Database Users

Collecting data about GPS for a fix. RAW data only, no anecdotal observations!

Hi! Some of you might recognize me from the WinMo/WinPho world. I just picked up a Captivate and i'm having the same GPS issues we all are. I noticed there's a distinct lack of raw, verifiable data regarding the issue. Nearly everything is anecdotal observations.
Please do NOT post any observational data here (example: I changed xxx and yyy setting, and now my GPS works great! Tracking is spot on!) - these posts, if not based on provided raw data, will be deleted.
Please DO post raw data (lbstestmode readings, NMEA strings, sensor output, adb logcat output, My Tracks, comparisons vs. other GPS with raw data, etc.) - also, feel free to post observations based on this raw data.
[size=+2]Apps i'm using to gather this data:[/size]
[size=+1]Android Apps[/size]
LbsTestMode: Comes with just about any ROM. Use "Any Cut" from the market to make a shortcut to it, make sure your settings match those below (unless intentionally testing others) and press "Get Position Test". If you are going to post satellite signal values, it's easiest to take a photo and type the values in from the photo, as the signal strengths vary from moment to moment.
My Tracks: This google app records a track of your GPS location, updating every 1s (unless you haven't moved beyond a small threshhold) - You can then upload this data to google maps and link it as I did in post 2. Unfortunately this method does not record any NMEA data (so we can't see signal strength, etc.) But serves will to see how accurately the tracking is.
Bluetooth GPS Provider: Used along with the "Allow Mock Locations" setting in Settings - Applications - Development to allow the phone to pull location data from a tethered bluetooth GPS such as the BT338. Note that this program will show satellite signal strengths as a bar, without actual dB readings. The program listed below provides more data, but performs a bit worse over time.
Bluetooth GPS: This program provides the same functionality as the app above. However, it provides much more diagnostic data such as signal strength readings, Satellite Constellation View, and raw NMEA output.
[size=+1]PC Apps[/size]
SirfDemo: Used along with a bluetooth module to connect to my BT338 GPS on the laptop. You can see a screenshot of this app in the photo below. It also displays raw NMEA data and allows logging.
[size=+2]GPS Settings[/size]
These are the settings that i'm using for testing. If your settings vary, please post along with the data so we can have a complete picture. The ROM/Radio/Kernel you're using is important too Also note any tweaks you may have made to gps configurations (using lbstestmode, jupiter.xml, et al.)
IMPORTANT: Use the "Delete GPS Data" button in LbsTestMode immediately prior to conducting any tests, and ensure to use standalone/cold start. This will remove all cached assistance data and ensure the GPS isn't affected by external effects.
Da_G said:
Application Settings
Session Type: Tracking
Test Mode (S/W or H/W): S/W Test
Operation Mode: Standalone
Start Mode: Cold Start
GPS Plus: ON
Dynamic Accuracy: ON
Accuracy: 50
GPS Logging: (optional)
SUPL/CP Settings
Server FQDN Type: AUTO Config
Server: (greyed out, default)
Server Port: (greyed out, default)
SUPL Secure Socket: ON
AGPS Mode: SUPL
Click to expand...
Click to collapse

I'll go first!
This is my test setup for the data in this post. You can see most of the output there on the screen. (I'll type it up for readability too) - There's about a 110 degree clear sky view, overcast. There is probably some minor multipath from the 2 story buildings nearby.
Phone setup is my own custom ROM (based on the recent Vibrant Froyo leak, with Cappy JH7 bits tacked on to address the various incompatibilities), JK3 baseband, Setiron 1.4.3 1200mhz kernel.
Format below is:
PRN of Satellite : Signal Strength : Used in Fix?
BT338 Standalone GPS (Tethered to laptop)
SV 5 : C/No 00
SV 6 : C/No 00
SV 15 : C/No 21 *
SV 16 : C/No 21 *
SV 18 : C/No 33 *
SV 21 : C/No 34 *
SV 22 : C/No 18 *
SV 25 : C/No 00
SV 26 : C/No 32 *
SV 29 : C/No 30 *
SV 30 : C/No 16 *
Captivate1 (Standalone internal GPS, to the left of BT GPS in photo)
SV 6 : C/No 13
SV 15 : C/No 23 *
SV 16 : C/No 19 *
SV 18 : C/No 32 *
SV 21 : C/No 12
SV 22 : C/No 19 *
SV 26 : C/No 21 *
SV 29 : C/No 33 *
SV 30 : C/No 11
Captivate2 (Standalone internal GPS, to the left of BT GPS in photo)
SV 6 : C/No 14
SV 15 : C/No 26 *
SV 16 : C/No 16 *
SV 18 : C/No 33 *
SV 21 : C/No 32 *
SV 22 : C/No 20 *
SV 26 : C/No 29 *
SV 29 : C/No 33 *
SV 30 : C/No 11
Some notes and observations:
I took 4 photos a few seconds apart from this angle to look at the signal fluctuations. You might notice Captivate1 SV 21 is an anomaly (12dB). The signal shows majorly degraded from the BT GPS and the Captivate2. Appx. 4 seconds before, the phone was getting 23dB on that satellite. The BT GPS and Captivate2 both remained at the same level (34dB, 32dB) - that might indicate some antenna issue after all, perhaps a marginal connection in some units. I'll do some more extensive comparison between these 2 units while I have them (2 or 3 days)
Here are 3 different MyTracks I took yesterday while driving: both GPS units on the dash.
AT&T Captivate Track:
http://maps.google.com/maps/ms?hl=e...26965825393652.000496745b69eeef7de5f&t=h&z=14
BT338 GPS Tethered to Captivate Track:
http://maps.google.com/maps/ms?hl=e...26965825393652.00049673fd50c60e8d690&t=h&z=14
Both Tracks overlaid (anyone know how to get one track as a different color for better visibility? I know you can in g earth..)
http://maps.google.com/maps/ms?hl=e...09,-117.346945&spn=0.053451,0.065746&t=h&z=14
[size=+2][Edit: Forgot to add, here is some data on fix times.][/size]
GPS in Standalone operation - Cold start
BT GPS - 59s
Cappy1 - 19s
Cappy2 - 16s
GPS in Standalone operation - Warm Start
BT GPS - 32s
Cappy1 - 12s
Cappy2 - 9s
GPS in AGPS operation - Hot Start
BT GPS (no AGPS, hot start only) - 3s
Cappy1 - 7s
Cappy2 - 5s
Interesting to note is that the 2 Captivates running the same software, used in the same position/enviornment, and trying to get a fix at the same time, Cappy2 consistently performs better than Cappy1. This again may point to some design-time issues with the GPS Antenna.

