Jump to content
mikejeep

SignalCheck Beta Crew Forum

Recommended Posts

7 hours ago, jefbal99 said:

New beta out?

Yes sir! I was talking to a few people elsewhere, forgot to mention it in here. I have been scrambling trying to come up with a solution for those who are suddenly landing on PLMN ID's other than 310120, thus rendering their site notes useless.. I think I figured it out. It might also clear up the duplicate GCI problem that's mostly prevalent when hopping between Sprint/Clear sites. Fingers crossed! If it solves that issue (along with a few other nice changes) it'll go public Soon.

-Mike

  • Like 1

Share this post


Link to post
Share on other sites

My plmn are back to 310120. How does this fix affect clearwire plmn? Would the notes I put in the db show up at the wrong time?

Share this post


Link to post
Share on other sites
On 10/11/2018 at 9:24 PM, mdob07 said:

Hey Mike, I recently noticed that on AT&T the B2 second carrier notation is only working on two sectors currently. I also have a secondary B4 carrier that isn't noted at all. See picture below.

(snipped)

In my area for GCI patterns for each band. We have a couple of 6 sector sites and now they are converting quite a few normal 3 sector sites to be 4 sector sites. The 4th sector follows the 6 sector pattern for the sector that split.

(snipped)

For B66, the only site I've hit B66 on used the same pattern as the B4 second carrier in my area, but I'm pretty sure that area doesn't have a B4 second carrier deployed. Not sure what they'll use for B66 here. I'm also seeing an additional 5x5 B2 carrier popping up in areas with a 20x20 carrier existing, using the 2nd carrier pattern noted about. Areas outside of the city that are not 20x20 already have a 2nd B2 carrier deployed. If they deploy that carrier outside of the city there may be sites with 3 B2 carriers, I haven't seen this yet though.

 

On 10/12/2018 at 10:43 PM, mdob07 said:

As luck would have it I actually found this new carrier today in an area where it is the 3rd carrier. I hit the north sector, so assuming the pattern follows the same as all of the other bands it should be D3, D4, D5 (D6, D7, D8 for 6 sector sites) for the 3rd PCS carrier. I'll try to hit this site again when I have time and confirm the other 2 sectors.

I swore I replied to your messages last month, but I don't see anything in here.. not sure what happened, sorry about that.

Your second carrier sectors for band 2 and band 4, as well as band 14, are significantly different from the information I have.. and it's the first info I have about a band 2 third carrier.

I have B3-B6 as a band 2 second carrier in some markets, and B5-B8 noted as a band 4 second carrier in some markets. The app will make an educated guess if EARFCN is available and there is a conflict. I have D7-D9 as band 14.

I am guessing you have a fairly substantial collection of logs to support the sectors you spelled out. Are you confident that they are in use across your area and not one-off sites? Things change over time, so I appreciate it when people get me information "from the field" to help keep things updated!

Thanks,
-Mike

Share this post


Link to post
Share on other sites
18 minutes ago, Flompholph said:

My plmn are back to 310120. How does this fix affect clearwire plmn? Would the notes I put in the db show up at the wrong time?

Until now, SignalCheck would query the database for a site note with an exact match on the current PLMN+GCI pair. If it didn't match an existing entry, it would pull no note (and if the Logger was running, it would create a new entry).

The new logic in this beta still carries out that step first, but if there is no match and it's a Sprint-owned PLMN (including Clear/Virgin/Boost/Prepaid/etc), it then queries the database again using all of the known Sprint PLMNs in active use and the given GCI. If it finds one match, it uses that note. If it finds multiple matches (which shouldn't really happen), it defaults to the note with the longest string of text--basically to avoid selecting a blank note.

The burst of 312530 IDs that showed up for a bit (mine are back to 310120 also) got me to think in this direction, but since flipping between Sprint/Clear sites sometimes results in a mismatched PLMN+GCI pair, this same solution should cover that oddity.

Hopefully that makes sense? Kind of tricky to explain. It's not perfect, but in theory it should resolve most issues.

-Mike

Share this post


Link to post
Share on other sites

Using the latest beta at the time, I disabled EARFCN using the modem and got these results:

b0Hk0ag.png?1

CD9SUqm.png?1

This is not an exhaustive list. My hope is for SCP to catch the unique Sequential GCI Sectors to give an idea of what is being looked at band-wise.  Perhaps this has already been fixed since that time.  Thanks!

