All things GPS

[George0211]

How positive are you with this info? Are the other cell providers getting updated hardware? Can anyone check on this?

That's because Samsung NEVER admited that GPS was broken. They've always said something along the lines of "Some People have problems with GPS accuracy" or something like that.
Hence, they are using the word "optimize".
Also, hardware being defective is just a speculation, and Samsung has said on many ocasions that it is software.
Also, AT&T and Samsung statements at first DID say that GPS fix was included with this update. I even spoke with level 3 tech rep from Samsung the other day and he said this update IS the GPS fix and no more updates are coming.
Of course now, people are reporting that Samsung has changed their story and supposedly another update is coming. I spoke with Level 1 person today, who assured me that another update is coming but as always there is no specific date for it.

 

[boomerbubba]

Good information seemingly has a difficult time breaching what appears to be device specific defensiveness.

But then again, no reason to sticky a workaround for a non-existant problem either .

Anyone who believes the GPS problem is non-existent is a noobie, or just hasn't been paying attention.

The bug, whose root cause is the mishandling of time-dependent GPS cache data, is present in every Epic 4G phone. It just does not occur for some users sometimes because of their particular usage patterns.

 

[Thuneau]

It looks like he didn't have to, it was already there, which either makes him very lucky or idk what welcome to the lost data club Thuneau, btw, how's the GPS now ?

The GPS Test app found and locked on to 10 satellites within seconds, so that was encouraging.
But, the directions app as I took a quick walk around the block was not very reliable. Had the arrow pointed the opposite way half of the time- as I walked north, it pointed south. Also, it lost the satellite sync couple times.
So my first impression is, GPS is partially fixed.
I think the location and map serving service is not configured right. GPS drivers/firmware are improved.

 

[Andyman10]

@Mr GPS

"This means AGC is adaptive to the patterns of GPS signals based on the settings of start point/duration/sampling interval on the AGC. Another factor we need to consider is that the perceived SNR values are from GPS TEST driver's interpretation, not from Samsung GPS driver's interpretation. "

I have had a reply from the author of GPS Test as follows:

"The app uses the standard Android GPS listener APIs to get the data (these are Java calls)."

So it's going to be seeing SNR data provided by the Samsung/Broadcom driver to the API, as will any other application using the SDK.
Cheers

 

[GaryStebbins]

My GPS is unusable. It takes forever to lock on satellites and will not maintain a lock. My previous cell phone (a Samsung Windows phone) was reliable and could get a lock quickly inside a car and maintain that lock.

I just spoke with Samsung (phone number given to me by AT&T support), and the customer service rep I spoke with confirmed that the GPS fix was not included in the JH7 update. They are still working on a fix. He said it will "hopefully be available next month."

For those of you that state your GPS is inaccurate because it shows you wandering from the sidewalk to the middle of the road... That's not bad. With a decent dedicated GPS, my accuracy is usually 15-30'. Unless you're walking next to a very wide road, you would expect to sometimes appear to be in the middle of the road... And I don't expect the GPS in a cell phone to better a dedicated GPS. If it just showed my position to be on the right road, I'd be happy...

 

[RickJ]

Anyone who believes the GPS problem is non-existent is a noobie, or just hasn't been paying attention.

The bug, whose root cause is the mishandling of time-dependent GPS cache data, is present in every Epic 4G phone. It just does not occur for some users sometimes because of their particular usage patterns.

NOOOOoooooooo, you forgot emotional defensivness, or ostrich syndrome in the choice list.

Seriously though, simply because I own an evo, (nevermind that Im patiently awaiting an epic pending a resolution to this and the other issues) I'm expecting to be called out as a troll soon, so I shall be on my merry way

 

[Aero1]

to get on topic, the Epic has several known GPS bugs. The main one's symptoms are widely reported. The reason why some don't experience the main GPS bug present on their Epic is due to usage habits. Lots of bugs present on every single handset of a given make/model do not manifest for everyone due to different usage habits.
We all have the bug, some may see no impact. this workaround it for those having the impact from the bug we we all have.

There is also built in masking of some symptoms becasue of the Harding of the accuracy rate, something Epic objectively incorrectly reports to programs (asd you do have that otehr GPS bug as well) , but which allows some programs to work even under accuracies that are low.

If my Gps works without using this fix, should I even bother with it? Don't want to mess anything up...

There are people saying they get over 200k on 3G uploads as well,a s well as people whose habits see not upload problem. They should also not apply a fix to that.

 

[redguardian]

Mine works perfect (without any fix) and im no noob either. I have done random tracking on my phone using Tasker and it always locates at a radius of about 20 yards.
Its pretty good in the car too.

 

[Othan1]

Mine works perfect (without any fix) and im no noob either. I have done random tracking on my phone using Tasker and it always locates at a radius of about 20 yards.
Its pretty good in the car too.

My last Epic was always spot on regardless whether I rebooted or not. I could never understand why people thought there was a problem. I was continually called a Troll and other names just because I had a phone that appeared to work. I ignored them, of course.