What software are you using to collect this data?

I'll post it in the first post, as i'm sure others will want to know too Look there shortly.

If we all post as replies to this thread that could get really cumbersome and hard to track.
Might I suggest another submission method?
Lol don't have time to do it but if you really wanted to get this done right you could code an algorithim that calculates average distance away from a street in my tracks... use that to generate a gps performance score... and use those gps performance scores to generate an average with standard deviations/ 95% confidence intervals.
That, would be so cool!
Sent from my SAMSUNG-SGH-I897 using XDA App

It would be great if there was so much data to be had that it was hard to track. Right now the situation is exactly opposite of that Bring it on! We can worry about managing the flood of data later, when there is a flood of data.
The problem with calculating an average distance like that is that it won't factor in enviornmental factors. The GPS might perform great with a 110 degree sky view and minimal multipath, with a 110 degree sky view and a mountain nearby, it might go bonkers. Due to the frequency range and transmission type, GPS signals are very prone to enviornmental factors. (These same weaknesses are also their strength in being so accurate in the right enviornment)
Updated first post with info about the apps I used.

This may count as anecdotal, but...
You asked for My Tracks data... something I have noticed is that, for me at least, My Tracks wil be pretty damb accurate if running in the background, while navigation or maps can't track well I the foreground at all. In fact, even My Tracks won't do well in the foreground. Not sure what this means, but liking forward to any developments on this front.
Sent from my SAMSUNG-SGH-I897 using XDA App

Dam Da_G great to see you. Long time since I have worked with wm but always enjoyed making roms from your leaked builds. I will work on collecting gps data. I do onsite IT work so I will log gps all day and post back when I get a few days of data.

Good to see you too shep Re: WM, The King is dead! Long live the King
Updated post 2 with some more interesting data regarding lock times. Re: My Tracks and performance in foreground/background, i'll slate that test for tomorrow. I think you're wondering if the CPU usage may have something to do with accuracy? It's certainly possible but if most of the calculations are handled in the baseband like I think, not likely. Either way, tomorrow's test will answer that

i'm not sure if this is the data you want but i matched all lbstestmode settings as yours, and did 3 position tests, took screen shots right after sat locked.
hope this useful...
here's my cappy's setup.
Rom : Perception Build 5
Kernel : SetiroN's v1.4.3 1350 Mhz OC'd
and here's the results
First Test
SV 22 : C/No 35 *
SV 06 : C/No 35 *
SV 03 : C/No 35 *
SV 09 : C/No 33 *
SV 19 : C/No 32 *
SV 14 : C/No 29 *
SV 21 : C/No 28
SV 27 : C/No 13
Second Test
SV 06 : C/No 37 *
SV 03 : C/No 36 *
SV 22 : C/No 35 *
SV 14 : C/No 33 *
SV 19 : C/No 32 *
SV 21 : C/No 30
SV 09 : C/No 28
SV 18 : C/No 11
Third Test
SV 03 : C/No 34 *
SV 22 : C/No 32 *
SV 06 : C/No 32 *
SV 19 : C/No 31 *
SV 09 : C/No 30
SV 14 : C/No 28 *
SV 21 : C/No 23
SV 18 : C/No 13
Screen Shots
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}

Perception Build 6 w/ JM9 gps files
Indoors - Currently clear skies
SV 18 : C/No 33 *
SV 12 : C/No 32 *
SV 22 : C/No 31 *
SV 25 : C/No 30 *
SV 09 : C/No 30 *
SV 14 : C/No 26 *
SV 21 : C/No 23 *
SV 27 : C/No 22 *
SV 15 : C/No 20 *
SV 31 : C/No 18 *

born_fisherman said:
Perception Build 6 w/ JM9 gps files
Seeing how well this works with xda app:
Sent from my SAMSUNG-SGH-I897 using XDA App
Click to expand...
Click to collapse
can't enlarge the image...

Great post, drancid. What was the enviornment like at the time? How far apart were the fix attempts? All on the same device?
I'm guessing it was indoors or with a partially blocked view of the sky?
@born_fisherman:
Looks like it was just a thumbnail that uploaded. No link to a larger photo, and the text is too small to read

Da_G said:
Great post, drancid. What was the enviornment like at the time? How far apart were the fix attempts? All on the same device?
I'm guessing it was indoors or with a partially blocked view of the sky?
Click to expand...
Click to collapse
thx.
yes, results from my cappy. just have one device.
about 2 minutes interval.
tested in my back yard. big tree and two story houses were near me but cappy was facing clear sky.

updated my post - The JM9 files work wonders. Oh, I also have SUPL Secure Socket ON