Share this post


Link to post
Share on other sites

I have a question is there a new entry added if a GCI is found to match an entry but not the PLMN?  Just asking because the size of the clearwire PLMN in the db may affect new sprint finds being added to the db.  If a new entry was added then the new logic may stop working because it would act as if the site note was blank the second time you went to a site and found a plmn+GCI match. Not sure if an option like the ignore PLMN like you have with neighbors would work. So that way if it is causing issues it could be turned on and off. 

Share this post


Link to post
Share on other sites
1 hour ago, dkyeager said:

Using the latest beta at the time, I disabled EARFCN using the modem and got these results:

(snipped)

This is not an exhaustive list. My hope is for SCP to catch the unique Sequential GCI Sectors to give an idea of what is being looked at band-wise.  Perhaps this has already been fixed since that time.  Thanks!

The sectors you hit (0B and 14) will appear in the app as band 26 without EARFCN info *if* you have enabled the "Use New Sprint/Samsung Indicators" option under General Settings. They are part of the new Samsung configuration that has been rolling out recently, but do not apply to all regions right now. 02 is also band 26, and there are several other sectors for the other bands/carriers.

-Mike

  • Thanks 1

Share this post


Link to post
Share on other sites
1 hour ago, Flompholph said:

I have a question is there a new entry added if a GCI is found to match an entry but not the PLMN?  Just asking because the size of the clearwire PLMN in the db may affect new sprint finds being added to the db.  If a new entry was added then the new logic may stop working because it would act as if the site note was blank the second time you went to a site and found a plmn+GCI match. Not sure if an option like the ignore PLMN like you have with neighbors would work. So that way if it is causing issues it could be turned on and off. 

A new database entry would be added because the app would see it as a new PLMN and GCI. So that would result in the same GCI having multiple entries, but the new entry would be copying in the old entry's site note. If the Logger is running, the new entry will be populated with all of the cell info available.

Hypothetical example.. you have a local Sprint site 123456AA that you logged months or years ago under 310120. Next week, your phone suddenly starts reporting everything as 312530 like several of us experienced recently. So, when you hit 123456AA for the first time:

  • Previous versions of SignalCheck would not find anything in the database, and consider it a brand new site. A new log entry would be created as 312530/123456AA. There would be no site note, because the app didn't look beyond 312530 for it.
  • The new logic implemented in this beta would initially not find anything in the database, but then in searching all of the Sprint PLMNs it would see the 310120 entry. A new log entry would be created as 312530/123456AA, using the site note pulled from the original log entry.
  • In both instances, you're getting a new (potentially bogus) database entry. The new method pulls the existing site note over.

Similar logic can be applied to Clear sites, which report a different PLMN. Many devices do not update the PLMN and GCI at the same time; therefore, there are probably duplicate entries in your log, half of which have no site note. I know I had a ton of Clear GCIs logged with Sprint PLMNs and vice versa. Yes, half of them are wrong. But until networks and/or devices behave differently, this problem is going to continue.

I see this new logic as a solution so that instead of blank site notes, you get an accurate site note stored for the "new" cell you hit. The only potential issue is that if GCIs are legitimately duplicated across the Sprint network under different PLMNs, and you happen to connect to both sites. Then a site note will be pulled that really shouldn't be. But that's a very very slim potential compared to the 99.99% of times this feature is going to benefit users IMHO.

It's tricky, because there's no easy way to know if a PLMN ID is accurate or a hiccup. But the app is only as good as the data that is passed to it. 

-Mike

  • Like 3

Share this post


Link to post
Share on other sites
25 minutes ago, mikejeep said:

The sectors you hit (0B and 14) will appear in the app as band 26 without EARFCN info *if* you have enabled the "Use New Sprint/Samsung Indicators" option under General Settings. They are part of the new Samsung configuration that has been rolling out recently, but do not apply to all regions right now. 02 is also band 26, and there are several other sectors for the other bands/carriers.

-Mike

1 phone out of 4 used for that test did not have the "Use New Sprint/Samsung Indicators" option turned on, which was the one these screen shots came from. Sorry for the error.  Thanks again

  • Like 1

Share this post


Link to post
Share on other sites

Very exciting!  I can't wait to try it out.  Does this change any of the behavior of the neighbor notes as I've suggested in the past, or is that a separate project? 