When I got this new one (returned due to loose screen), it just refuses to lock at all when I'm using CardioTrainer. No matter what I did, reboot or not, it will not connect. So maybe a tiny fraction of phones actually work (they were the batch tested by QA?) but they are by far the exception and not the norm.

I just did the workaround provided and it appears to actually work. CardioTrainer is locking on almost immediately. I will try it more later today when I go out for my hike and see if it stays on.

 

[NJR]

Mr. GPS I am copying this from another thread as it seems perintent here..do you concur with my views:

I have come to the conclusion that the GPS reception is just not as strong as others. I had a BB Bold on my front seat running ATT Nav. I had the Captivate with update, on the dashboard running Google Nav for 30 minute trip. As long as the Captivate stayed on the dash my navigation was pretty good although there were a couple of drops. When I put it on the seat next to the BB it lost lock very often. The BB never dropped. That seems to me to indicate a week GPS antenna system-- hence a hardware problem so I think, although as others have said the GPS is now optimized, and does perform better, I don't think it can ever be as good as say a BB or other smart phone. I hope I am proven wrong.

 

[s.m.knipe]

Also, AT&T and Samsung statements at first DID say that GPS fix was included with this update.

Actually, on the night between September 21st- 22nd when the update was revealed, and most of the morning on the 22nd, the info page on AT&T's website, and the press release used by the original articles/blogs written about it said nothing about the GPS fix. They listed the Quickoffice update, the media scanning, exchange 2003, MMS fix, and some other fixes, but NOTHING about the GPS, even on this very forum check out the comments prior to Samsung's Sept. 23rd announcement via facebook that the update fixes the GPS such as:

http://androidforums.com/samsung-captivate/181321-official-ota-update-discussion.html#post1622212

09-21-2010, 10:25 AM #9 (permalink)
emuneee
Member

Join Date: May 2010
Posts: 166

Device(s): Samsung Captivate (Galaxy S)
Thanks: 9
Thanked 23 Times in 19 Posts
Samsung Captivate Software Update to Rollout Today?
Saw this posted at XDA. This looks official.

Answer Center | AT&T

Quote:
09/21/2010 - Firmware Over the Air (FOTA) software update changing software version from I897UCJF6 to I897UCJH7. A 12MB download is being pushed to Samsung Captivate owners starting on 09/21/2010 and will continue Tuesday through Friday at 15K customers per day until complete. This is NOT an operating system update. This update includes the following software updates:
Preinstalled software - QuickOffice for Android will allow users to edit various Microsoft Office documents.
Messaging - Exchange 2003 support, MMS and other messaging enhancements.
Voice - Improved calling experience.
Media - media scanning time improvements.

Sadly enough, no mention of any type of GPS fixes...smh.


or one of the first articles that broke about the update:

Samsung Captivate OTA update begins rolling out - San Francisco Gadgets | Examiner.com

Also consider that this video and article came out prior to Samsung's "official" ( and more like hopeful) announcement

Here is AndroidCentral's own article, again listing no GPS fixes announced:
Update coming this week for the Samsung Captivate | Android Central

and here is the original and official press release from AT&T:
Answer Center | AT&T

But it was modified on Sept. 22nd to include the last line of "GPS performance improvements" (notice the lack of the word "fix"). The original did not include that line, and was actually laid out in reverse order than it is in that link. The update information was at the bottom of the page with the instructions for checking firmware version number above it. I do not have a cached image of the page, but will try to find one...

I'm sorry to burst your bubble, but nothing was said about GPS fixes surrounding the update until after that careace video and a few other blog posts announced that it was surprisingly fixed, and miraculously the next morning, Samsung announces that it is the fix.

 

[Mr. GPS]

@Mr GPS

"This means AGC is adaptive to the patterns of GPS signals based on the settings of start point/duration/sampling interval on the AGC. Another factor we need to consider is that the perceived SNR values are from GPS TEST driver's interpretation, not from Samsung GPS driver's interpretation. "

I have had a reply from the author of GPS Test as follows:

"The app uses the standard Android GPS listener APIs to get the data (these are Java calls)."

So it's going to be seeing SNR data provided by the Samsung/Broadcom driver to the API, as will any other application using the SDK.
Cheers

Right, I think they use getSnr() method in Class GPSSatellite to get SNR in standard Android GPS API
GpsSatellite | Android Developers

After we have clarified this factor. we can design test schemes to figure
out what conditions that AGC/GPS chip firmware/GPS driver will get fooled to perceive (or compute) low SNR values. But they shouldn't do
so and can be avoided by using smart policies or procedure.

The conditions of other component interferences can be from either
external(my hands cover case) or internal (your flight mode case).

 

[boomerbubba]

Mine works perfect (without any fix) and im no noob either.

Well then you may be in the other class of users I mentioned -- those who haven't been paying attention to the forums. Have you tested using the scenarios that have been documented to replicate the bug? (If you have to ask what those scenarios are, that clinches it.)

 

[George0211]

I'm sorry to burst your bubble, but nothing was said about GPS fixes surrounding the update until after that careace video and a few other blog posts announced that it was surprisingly fixed, and miraculously the next morning, Samsung announces that it is the fix.