Ok, I'm indoors. Waited 5 minutes between tests. One device.
ROM - di11igaf's UGJK3, MODEM - JK3, KERNEL - SetiroN 1280 1.4.3
Gps Fixes - Jupiter V006 + JM9 Files
TEST 1
PRN: 25 CNo: 35
PRN: 31 CNo: 34
PRN: 22 CNo: 34
PRN: 18 CNo: 29
PRN: 14 CNo: 20
PRN: 30 CNo: 19
TEST 2
PRN: 25 CNo: 32
PRN: 22 CNo: 31
PRN: 31 CNo: 28
PRN: 30 CNo: 28
PRN: 24 CNo: 23
PRN: 18 CNo: 22
PRN: 14 CNo: 18

majinzen said:
Ok, I'm indoors.
ROM - di11igaf's UGJK3, MODEM - JK3, KERNEL - SetiroN 1280 1.4.3
Gps Fixes - Jupiter V006 + JM9 Files
TEST 1 TEST 2
PRN: 25 CNo: 35 PRN: 25 CNo: 32
PRN: 31 CNo: 34 PRN: 22 CNo: 31
PRN: 22 CNo: 34 PRN: 31 CNo: 28
PRN: 18 CNo: 29 PRN: 30 CNo: 28
PRN: 14 CNo: 20 PRN: 24 CNo: 23
PRN: 30 CNo: 19 PRN: 18 CNo: 22
PRN: 14 CNo: 18
Click to expand...
Click to collapse
Since you are using the JM9 files, use Supl.google.com / 7276 to get better results

I've tried both and I get better or equal results with this config. Sometimes supl.google.com takes forever to lock on for me.

See, that is really what I would like to find out more about the GPS - why it's different for others and why it isn't consistent!

Assonance 2.0 (stock = supl.google.com 7276, secure socket off, standalone, hot start, http://forum.xda-developers.com/showthread.php?t=856877)
#1007
sitting near a window with a clear view of the sky
TTFF 15s
Accuracy: 25.0 m
PRN 31 = 28*
PRN 23 = 22*
PRN 32 = 21*
PRN 24 = 21*
PRN 20 = 21*
PRN 11 = 21*
PRN 14 = 18*
PRN 30 = 17*
5 min later:
TTFF 15s
Accuracy 15.0 m
PRN 32 = 30*
PRN 20 = 29*
PRN 31 = 25*
PRN 24 = 25*
PRN 11 = 24*
PRN 30 = 20*
PRN 23 = 14
PRN 14 = 13
FWIW, this setup tracks very well on some car trips, less well on others.
What clues can you get from seeing our signal strengths, besides that it's erratic..?

[BOUNTY] NOW$50! Enable wifi channels 12, 13, 14 (wireless regulatory domain setting)

OK, hopefully this thread isn't in the wrong place... (If it is, feel free to flame and call me whatever, I'm just trying to get something done and I know all the devs watch this section)
I am in the US, and I am fully aware of the legalities of broadcasting on channels other than 1-11. I have several routers, running dd-wrt firmware, and three of the four are running as repeater bridges. Due to the extremely noisy wireless environment caused by neighboring networks, the only way I can get a clear signal between all of them is by setting the wireless channel to channel 14. Unfortunately, as soon as I do that, my Captivate no longer sees the network. This is exceedingly upsetting when my fiance shows me her iphone... which is still connected to the network!
I know it's not much, but I'm hoping that others will chip in or that some dev will take pity on me, but I've got $10 on a solution, paypal to whoever can make my Captivate connect to my wireless network on channel 14.
In hopes that others contribute as well, I'll make a list with total bounty:
Total: $15 - wifi channels 12, 13, 14 enabled
richhed - $10
ccdoggy - $5
BONUS!!!
Evidently, while we're messing with the wpasupplicant and wireless drivers, others have chimed in with an interest in the 5GHz wifi channels. Hopefully while tweaking the drivers that run the wireless radio to enable channels from other regions, a way to enable the 5GHz frequency will be discovered as well!
TOTAL: $35 - Enable 5GHz frequency wifi channels
ccdoggy - $5
Shammyh - $20 --- THANK YOU
DarkSi08 - $10

I never tested it, but the CyanogenMod 6.1 12-27 nightly definitely allows you to choose the channel (I saw 11-14 if I remember correctly), so if you run that you should be able to have the option to change it.
I know this isn't quite what you're looking for, but it's the only immediate solution that I know of.

root your phone and Install a custom ROM, than you need to choose a working modem.bin (european versions support those extra channels, please go to the modem thread to get more information)
the problem is: to enable 850MHz band on those modems... it is not impossible
good luck