Relatedly, does this add sectors DA and DD-DF to LTE 750/Band 14 for AT&T?  I sent you a few diagnostics showing those in use.

- Trip

Share this post


Link to post
Share on other sites

Already loving the LTE 750 designation for band 14. Makes it much easier to spot while driving. 7e5ffbe124876bf756ab8e4fe06ce04e.jpg

Sent from my SM-G965U using Tapatalk

  • Like 3

Share this post


Link to post
Share on other sites
15 hours ago, Trip said:

Very exciting!  I can't wait to try it out.  Does this change any of the behavior of the neighbor notes as I've suggested in the past, or is that a separate project? 

Relatedly, does this add sectors DA and DD-DF to LTE 750/Band 14 for AT&T?  I sent you a few diagnostics showing those in use.

Hmm I don't remember specifically what you suggested changing about neighbor notes? I looked back at my flagged e-mails and I didn't see anything.. I'm sure it's somewhere but refresh my memory!

I didn't realize you had sent diagnostics, I just searched and found a DA and a DD from you -- if you're certain that those sectors are band 14, I'll add them since they don't conflict with any other information I have. No DB or DC?

-Mike

Share this post


Link to post
Share on other sites
On 11/28/2018 at 6:05 PM, mikejeep said:

 

I swore I replied to your messages last month, but I don't see anything in here.. not sure what happened, sorry about that.

Your second carrier sectors for band 2 and band 4, as well as band 14, are significantly different from the information I have.. and it's the first info I have about a band 2 third carrier.

I have B3-B6 as a band 2 second carrier in some markets, and B5-B8 noted as a band 4 second carrier in some markets. The app will make an educated guess if EARFCN is available and there is a conflict. I have D7-D9 as band 14.

I am guessing you have a fairly substantial collection of logs to support the sectors you spelled out. Are you confident that they are in use across your area and not one-off sites? Things change over time, so I appreciate it when people get me information "from the field" to help keep things updated!

Thanks,
-Mike

That is what I am observing over a fairly large area, mostly KY, some southern IN, and Southern IL. The 2nd and 3rd PCS as well as the 2nd AWS carrier do seem to be limited to my GCI zone, Looking through my logs I don't really see it for other areas. I've got data in my log from trips to Chicago, New Orleans, and Cincinnati all driving from Louisville, just haven't had time to do more than glance over it yet. From what I've seen glancing through it most looks to be the same as here except for B14 seems to be different in each major city. (I can send you my log file if you'd like to look over it, I've logged almost 7,000 rows of data.)

The 3rd PCS carrier is only in areas where Bluegrass Cellular owns part of PCS A block. From just south of Louisville going northward AT&T has a 20x20 carrier live on 08-0D DL EARFCN 700. They've just recently turned up an additional 5x5 PCS carrier on sites that have received the recent network upgrades centered on EARFCN 1125 using B5-BA. South of Louisville Bluegrass owns 5 MHz in the middle of the 20x20 carrier deployed in Louisville. Its split up around it to a 10x10 on EARFCN 650 (08-0D) and a 5x5 at 775 (B5-BA). The additional carrier that is being deployed now at EARFCN 1125 is the 3rd PCS carrier on D3-D8 . Its only on upgraded sites, so it must require a new or different radio configuration. I don't have a very large sample size yet as they seem to be focusing on urban areas first for the upgrades. Here is the B2 I've hit from a site in Elizabethtown KY: 

  I74HWV6.jpg

  • Thanks 1

Share this post


Link to post
Share on other sites

I've observed GCI endings 23, 3F, and 63 for AT&T B2 small cells, very small sample so far though. Hopefully I keep finding more of these. 

Ii1YRrM.jpg

Also posted in the AT&T discussion: 

 

  • Like 1
  • Thanks 1

Share this post


Link to post
Share on other sites
6 hours ago, mikejeep said:

Hmm I don't remember specifically what you suggested changing about neighbor notes? I looked back at my flagged e-mails and I didn't see anything.. I'm sure it's somewhere but refresh my memory!

I didn't realize you had sent diagnostics, I just searched and found a DA and a DD from you -- if you're certain that those sectors are band 14, I'll add them since they don't conflict with any other information I have. No DB or DC?

