Jump to content

RAvirani

S4GRU Staff
  • Posts

    3,457
  • Joined

  • Last visited

  • Days Won

    159

Posts posted by RAvirani

  1. Hey Mike,

    I’ve noticed a location bug several times recently—it seems like the location reported by SCP lags my actual location by several minutes.

    The outcome is something similar to the “stuck location” bug we used to have. Do you have any idea why this might be happening? @S4GRU and @Trip have see  this recently as well, I believe. 

    • Like 2
  2. 45 minutes ago, Trip said:

    Yesterday, I saw a post on Facebook from an old friend about the Evolve III Maestro, a $60 laptop (at Micro Center) that apparently has an LTE modem in it.  Specifically, the Quectel EC25-AF*.  The nearest Micro Center was sold out of new ones, but offered open box at $48.  So for $51 after tax, I walked out this morning with one.  An absolute steal.

    I backed up the Windows product key and then put Kubuntu 22.04 on it.  I spent a lot of time today trying to get the build tools installed so I could get the wifi module installed and running, and once I did that, I pretty quickly got to sending it modem commands, which I've never done before.  I don't have a SIM in it yet, but it does seem to function as it should.  (I figured out the command to get Timing Advance values out of it, but without a SIM, I doubt it would have such values.)

    Now I'm trying to puzzle out how to write some kind of script or small program to use it for logging purposes.  I think I've worked out how to send commands to it from bash using socat, but I need to pick specific commands to run.  I'd actually like to replace some of my phones with this if I can get it to record cells, with GPS, and cycle through a set of bands repeatedly. 

    My current phone collection consists of 9 phones other than my main S21FE, and they require babysitting while I'm logging.  It'd be nice to pare that down to a more reasonable number and let the computer do the work, hopefully in the back seat with less babysitting.

    - Trip

    =====

    * Per the spec sheet, it supports B2/4/5/12/13/14/66/71.  That'd cover all of Verizon and US Cellular LTE, all but B29/30 on AT&T, and all but B41 and Sprint Keep on B26 for T-Mobile. 

    Sounds like an awesome project—keep us posted on how it’s going!

    Also—if you’re able to figure out logging, I can work with you to get the data uploaded to SignalCheck. 

    • Like 1
  3. 3 hours ago, PedroDaGr8 said:

    Hmm, I haven't tried that particular one yet. I will test it the next time I am down in that area. For whatever reason, I can never break 1gig at the Gregg's bicycle site even though I know it is gig+. I can get close (850 range) but never cross it.

    I’ve actually pulled two gigs on that site. 1.6 more recently. 

    https://imgur.com/a/1t7cVVg

    I’ve pulled multi-gig at the site on the 148th street exit off 520 as well. 

  4. 18 minutes ago, PedroDaGr8 said:

    @RAvirani Do you think you could implement a filter that converts all of the T-Mobile n7 to n41 (both ready logged and incoming)? At least in the meantime, until Mike implements his fix in the app. Now that n41-SA has finally gone live across the Seattle market, not just a couple choice sites, I will be doing a lot of logging of n41.

    Do the uploaded n7 entries have the correct NR-ARFCN?

  5. 1 hour ago, mikejeep said:

    The current version does hit the database each time, because that was the only method I could get to work consistently. Started re-visiting an overhaul of it a few weeks ago when I realized it might be contributing to ANRs, it's just unfortunately not as simple as you described. The background service that does the logging is separate from the activity that displays data. Might end up settling for a "close approximation" approach, we'll see..

    Would it be possible to store a variable accessible by the background service that can be incremented? I’d think you could even leave it on the disk (not in RAM) when the background service is running in the absence of the display being visible…then maybe pull it into RAM for quicker updates when the app is displayed?

  6. 8 hours ago, mikejeep said:

    I like it too -- but I believe that feature is contributing to the app instability that many users are experiencing. I don't intend for it to be permanent; I'm working on overhauling it with a different approach. My first few attempts have not produced a great alternative, I just don't want to hold up releasing an update just for this. If I don't have a replacement working when I'm ready to release an update, maybe I'll try to continue offering the current option with a warning about potential issues.

    Are you querying the SCP database every time you update the displayed number of points? If so, it might be less expensive to just store a variable and increment it every time you write a trail entry to the DB. You could sync the variable with the DB via a query every 128 points or something just in case they somehow come out of sync. 

    8 hours ago, mikejeep said:

    Ah, that is annoying. The only reports I found were both from you, but they were LTE-only and every cell was identifying itself as AT&T.

    EDIT: I found one other Dish report from another user -- SCP was identifying it as AT&T LTE with Dish NR.. so who knows.

    I have momentarily seen SCP identify Dish with AT&T LTE—see below:

    https://imgur.com/a/NxWXLbD

    Although 99% of the time it’s 5G on the secondary SIM. 