Currently working on this
Code:
static struct ieee80211_channel __wl_2ghz_channels[] = {
CHAN2G(1, 2412, 0),
CHAN2G(2, 2417, 0),
CHAN2G(3, 2422, 0),
CHAN2G(4, 2427, 0),
CHAN2G(5, 2432, 0),
CHAN2G(6, 2437, 0),
CHAN2G(7, 2442, 0),
CHAN2G(8, 2447, 0),
CHAN2G(9, 2452, 0),
CHAN2G(10, 2457, 0),
CHAN2G(11, 2462, 0),
CHAN2G(12, 2467, 0),
CHAN2G(13, 2472, 0),
CHAN2G(14, 2484, 0),
};
static struct ieee80211_channel __wl_5ghz_a_channels[] = {
CHAN5G(34, 0), CHAN5G(36, 0),
CHAN5G(38, 0), CHAN5G(40, 0),
CHAN5G(42, 0), CHAN5G(44, 0),
CHAN5G(46, 0), CHAN5G(48, 0),
CHAN5G(52, 0), CHAN5G(56, 0),
CHAN5G(60, 0), CHAN5G(64, 0),
CHAN5G(100, 0), CHAN5G(104, 0),
CHAN5G(108, 0), CHAN5G(112, 0),
CHAN5G(116, 0), CHAN5G(120, 0),
CHAN5G(124, 0), CHAN5G(128, 0),
CHAN5G(132, 0), CHAN5G(136, 0),
CHAN5G(140, 0), CHAN5G(149, 0),
CHAN5G(153, 0), CHAN5G(157, 0),
CHAN5G(161, 0), CHAN5G(165, 0),
CHAN5G(184, 0), CHAN5G(188, 0),
CHAN5G(192, 0), CHAN5G(196, 0),
CHAN5G(200, 0), CHAN5G(204, 0),
CHAN5G(208, 0), CHAN5G(212, 0),
CHAN5G(216, 0),
};
static struct ieee80211_channel __wl_5ghz_n_channels[] = {
CHAN5G(32, 0), CHAN5G(34, 0),
CHAN5G(36, 0), CHAN5G(38, 0),
CHAN5G(40, 0), CHAN5G(42, 0),
CHAN5G(44, 0), CHAN5G(46, 0),
CHAN5G(48, 0), CHAN5G(50, 0),
CHAN5G(52, 0), CHAN5G(54, 0),
CHAN5G(56, 0), CHAN5G(58, 0),
CHAN5G(60, 0), CHAN5G(62, 0),
CHAN5G(64, 0), CHAN5G(66, 0),
CHAN5G(68, 0), CHAN5G(70, 0),
CHAN5G(72, 0), CHAN5G(74, 0),
CHAN5G(76, 0), CHAN5G(78, 0),
CHAN5G(80, 0), CHAN5G(82, 0),
CHAN5G(84, 0), CHAN5G(86, 0),
CHAN5G(88, 0), CHAN5G(90, 0),
CHAN5G(92, 0), CHAN5G(94, 0),
CHAN5G(96, 0), CHAN5G(98, 0),
CHAN5G(100, 0), CHAN5G(102, 0),
CHAN5G(104, 0), CHAN5G(106, 0),
CHAN5G(108, 0), CHAN5G(110, 0),
CHAN5G(112, 0), CHAN5G(114, 0),
CHAN5G(116, 0), CHAN5G(118, 0),
CHAN5G(120, 0), CHAN5G(122, 0),
CHAN5G(124, 0), CHAN5G(126, 0),
CHAN5G(128, 0), CHAN5G(130, 0),
CHAN5G(132, 0), CHAN5G(134, 0),
CHAN5G(136, 0), CHAN5G(138, 0),
CHAN5G(140, 0), CHAN5G(142, 0),
CHAN5G(144, 0), CHAN5G(145, 0),
CHAN5G(146, 0), CHAN5G(147, 0),
CHAN5G(148, 0), CHAN5G(149, 0),
CHAN5G(150, 0), CHAN5G(151, 0),
CHAN5G(152, 0), CHAN5G(153, 0),
CHAN5G(154, 0), CHAN5G(155, 0),
CHAN5G(156, 0), CHAN5G(157, 0),
CHAN5G(158, 0), CHAN5G(159, 0),
CHAN5G(160, 0), CHAN5G(161, 0),
CHAN5G(162, 0), CHAN5G(163, 0),
CHAN5G(164, 0), CHAN5G(165, 0),
CHAN5G(166, 0), CHAN5G(168, 0),
CHAN5G(170, 0), CHAN5G(172, 0),
CHAN5G(174, 0), CHAN5G(176, 0),
CHAN5G(178, 0), CHAN5G(180, 0),
CHAN5G(182, 0), CHAN5G(184, 0),
CHAN5G(186, 0), CHAN5G(188, 0),
CHAN5G(190, 0), CHAN5G(192, 0),
CHAN5G(194, 0), CHAN5G(196, 0),
CHAN5G(198, 0), CHAN5G(200, 0),
CHAN5G(202, 0), CHAN5G(204, 0),
CHAN5G(206, 0), CHAN5G(208, 0),
CHAN5G(210, 0), CHAN5G(212, 0),
CHAN5G(214, 0), CHAN5G(216, 0),
CHAN5G(218, 0), CHAN5G(220, 0),
CHAN5G(222, 0), CHAN5G(224, 0),
CHAN5G(226, 0), CHAN5G(228, 0),
};

THANK YOU DG! BTW, about to flash Cog3, Love your ROMs!

designgears said:
Currently working on this
Click to expand...
Click to collapse
get the 5 GHz frequency working (somehow?) and I'll be pleased. tired of my dual band router's N not being used.

For the bridges you might seriously consider using a 5 or 10mhz channel.. Dont know the bandwidth requirements but by using a smaller channel that is not one of the primary non overlapping channels you cut noise by almost 50%, thus allowing the bridges to get a clear signal to each other.. however normal 802.11 devices dont support < 20mhz channels, so if the bridges arent using a dedicated bridge link with 1 other ssid for the wireless clients, obviously there will be problems. I think dd-wrt will support the smaller channels (down to 5 mhz i believe) on certain hw configurations. The other option is to use a 5ghz bridge link and 2.4 client ap.
In extremely noisy situations, the only fix is a better antenna. The idea being, more power going out, more power to your devices, and with an antenna the devices signal gets amped back up too.. However the noise will also, but due to proximity the SNR at the AP should be within acceptable parameters for a decent link on 802.11n with a 20mhz channel. Do not use 40mhz channels this will only make it worse. For a good indoor solution in noisy environments (6 or more SSID's with a signal of -75 or better dbm) I'd get a Ubiquity Bullet with a 9-13dbi omni OR 16 dbm sector aimed in the direction that makes sense. This will increase usability enormously.

Wait... Does Dg' post mean the cappy supports 5 ghz wireless N?
Sent from my SGH-I897 using XDA App

No his post means he is working on channels 12-14, where does it ever say he is working on 5 ghz N??

silverslotcar said:
No his post means he is working on channels 12-14, where does it ever say he is working on 5 ghz N??
Click to expand...
Click to collapse
I think he just misunderstood my out of the park request (and the fact that the "code" quoted part of DG's post mentions a 5GHz part). the hardware in our phones doesn't support the 5 GHz frequency for wifi, so (unless the functionality is there and they just don't have it enabled?) that's just a pipedream in my mind that I jokingly requested.