They changed their story litteraly within a day or 2. Once people started calling AT&T and Samsung, everyone's been getting the same answer, that this is the fix.
Then a few days later they changed their story again.
So, you're not bursting my bubble, my point is that we didn't get what we were told, because the story keeps changing every day.

 

[s.m.knipe]

Ah, I see. Yeah, it is weird that the story seems to change not only from day to day, but also from who is told what. this is totally fixable for them, as I am sure it is a Firmware and driver issue, but it really worries me that it is taking this long. The reason I am so sure is because I have tested the Captivate repeatedly side-by-side with my N1 and G1, and they show very similar SNR values for the satellites seen, the Captivate simply takes forever to lock, whereas the N1 is nearly instantaneous and the G1 is usually under 30 seconds. I am beginning to wonder about the inferior I/O file system used on the Galaxy S phones, and if that is impacting it...

 

[Andyman10]

Right, I think they use getSnr() method in Class GPSSatellite to get SNR in standard Android GPS API
GpsSatellite | Android Developers

After we have clarified this factor. we can design test schemes to figure
out what conditions that AGC/GPS chip firmware/GPS driver will get fooled to perceive (or compute) low SNR values. But they shouldn't do
so and can be avoided by using smart policies or procedure.

The conditions of other component interferences can be from either
external(my hands cover case) or internal (your flight mode case).

Hi,

Just as an aside ..since setting up the GPS with wireless off it has performed much better (no power-off in between) and I'm getting 8 in view, 6 in use even indoors in my computer room, and there's lots of background noise there! I am no doubt being very simplistic here but having seen issues like this before in demodulator and DSL design that were due to something simple like incorrect AGC setup and tracking I have real hopes that it is at least partly fixable in firmware. Still a pretty poor design I suspect, but then as I said before these guys have to cram everything but the kitchen sink into a tiny area and with an equally tiny power budget; not easy.

 

[askpcguy]

I just saw this over on xda-developers forum, seems like the guy done quite a bit of leg work. [GPS/AGPS] REAL solution analysis thread (No AGPS - Jupiterv002!) - xda-developers

 

[toneflem]

mine has been pretty accurate. heres my map from coming back from campus.
Purdue to home - Google Maps

 

[armeck]

Here's my before the update:

And after:

Definitely better... but still not too accurate.

 

[simulator]

One of my coworkers just got a captivate on Friday. He walked in my office today, and he said: "Great phone. But what the f*** is the problem with the GPS?"

We really should get a good class action lawsuit going...

 

[Mr. GPS]

I. One common thing you have to do with the host processor assisted type designs that are being widely used now is control, via a Java or direct API, the tracking of the satellite signal being received and the initiation of the AGC or Auto Gain Control circuits. Basically a teeny weeny signal from the satellite comes in to the GPS
Cheers

Need a more clear picture of possible GPS update from Samsung in future.

From the surface, we know that Samsung GPS driver did not implement correct and high performace of Android GPS API for location-based applications. Some errors are purely at level of GPS software driver. Some like getSnr() may go down to the GPS firmwire and/o chip hardware.
For the Java & Direct API settings of AGC stuffs you mentioned, I can image the possible cases in Captivate phones

1. Hardcoded settings of start point/duration/sampling interval on
the AGC circuits

2. In Bradcom GPS firmware, GPS driver or phone system
(system firmware) is responsible to do the settings

3. In Bradcom GPS chip firmware, a separate application is
responsible to initialize it with proper settings.This application
from Broadcom may not run under Android environment.

For case 1, the phones need a recall. For case 2, this requires phone system firmware or GPS driver software update.

For case 3, what is a normal practice for changing this GPS firmware?

 

[Terrigno]

Hahahaha.

Here is my tracks from today, after 4 days of GPS not working once -

lol. Good thing I know how to count. "Hey, what street do you live on?" "Im on 1, 967th street". "OH, ok!" lol

 

[s.m.knipe]

That actually isn't too bad. If you look at real tracks from any good old fashioned GPS unit, that is what the raw data usually looks like when the unit is stationary. They then employ tricks like averaging and snapping to roads etc, to make things seem more "normalized." I'm not saying the Captivate GPS doesn't suck, just that kind of track is to be expected.

My Garmin GPSMap 76 doesn't do that...

 

[Thuneau]

It looks like he didn't have to, it was already there, which either makes him very lucky or idk what welcome to the lost data club Thuneau, btw, how's the GPS now ?

I think I figured out the problem. Before doing the factory reset you should undo the one-click lag fix or you lose all data- there is a little paragraph about it during installation of OCLF.
I had the lag fix and did not undo it before applying the reset.
So now I'm screwed.
I can't even get the OCLF to work again- it tells me it's installed but my quadrant score is back to 900's- it was 2200 before.
I might have to get Odin and re-flash everything and start from scratch.

 

[s.m.knipe]

And I don't expect the GPS in a cell phone to better a dedicated GPS. If it just showed my position to be on the right road, I'd be happy...

I thought I read somewhere that the same Broadcom chip Samsung used for the Galaxy S's is used in TomTom's most recent units?

If so, shouldn't performance be the same?- provided everything is coded correctly on the firmware end?