-Mike

I had proposed having a "Sprint" mode as opposed to the current behavior options of either locking the neighbor notes to the current PLMN (which would prevent Sprint neighbor cells from appearing when connected to Clear sites, for example) or turning off PLMN matching entirely (which could allow US Cellular or other carrier sites in the log from showing even though they would not ordinarily appear in neighbor cells).  The idea would be that if you were connected to a Sprint-related PLMN, neighbor notes would be restricted to other Sprint-related PLMNs and not just the current PLMN, so you could see a neighbor cell for 311870 (Clear) even though you're connected to 310120 (Sprint) but without seeing 311580 (US Cellular), for instance.  The catch, of course, is what to do about 311490, which is a Clear PLMN but seems to be used for T-Mobile roaming.

DA is fourth sector of the D7-D9 set; spotted locally on a single AT&T tower that has a fourth sector.  DD is part of a DD-DF set that seems to be used in the Virginia market (as opposed to the DC market).  I'm not sure why the inconsistent sector between DC/Northern Virginia and the rest of Virginia.

- Trip

  • Like 1
  • Thanks 1

Share this post


Link to post
Share on other sites

One other note; I sent you diagnostics today for a Shentel site that added Band 41 but was incorrectly flagged as "B41 MM".  0x0FE615 + 0x1450 = 0xFFA65, part of the normal pattern, and it had three carriers, indicating it's not a Mini Macro.  See here:

First is the third carrier:

https://imgur.com/a/qsBtPaB

Then here's first carrier of the same sector of the same tower noted as "B41 MM":

https://imgur.com/a/UnT0CID

- Trip

Edited by Trip
Added screenshots.
  • Like 1
  • Thanks 1

Share this post


Link to post
Share on other sites
On 11/30/2018 at 5:14 AM, Trip said:

I had proposed having a "Sprint" mode as opposed to the current behavior options of either locking the neighbor notes to the current PLMN (which would prevent Sprint neighbor cells from appearing when connected to Clear sites, for example) or turning off PLMN matching entirely (which could allow US Cellular or other carrier sites in the log from showing even though they would not ordinarily appear in neighbor cells).  The idea would be that if you were connected to a Sprint-related PLMN, neighbor notes would be restricted to other Sprint-related PLMNs and not just the current PLMN, so you could see a neighbor cell for 311870 (Clear) even though you're connected to 310120 (Sprint) but without seeing 311580 (US Cellular), for instance.  The catch, of course, is what to do about 311490, which is a Clear PLMN but seems to be used for T-Mobile roaming.

Oh I like this.. sorry if you mentioned it before and it slipped through the cracks. I think matching any Sprint-related PLMN is a good idea, I'm very happy with how that change has worked out so far on active cell. Carrying the concept over to the neighbor cells does make sense.

 

Quote

DA is fourth sector of the D7-D9 set; spotted locally on a single AT&T tower that has a fourth sector.  DD is part of a DD-DF set that seems to be used in the Virginia market (as opposed to the DC market).  I'm not sure why the inconsistent sector between DC/Northern Virginia and the rest of Virginia.

Thanks -- I'll add DA and then DD-DF!

 

On 12/4/2018 at 10:55 PM, Trip said:

One other note; I sent you diagnostics today for a Shentel site that added Band 41 but was incorrectly flagged as "B41 MM".  0x0FE615 + 0x1450 = 0xFFA65, part of the normal pattern, and it had three carriers, indicating it's not a Mini Macro.  See here:

That is showing MM because the GCI has an "even" third character and ends in 31.. the typical Sprint configuration for a MM. But I know Shentel uses 0FE and 0FF, so those even/odd rules might not apply in that region. Does Shentel have any mini macros? (oops, see below)

-Mike

Share this post


Link to post
Share on other sites
3 hours ago, mikejeep said:

That is showing MM because the GCI has an "even" third character and ends in 31.. the typical Sprint configuration for a MM. But I know Shentel uses 0FE and 0FF, so those even/odd rules might not apply in that region. Does Shentel have any mini macros?

-Mike

FF is odd FE is the base(even) for that market. 0xFFA6531 should not show as a MM.

 

Also I am having an issue with SCP beta not starting on start up of the phone. It shows a android app crash display"SignalCheck Pro has stopped." after a few minutes. If I click the feedback would that go to you or just google.

  • Thanks 1

Share this post


Link to post
Share on other sites
6 hours ago, Flompholph said:

FF is odd FE is the base(even) for that market. 0xFFA6531 should not show as a MM.

You are correct, my bad! I dug around and found a loophole in the routine that identifies Mini Macro and Small Cell sites that I think will resolve this.