There are 5 ghz channels listed in his post above. i thought that maybe enabling all of the 2.4 channels would also enable 5 ghz channels. I was just asking as from what i can tell from dg's post 5 ghz was listed. Yes i know the hardware needs to support it, it was worth just asking.
Sent from my SGH-I897 using XDA App

Kaik541 said:
I think he just misunderstood my out of the park request (and the fact that the "code" quoted part of DG's post mentions a 5GHz part). the hardware in our phones doesn't support the 5 GHz frequency for wifi, so (unless the functionality is there and they just don't have it enabled?) that's just a pipedream in my mind that I jokingly requested.
Click to expand...
Click to collapse
Exactly. Looking at it more i see that dg's post is just the regulatory domain regulations. 5 ghz is included because its he regulations for the country. My bad.
Sent from my SGH-I897 using XDA App

ccdoggy said:
Wait... Does Dg' post mean the cappy supports 5 ghz wireless N?
Sent from my SGH-I897 using XDA App
Click to expand...
Click to collapse
Im not 100% sure if it does or not, but the source is there for it. I am pretty sure the same driver is used on the models that do support it. That would be my guess..
But then I found that the Galaxy Tab has dual band support and runs the exact same chip..

Still at a whopping $10! Come on folks, I can't be the only person that needs this functionality? Let's get a few more pledges and make this worth DG's (or another dev's) while!

designgears said:
Im not 100% sure if it does or not, but the source is there for it. I am pretty sure the same driver is used on the models that do support it. That would be my guess..
But then I found that the Galaxy Tab has dual band support and runs the exact same chip..
Click to expand...
Click to collapse
Now, THAT would interest me. 5.0 Ghz just for the ability to do network scans for conflicting networks for that range also.
$5 for upper 2.4 channels
$5 for enabling 5.0 if possible

Isn't there anyone else who can throw some more money in on this and make it worth DG's (or someone else's) time? Anyone who might want to, say, travel to Europe or Japan, and be able to take their phone with them, and connect to these upper channels?
Sent from my SAMSUNG-SGH-I897 running Cognition 3.02, Firebird2 kernel, and i9000UGJK4 modem, using XDA App

I'd put up a solid $20 if someone figures out how to get the Captivate to work w/ 5 GHz 802.11a/n networks.

thekurrgan said:
For the bridges you might seriously consider using a 5 or 10mhz channel.. Dont know the bandwidth requirements but by using a smaller channel that is not one of the primary non overlapping channels you cut noise by almost 50%, thus allowing the bridges to get a clear signal to each other.. however normal 802.11 devices dont support < 20mhz channels, so if the bridges arent using a dedicated bridge link with 1 other ssid for the wireless clients, obviously there will be problems. I think dd-wrt will support the smaller channels (down to 5 mhz i believe) on certain hw configurations. The other option is to use a 5ghz bridge link and 2.4 client ap.
In extremely noisy situations, the only fix is a better antenna. The idea being, more power going out, more power to your devices, and with an antenna the devices signal gets amped back up too.. However the noise will also, but due to proximity the SNR at the AP should be within acceptable parameters for a decent link on 802.11n with a 20mhz channel. Do not use 40mhz channels this will only make it worse. For a good indoor solution in noisy environments (6 or more SSID's with a signal of -75 or better dbm) I'd get a Ubiquity Bullet with a 9-13dbi omni OR 16 dbm sector aimed in the direction that makes sense. This will increase usability enormously.
Click to expand...
Click to collapse
if the signal strength is a big factor then couldnt the op just set the antenna strength higher? i have 2 routers with dd-wrt one set as a repeater and the default antenna power setting is 70-71mw or something, they can both be turned up to 251mw. my entire block and a neighboring one can pickup my wifi if i leave it unsecured with the antenna power at max but i use the default settings, no need to broadcast free internet to 50+ people. i dont do that but my brother did with his router when he lived here, combined with some bigger antennas he used to share his wifi with his friend 3 houses up the road.

Perhaps change the country code in nvram_mfg.txt or nvram_net.txt and see what happens. The files are in /system/etc/wifi/:
Code:
# country code
ccode=EU
Change it to JP, since channel 14 is allowed in Japan.

Dani897 said:
if the signal strength is a big factor then couldnt the op just set the antenna strength higher? i have 2 routers with dd-wrt one set as a repeater and the default antenna power setting is 70-71mw or something, they can both be turned up to 251mw. my entire block and a neighboring one can pickup my wifi if i leave it unsecured with the antenna power at max but i use the default settings, no need to broadcast free internet to 50+ people. i dont do that but my brother did with his router when he lived here, combined with some bigger antennas he used to share his wifi with his friend 3 houses up the road.
Click to expand...
Click to collapse
Well yes, if it were that easy... And if I had about 30 spare WRT54G routers lying around... I appreciate your trying to help, but if you increase the Tx power above around 90mW (on Linksys WRT54G routers) you can plan on replacing it by next week or so. The board and the radio are inside a plastic housing that barely has adequate ventilation slots for the amount of heat produced by the radio at the stock Tx power levels, and I know from experience that just because the firmware allows it doesn't mean that you should do it...
Even if that wasn't the case, OR if I wanted to solder in a case fan on to of the radio chipset, the more you increase Tx Power, the more you increase noise. I once heard the analogy of a noisy wireless network being like a crowded concert hall. If you were sitting next to me in a concert hall, and we were having difficulty hearing one another, I could put my hands to my mouth, cup them around your ear, and scream at the top of my lungs right into your ear, and you would definitely hear me, but it's doubtful you would understand what I said (doubtful you would understand much of anything coming from that direction for the next ten minutes or more, actually). In the same fashion, if there lots of other wireless networks in range, increasing the Power of your network without trying a channel with less noise is actually counter-productive. The more noise, the worse your SNR, and SNR in any connection IS your effective link quality. I have 8, (yes, 8) other networks in range of my primary AP (gateway), all range between channels 1 and 9... There are two other networks on channel 11 in range of my first repeater, and then one more on channel 6 in range of the second repeater. I wish I could solve the signal quality issue as easily as cranking the volume up to the max, but when I increase it even past 84mW my clients in range of the primary AP start getting packet loss and terribly intermittent connectivity loss.
___________________
ppmz said:
Perhaps change the country code in nvram_mfg.txt or nvram_net.txt and see what happens. The files are in /system/etc/wifi/:
Code:
# country code
ccode=EU
Change it to JP, since channel 14 is allowed in Japan.
Click to expand...
Click to collapse
I appreciate you trying to help out, I was thinking along the same lines at first. I have tried editing the settings.db file, changed the wifi_num_allowed_channels to 14 from 11, I edited the build.prop ro.wifi.channels also from 11 to 14, and I changed the ro.product.locale.region from US to JP. I also changed the country code to JP in both of the system/etc/wifi/ files you are referring to, all to no avail... Trust me, if it was something that easy (simple sqlite .db editing with root explorer, move back to /dbdata/ and chown system system, fix permissions... I'd have been posting a fix rather than a bounty. It seems to be deeper-rooted than that... Seems to be related to the now missing wireless regulatory domain setting that was once present in the settings for android devices, but has apparently been removed in (eclair? donut?) and newer OS revisions.
Thank you all for your input, and for your attempts at a solution. I am sure that if we keep at it we'll find the solution, and as of right now the first person who does will be the recipient of a whopping $40!
Sent from my SAMSUNG-SGH-I897 using XDA App