    This might also help you to understand the setup better:

    https://imgur.com/a/yK16Nmo

    • Like 1
  7. On 7/21/2022 at 9:43 PM, mikejeep said:

    New beta coming soon that removes that and updates some internal libraries that seems to have resolved all issues on my devices.

    Wait - to clarify are you removing the display of pending web data points? I do really like that feature…

    2 hours ago, mikejeep said:

    Working on crushing all of the latest bugs, but I need more data to help improve the app for Dish users. 

    You’ll have to effectively add dual SIM support to SCP to record Dish points. I played around a bit in android studio (trying to display Dish data on my Samsung 22) out of curiosity and it looks like all data is reported as coming from a secondary SIM. 

    • Like 1
  8. 6 hours ago, S4GRU said:

    On another note, I found a treasure trove of City of Everett archives online and added every missing Tmo site to the database from there, as well as update any missing Site numbers/names.  I have added or updated probably a 1,500 sites in the 2 months as i scour data sources as I have free time or insomnia.

    Oh awesome! Is there a link you could provide so I could take a peek?

  9. SCP doesn’t appear to work with Dish service. Dish SIMs contain two virtual SIMs - SIM 1 camps on AT&T and SIM 2 connects to Dish NR when available. 

    SCP knows the phone is connected to NR (it shows connected to NR in the top right corner), but the app only displays AT&T LTE information. 

    Screenshots:
    https://imgur.com/a/48PuJ1b

    I send a diagnostic report - any ideas Mike?

    • Confused 1
  10. 14 hours ago, Trip said:

    Interesting twist.  Today, I took off for southern Virginia and had my T-Mobile B71 phone (Revvlry on Android 10) and I noticed the GPS was bad as compared to my other two devices on SCP on the dash.  Couldn't get it to clear by toggling the Location Services box, so I quit and restarted SCP.

    When I did so, it first came up with the bad coordinates, then extremely rapidly zipped through from where it was stuck all the way up to where I was currently, like the coordinates were backlogged and then suddenly flushed out within a few short seconds.  More bizarrely, I noticed the same behavior on two other devices (Pixel 3a and Moto One 5G Ace, both on Android 11) and had to do the same thing to those.  I also dumped my trails on all three, assuming they were all bad by then.  The rest of my devices were correct.

    I was barreling down the highway at the time, so I couldn't investigate further.  Other than my quoted message above, I've never seen this behavior before, and now it seems to be widespread. 

    - Trip

    Almost the exact same thing happened on my Verizon Galaxy Flip a couple days ago. 

  11. On 6/6/2022 at 7:46 AM, mdob07 said:

    @mikejeepI've been noticing SCP freezing and unexpectedly closing a lot more recently since that latest beta update. It used to only do it after it had been running in the background for several hours. Now it's happening fairly frequently when I've got SCP on screen as I'm driving. Sometimes it will freeze but the system notification that it's not responding won't pop up for 15-20 minutes, or sometimes it's immediate. I have to go into settings-apps and force stop scp to get it to work again if I notice it before the system notification. 

    @mikejeep have you had a chance to look into this issue yet?

  12. 17 hours ago, S4GRU said:

    They have been coming on live quickly in the South Sound when completed.  I wonder what the delay is up that way?  Maybe RF integration?

    Robert

    The South Sound’s network is a lot more sparse which maker integration much simpler. 

    14 hours ago, PedroDaGr8 said:

    I honestly have no clue, maybe someone like Ravirani could offer some insight. The only thing I can think of is that they are waiting on additional sites to come online so that they can optimize all of the sites at once. 

    I’m guessing these sites (with the exception of #1) are delayed for cluster optimization. Optimizing things like tonnage config, etilt, mobility/HO params, etc is complex in areas with a high number of neighbors.

    Also, if nearby new sites are slated to come online in the near future, it often makes more sense to push the new config when all eNBs are ready to go. Otherwise, you’re essentially designing the network twice for a given area. 

    I’ll get back to you on site #1. 

    • Like 2
  13. 22 hours ago, Trip said:

    I don't anticipate returning for a long time.  I made that trip specifically because there was otherwise no reason for me to ever go there.

    I can clearly see some bad points on a number of sites.  I've located some (not all) of them, particularly in/near Luray, Front Royal, and Strasburg.  Up to you what you want to do with them.  I can make edits which I'm capable of making, or you can make changes as you see fit.  Whichever.  I'm fine either way.

    - Trip

    I’d recommend deleting the bad sectors. 

    This way, next time the area is mapped, the data will look good - the sites will remain where you located them (even if you delete all the sectors) and the new sector coverage polygons will be accurate. 

  14. 8 minutes ago, Trip said:

    Should it be rolled back, or should I just locate the sites properly with the bad sectors in place or something?

    - Trip

    I can try to isolate the “bad” points in each sector, although if you also uploaded a lot of good data at around the same time, that could get really time consuming. Alternatively, you can delete the bad sectors and remap them next time you’re in the area…

    I’d definitely leave the sites in their correct locations. 

×
×
  • Create New...