 

6 hours ago, Flompholph said:

Also I am having an issue with SCP beta not starting on start up of the phone. It shows a android app crash display"SignalCheck Pro has stopped." after a few minutes. If I click the feedback would that go to you or just google.

Yes, this is a known issue with the current SignalCheck beta -- all of the System Shortcuts and the Reset Connection feature are also broken. I'm just about done repairing all of that damage.. a new beta will be up tonight that resolves most of it. I will continue working on some less-prevalent bugs that have come up. Moving from targeting Android 7.1 to 9.0 was an enormous change behind the scenes in this release, which I anticipated.. I hope that some of the changes will help improve the app's efficiency.

I encourage people to hit  "Send Feedback" whenever the app crashes and that screen appears; it files a crash report via Google Play that I have access to. However, sometimes those reports don't contain enough information to help me pinpoint the issue. I have a secondary error collector that does almost the same thing but includes more info that indicates specifically what part of the code triggered it, but depending on the manner of the crash it might not actually have a chance to execute. The more data I can gather about problems, the easier it is to track them down. I monitor a lot of statistics about how the app is performing, not just force closes but also routines that take too long to execute. I'm always trying to improve the numbers.. when a new public version is released it's fun to see the crash numbers plummet when I fix something.. less fun to see them spike.. ;)

-Mike

  • Like 3

Share this post


Link to post
Share on other sites

New beta should be available on Google Play within the hour. Fixed numerous bugs, several which first appeared in the beta from last week. Added some of the new stuff, will be adding more and continuing to squash bugs..

-Mike

  • Like 3

Share this post


Link to post
Share on other sites

Updated the beta again.. fixed a glitch that was causing force closes at startup if you had the Background Service disabled. Figured most people wouldn't have even seen the first update overnight so I'd sneak another one in there..

-Mike

  • Like 2

Share this post


Link to post
Share on other sites
21 hours ago, mikejeep said:

Oh I like this.. sorry if you mentioned it before and it slipped through the cracks. I think matching any Sprint-related PLMN is a good idea, I'm very happy with how that change has worked out so far on active cell. Carrying the concept over to the neighbor cells does make sense.

Does Shentel have any mini macros? (oops, see below)

 

-Mike

I look forward to having the option for neighbor cells.  It will help me in a certain specific case; the US Cellular tower on Willis Mountain uses the same PCIs as the new Shentel site in Dillwyn or Alpha (I forget which off-hand), and there's some amount of overlap between them.  Because I know in my head about that overlap, I can tell what I'm actually seeing.  The average person or even the advanced person unfamiliar with a local area would not.

I know you struck it out, but the question you asked is an interesting one.  I've not found any first-hand evidence that Shentel uses Mini Macros, but when I met with their lead RF engineer back in the summer, he indicated that they were in use.  I'm just not sure where.

- Trip

  • Like 3

Share this post


Link to post
Share on other sites

Another beta release was just pushed out and should appear on Google Play within the hour.. resolves an issue from 4.492b that caused Sprint LTE sites to get mislabeled (missing carriers and MM/SC/10x10 indicators). Mobile IP address display is also back--nobody told me it was broken, but I just discovered that it hadn't worked for 2 months!

This release also *might* resolve the significant bug that caused LTE band information to disappear when the PLMN ID changed (you would just see "LTE" until it connected to another cell).. fingers crossed..

-Mike

  • Like 6

Share this post


Link to post
Share on other sites

Hey Mike,  would it be possible to change SCP to identify AT&T B14 by EARFCN vs GCI pattern? After doing some analysis on my log, looking the the B14 thread on Howard Forums and the netmapper map it appears to wildly vary by market what AT&T used for B14 sectors. The EARFCN should be consistent across the country. This would keep you from having to set several different patterns for B14. 

Here is what I found:

Louisville KY, Lexington KY, and Nashville TN use BF-C4

Louisiana, Ohio, Texas - D7-D9

Alabama - C4-C6

some of California, North Carolina, South Carolina, New Hampshire, Vermont, Washington St,: AB-AD

New York, New Jersey - D2-D3

Chicago & Northern Indiana, 2B- 2D

Oklahoma and Missouri - D4-D6

Some California - C2- C5

I've also noticed that in NYC the B2 2nd carrier is AB-AD, which is the same as my B4 2nd carrier here in Louisville and the same as B14 in several markets. 

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×