[REF]Analysis of build.prop cellular speed tweaks

Introduction:
This thread suggests several build.prop entries that supposedly increase your 3g speeds:
[mod]Boosting Signal Strength (hspa/gprs)
After doing nearly 100 benches with the speedtest app, I have come to a couple conclusions:
The server your connect to for speedtest has the greatest impact on your speeds (network congestion)
The majority of the values given dont even exist by default on the device
Your battery charge might have more influence over speeds then any tweaks
Device defaults:
Code:
ro.ril.hsxpa=2
ro.ril.gprsclass=12
ro.ril.hep=" "
ro.ril.enable.dtm=" "
ro.ril.hsdpa.category=" "
ro.ril.hsupa.category=" "
ro.ril.enable.a53=" "
where " " means that the value isnt defined.
ro.ril.hsxpa: Defines weither to use UMTS, HSPDA/UMTS, or HSDPA/HSPUA respectively.
ro.ril.gprsclass: Defines what class speeds to use with edge. (12 is maximum defined in standard?)
ro.ril.hep: unknown and undefined
ro.ril.enable.dtm: Dual transfer mode (only relevent for EDGE connections, likely always enabled for devices built after 2010)
ro.ril.hsdpa.category: class 8 matches the max download rate
ro.ril.hsupa.category: class 6 matches the max upload rate
ro.ril.enable.a53: supposedly related to GSM security, no actual info (may not even exist)
Anything not listed is completely irrelevent.
Test setup:
PC on wireless-n
Streak 5 on wireless-g
Streak 5 switched to HSPA
ROM: CM7.2-rc0
Kernel: Phoenix kernel for CM7.2
BB: 406-2 (407)
Testing methodology:
Test 5/10 times on control
Make change via terminal and setprop
Enter and exit airplane mode
Do run
Undo mod and do control
Repeat for all
Do again, but make change via build.prop edits
Restart 3 times
Do run
Do control tests
Controls:
PC over landline:
Code:
Ping Down Up
77 2.18 4.8
63 2.11 7.98
97 2.49 5.85
63 3.07 7.6
92 3.26 7.55
49 2.71 7.17
50 2.54 7.49
55 2.8 7.28
65 2.39 7
40 2.51 4.91
Phone over wifi over landline:
Code:
Ping Down Up
90 2.91 7.06
54 3.23 6.53
84 3.38 7.73
72 3.7 9.03
51 4.04 7.62
Phone over 3g while charging:
Code:
Ping Down Up
146 0.45 0.24
786 0.29 0.18
95 0.29 0.23
966 0.28 0.29
114 0.63 0.32
886 0.27 0.15
100 0.27 0.19
837 0.38 0.34
826 0.44 0.51
896 0.21 0.56
ro.ril.hsdpa.category=8
Code:
96 1.58 1.24
106 1.56 1.56
911 1.55 1.35
101 1.96 1.82
916 1.63 1.41
Control after testing above:
Code:
125 0.48 1.01
914 0.27 0.61
110 0.69 0.9
146 0.17 0.18
108 1.1 0.73
ro.ril.hsupa.category=6
Code:
118 0.86 1.45
807 0.73 1.23
116 0.94 1.16
896 0.95 1.12
102 0.93 0.94
Control after testing above:
Code:
778 0.8 1.24
792 0.79 1.36
124 0.75 1.28
828 0.72 1.21
109 0.72 1.31
ro.ril.hsdpa.category=8 and ro.ril.hsupa.category=6
Code:
97 1.16 1.42
932 0.81 1.42
98 0.86 1.41
876 0.88 1.33
829 0.8 1.21
Control after testing above:
Code:
778 0.8 1.24
792 0.79 1.36
124 0.75 1.28
828 0.72 1.21
109 0.72 1.31
ro.ril.hsdpa.category=8 and ro.ril.hsupa.category=6 and ro.ril.hsxpa=5
Code:
126 1.16 1.41
824 0.87 1.08
795 1.08 1.06
127 0.98 1.05
116 0.73 1.18
Control after testing above:
Code:
1019 0.56 1.19
789 0.75 0.64
117 0.46 0.92
895 0.68 1.27
916 0.31 0.46
Conclusion:
A google search shows that ro.ril.enable.a53 may not even exist besides self references to it's own speed tweaks, it may very well not even exist.
The same can be said about ro.ril.hep.
ro.ril.enable.dtm may be irrelevent as DTM allows you to use voice and data at the same time, and this technology only applies to edge.
This is likely implicitly always enabled for any device that supports it past 2010, as only the earliest 3g devices may not have had support for it.
ro.ril.hsxpa=2 and ro.ril.gprsclass=12 are likely already correctly definied for it (with it being a HSPA device)
Bar ro.ril.hsdpa.category=8 and ro.ril.hsupa.category=6, the rest of them are not valid to begin with.
I was able to achieve much higher speeds by switching servers, and all tests were intentionally done on the same server for consistancy.
Those values do canonically exist on other devices, but they're never defined on the S5.
Just as much my speeds varied wildly even from control to control, as much as doubling from the lowest common to highest common.
ro.ril.hsdpa.category and ro.ril.hsupa.category might potentially give an increase in speeds, but as they dont even have default values, they might not even be valid.
I cant even reproduce some of the results from enabling only ro.ril.hsdpa.category=8, as I restart the device after each change to be sure.
I may have inadvertantly connected to a faster server for that one instance.
tl;dr: you get as much speed increases from CHANGING THE THEME ICON TO DISPLAY 4G, you can also make stuff up and pretend it does something and let the placebo effect do all the work for you.
ro.ril.hsdpa.category=8 and ro.ril.hsupa.category=6, might do something, but they could very well be invalid.

