.

Wasn't this specifically stated as a BingeOn feature during the original announcement? I seem to remember that enabling BingeOn would cut video bandwidth down to one-third on services that were *technically* compatible, even if they weren't *officially* compatible, thereby allowing you to reduce your metered data usage. The excess buffering is undoubtedly due to capacity issues on the backend of that service. This will likely clear up once YouTube is *officially* zero-rated, since (1) they will likely be able to pull lower-bitrate videos directly from Google rather than transcoding them, and (2) YouTube is one of the most-used video services, so the capacity freed up should be enough for other services. Anyone who remembers performing the "speed hack" of zeroing the RTSP proxy can attest that Sprint has been doing the same for years. The reason that modification worked was because the proxy was congested. But unlike with T-Mobile, you can't turn it off without hacking your phone.

Mike, this month makes two years since I moved halfway across my state, and I have to tell you, SignalCheck Pro was the first thing I consulted during every apartment tour. Didn't even make it into the unit at one place before I had seen enough.

The interesting thing about these examples is that the band 4 is the PCC while band 12 is the SCC. It has been stated previously that Sprint would need to use band 26 as the PCC to aggregate band 25 (or band 25 as the PCC to aggregate band 41) since you don't want to lose the PCC; but these examples seem to indicate otherwise.

One more quick interjection and I'll stop hijacking this thread. Here is an example of all six sectors on a 2C band 25 site: Here is an example of all six sectors on a 2C band 41 site:

Couldn't agree more. Many of us use both apps extensively across multiple providers. One of the things that can be confusing about CellMapper is that it uses the eNodeB ID (the first 6 hex digits of the GCI) as the site identifier. Sprint uses different eNodeB IDs for different bands, so each band will show up as a separate site. This is a business decision on Sprint's part, but there allegedly is a technical limitation to how many sectors can be addressed under one eNodeB ID. Other providers will generally use the same eNodeB ID for multiple bands to a certain extent. As has undoubtedly been mentioned elsewhere in this thread, we have seen AT&T do something odd when adding a third carrier to a site: they will reassign the gamma sector of one or more bands to a new eNodeB ID equivalent to one of the secondary UMTS Node B IDs for the site. Sometimes it's 8 sectors on the first and one sector on the second, and sometimes it's 6 sectors on the first and 3 sectors on the second. Verizon has been doing something similar with their third carrier where they use an eNodeB ID at a calculated offset from the original, but they are generally keeping the entire carrier on the same ID like Sprint does rather than the mess AT&T is making. One nice thing CellMapper does do is draw boundaries around the observed points within a given sector. Here is an example of an approximated site location which shows the boundary areas of data points collected from two different sectors: The actual site location is indicated by the pink dot (I have injected known site locations into the page using JavaScript). As a logged-in user, I can drag the site pin to its actual location; and after doing so, the boundaries are extended back to the pin, showing you the approximate sectorization of the cell: Oh, and in case it wasn't clear in my earlier post, the idiot who reconfigured the site after the RRUS 11/31 swap input Sprint's PLMN (310120) for the eNodeB ID!

It's band 12. I'll run through it again later.

Group 1 = Band 5 Group 2 = Band 5 + Band 2 Group 3 = Band 5 + Band 2 + Band 12 Group 4 = Band 2 + Band 25 EDIT: And yes, there is quite a bit of LTE Roaming+ which only appears in Group 4 (meaning it's band 25 only) -- CCA / RRPP?

Please update your link to np.reddit.com so that people don't get banned for following it and voting.

This, a thousand times this. Last week someone on Reddit was trying to convince me that Ookla proves Sprint has reliable service in the New Orleans market. On the T-Mobile side, what's up with the VoLTE call blocking? Does it not try to fall back to HSPA to complete the call (if it's available)?

Click thumbnail for full resolution. Nationwide LTE roaming (without Roaming+): Kansas LTE Roaming+ (without extended roaming): Mississippi LTE Roaming+ (without extended roaming): As of right now, selecting device-specific coverage reverts to showing LTE coverage without roaming; however, I found some additional tile sets, denoted only as groups 1-4, which may indicate device-specific, partner-specific, or account-specific coverage. Nationwide LTE+ (no group): Nationwide LTE+ (group 1): Nationwide LTE+ (group 2): Nationwide LTE+ (group 3): Nationwide LTE+ (group 4): Kansas LTE+ (no group): Kansas LTE+ (group 1): Kansas LTE+ (group 2): Kansas LTE+ (group 3): Kansas LTE+ (group 4): Mississippi LTE+ (no group): Mississippi LTE+ (group 1): Mississippi LTE+ (group 2): Mississippi LTE+ (group 3): Mississippi LTE+ (group 4):

Does VoLTE have QoS priority?

Indeed. Let's hope they make good on their plan to over-build in those areas.

Mike, I sent you a diagnostic report about this one earlier: Last year the GCI changed after an equipment swap, and SignalCheck has mistakenly shown it as LTE 2500 ever since. Something about the site in CellMapper caught my eye earlier: Data still works, so it's probably the only one with that mistake, but damn. I'm surprised they even activated it at the switch like that.