thank you, great work
I'll get right on that! I'll make a sticker for my DS5 that says its 4G LTE+ ;P

Thankyou for the thorough testing, now what about vmheap?
Sent from my Dell Streak using Tapatalk 2

I dont intend on testing it, it's default size never changes for any device, regardless of ram size. I assume goog knows what they're doing in respect to that, as manus can and do add in their own optimizations into their roms, and they still leave those values alone more or less.

Phone not staying connected to cell service

I need help figuring out why my phone keeps dropping connection and fixing it.
The phone will suddenly kill its connection (the icon for the connection drops the data indictor H, or LTE) whenever I start an app that needs data, like iHeartRadio. Also I can't get LTE anymore. Yesterday it was on LTE all day with no trouble, but a restart seems to have broken it.
The phone is an rma phone I just received. I'm on pureNexus and will post there too.
What can i do to find more information? I tried looking through logcat, but i don't know what to look for.

Personally I would go back to stock by flashing the full stock image with flash-all.bat command and see if the issue persists. If it does your RMA phone probably has a hardware issue. You could also borrow someone's sim chip and see if the issue persists.

jhs39 said:
Personally I would go back to stock by flashing the full stock image with flash-all.bat command and see if the issue persists. If it does your RMA phone probably has a hardware issue. You could also borrow someone's sim chip and see if the issue persists.
Click to expand...
Click to collapse
Back to stock, it's persisting.. I looked in catlog and see this:
05-04 09:11:01.380 1431 3560 D ConnectivityService: releasing NetworkRequest [ REQUEST id=28, legacyType=3, [ Transports: CELLULAR Capabilities: SUPL&NOT_RESTRICTED&TRUSTED&NOT_VPN] ]
05-04 09:11:01.380 1431 1443 I GnssLocationProvider: WakeLock released by handleMessage(15, 2, null)
05-04 09:11:01.381 1431 3560 D ConnectivityService: Sending DISCONNECTED broadcast for type 3 NetworkAgentInfo [MOBILE (HSDPA) - 111] isDefaultNetwork=false
05-04 09:11:01.411 13811 13811 D MusicLifecycle: com.google.android.music.net.NetworkConnectivityMonitor$NetworkChangedReceiver generated event: Broadcast received with context [email protected] and intent Intent { act=android.net.conn.CONNECTIVITY_CHANGE flg=0x4000010 (has extras) }
05-04 09:11:01.419 3863 3863 D MccTable: updateMccMncConfiguration mccmnc='null' fromServiceState=true
05-04 09:11:01.419 3863 3863 D MccTable: WIFI_COUNTRY_CODE set to
05-04 09:11:01.419 1431 14829 I WifiService: WifiService trying to set country code to with persist set to false
05-04 09:11:01.426 18316 18316 V BugleThrottledAction: CountryCodeDetectorAction scheduled to run now.
05-04 09:11:01.426 18316 18316 D BugleThrottledAction: CountryCodeDetectorAction already scheduled in 0ms, not scheduling it again.
05-04 09:11:01.426 14746 14746 D HeadsetPhoneState: sendDeviceStateChanged. mService=0 mSignal=0 mRoam=0 mBatteryCharge=4
05-04 09:11:01.426 14746 14746 D HeadsetPhoneState: sendDeviceStateChanged. mService=0 mSignal=0 mRoam=0 mBatteryCharge=4
05-04 09:11:01.434 1431 1431 I GnssLocationProvider: WakeLock acquired by sendMessage(12, 0, null)
05-04 09:11:01.434 5977 5977 D RcsService: Never provisioned and provisioning is not allowed, skipped config update request
05-04 09:11:01.438 13811 13811 I NetworkConnectivity: Network state changed: [type: MOBILE[HSDPA], state: CONNECTED/CONNECTED, reason: connected, extra: iijmio.jp, failover: false, available: true, roaming: false, metered: true]
05-04 09:11:01.451 18316 18316 V BugleThrottledAction: SelfParticipantsRefreshAction scheduled to run now.
05-04 09:11:01.465 3863 3863 I Telephony: PstnIncomingCallNotifier: Unregistering: Handler (com.android.internal.telephony.GsmCdmaPhone) {4be86a2}
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: GPS configuration:
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: #Thu May 04 09:11:01 GMT+09:00 2017
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: NTP_SERVER=north-america.pool.ntp.org
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: INTERMEDIATE_POS=0
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: A_GLONASS_POS_PROTOCOL_SELECT=0
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: USE_EMERGENCY_PDN_FOR_EMERGENCY_SUPL=1
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: DEBUG_LEVEL=3
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: SUPL_VER=0x20000
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: NMEA_PROVIDER=0
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: SGLTE_TARGET=0
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: CAPABILITIES=0x37
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: ERR_ESTIMATE=0
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: XTRA_VERSION_CHECK=1
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: SUPL_PORT=7275
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: LPP_PROFILE=0
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: XTRA_SERVER_3=https\://xtrapath3.izatcloud.net/xtra3grc.bin
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: SUPL_MODE=1
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: XTRA_SERVER_2=https\://xtrapath2.izatcloud.net/xtra3grc.bin
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: XTRA_SERVER_1=https\://xtrapath1.izatcloud.net/xtra3grc.bin
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: SUPL_HOST=supl.google.com
05-04 09:11:01.476 1431 1443 D GnssLocationProvider: SUPL_ES=0
05-04 09:11:01.476 1431 1443 I GnssLocationProvider: WakeLock released by handleMessage(12, 0, null)
05-04 09:11:01.479 1431 3814 E LocSvc_ApiV02: E/virtual enum loc_api_adapter_err LocApiV02::setServer(const char *, int):1027]: error status = eLOC_CLIENT_SUCCESS, set_server_ind.status = eQMI_LOC_ENGINE_BUSY_V02
05-04 09:11:01.485 14517 14517 D CellBroadcastReceiver: onReceive Intent { act=android.intent.action.SERVICE_STATE flg=0x10 cmp=com.android.cellbroadcastreceiver/.CellBroadcastReceiver (has extras) }
05-04 09:11:01.485 14517 14517 D CellBroadcastReceiver: Intent: android.intent.action.SERVICE_STATE
05-04 09:11:01.486 14517 14517 D CellBroadcastReceiver: Service state changed! 1 Full: 1 0 voice home data home null null null null null null Unknown HSDPA CSS not supported -1 -1 RoamInd=-1 DefRoamInd=-1 EmergOnly=false IsDataRoamingFromRegistration=false IsUsingCarrierAggregation=false Current state=0

Your RMA phone likely has a hardware issue. You could try flashing a different radio. Some people have claimed to get better coverage with radio 3.75 than with the newer radios. I had a similar problem but it had something to do with AT&T's network coverage in my area. When I switched carriers the problem went away.

jhs39 said:
Your RMA phone likely has a hardware issue. You could try flashing a different radio. Some people have claimed to get better coverage with radio 3.75 than with the newer radios. I had a similar problem but it had something to do with AT&T's network coverage in my area. When I switched carriers the problem went away.
Click to expand...
Click to collapse
I think I fixed it, I just removed my username and password from my APN settings

Firestick 4k no wifi scan

Firestick 4k stopped working on wifi for no apparent reason... (apparent to me )
After I had to disconnect my access point for some time (Cisco 2702) the firestick never worked. No settings changed at all. I just changed the POE cable to the AP so nothing related to that.
I did several restarts of the firestick, the AP, removed the DHCP lease from the router...
Ive seen on the forum that people had problems with some 5ghz channels so I checked both lower channels and hardcoded 104 on the AP.
But it seems that the firestick wont display any available wifi stations (Not even from neighbours that run default setting and basic APs). On the menu it just gives me the option to "troubleshoot" by rebooting the firestick.
I connected an Eth and got some logs but I dont know how to get logs without an ethernet connection so Im not sure if the wifi logs below are not relevant with the problem.
01-26 08:13:28.720 4254 4254 I ScannerController: allowWifiManagerScanning() calling WifiManager.stopReconnectAndScan(false, 0)
01-26 08:13:28.723 280 412 I DolbyAudioStreamOut: getLatency() Total latency:187ms (Downstream:104ms, DMA:158ms, tuning:-75ms, tuningBT:0ms), input format:0x1, sink:0x400, UxOutSetting:0xa000000, CurOutFormat:0xa000000
01-26 08:13:28.724 611 620 I art : Background sticky concurrent mark sweep GC freed 46422(3MB) AllocSpace objects, 1(20KB) LOS objects, 0% free, 25MB/25MB, paused 7.810ms total 64.883ms
01-26 08:13:28.725 4254 4254 I NetworkManager: isInternetLastKnownUp: true
01-26 08:13:28.732 280 412 I DolbyAudioStreamOut: getLatency() Total latency:180ms (Downstream:97ms, DMA:158ms, tuning:-75ms, tuningBT:0ms), input format:0x1, sink:0x400, UxOutSetting:0xa000000, CurOutFormat:0xa000000
01-26 08:13:28.732 280 412 I DolbyAudioStreamOut: getLatency() Total latency:179ms (Downstream:96ms, DMA:158ms, tuning:-75ms, tuningBT:0ms), input format:0x1, sink:0x400, UxOutSetting:0xa000000, CurOutFormat:0xa000000
01-26 08:13:28.733 611 662 D WifiStateMachine: CMD_SET_WIFI_DONT_RECONNECT_AND_SCAN { when=-11ms what=131247 target=com.android.internal.util.StateMachine$SmHandler }
01-26 08:13:28.733 611 662 D WifiStateMachine: reset dont_reconnect_scan flag
01-26 08:13:28.733 611 662 D WifiStateMachine: mDontReconnectAndScan=false
01-26 08:13:28.733 611 662 E WifiStateMachine: Error! unhandled message{ when=-12ms what=131247 target=com.android.internal.util.StateMachine$SmHandler }
01-26 08:13:28.738 4254 6381 I ST-NetworkHandler: Fetching Configured Networks
01-26 08:13:28.738 4254 6381 I WifiManager: Wifi getConfiguredNetworks called by com.amazon.tv.settings.v2
01-26 08:13:28.750 4254 4254 E SettingsLayoutFragment: setLeftPaneTitle - getView() returned null
01-26 08:13:28.750 4254 4254 E SettingsLayoutFragment: setLeftPaneTitle - failed to get left pane title text view
01-26 08:13:28.753 4254 4254 I ST-NetworkActivity: Ethernet is connected, only showing ethernet configure setting.
01-26 08:13:28.770 280 412 I DolbyAudioStreamOut: getLatency() Total latency:182ms (Downstream:99ms, DMA:158ms, tuning:-75ms, tuningBT:0ms), input format:0x1, sink:0x400, UxOutSetting:0xa000000, CurOutFormat:0xa000000
01-26 08:13:28.772 611 631 V MouseModeService: packageName=com.amazon.tv.settings.v2 activityName= mouseModeEnabled=false
01-26 08:13:28.773 4254 6381 I ST-NetworkHandler: Fetched configured networks.
01-26 08:13:28.777 280 1198 D DolbyAudioHwDeviceProxy: setParameters(0xa9bd9000): continuous_audio_mode=0
Any ideas?