-
After 15+ years, we've made a big change: Android Forums is now Early Bird Club. Learn more here.
All things GPS
- Thread starter enjoimark
- Start date
-
- Tags
- samsung captivate
tylerdurdin
Android Enthusiast
Now if only those "benchmarks" actually meant anything.
(Quadrant scores on any phone with a "lagfix" hack are absolutely meaningless at that point. It's the equivalent of putting your car in neutral, revving the RPMs to 4000 and thinking you're going "fast".
That's really not fair and kind of untrue. If you are seeing real world results how is it not a legit fix? I agree that it is a hack, but if Samsung implemented it from the release of the phone you would be on here telling people that no phone compares to the galaxy on Quadrant. If you read about how it works it is nothing short of brilliant.
I think it is fair to say that Quadrant can be gamed - sometimes, it's accurate, other times a kernel dev can tinker for meaningless scores. I mention this because I exchanged with kernel dev on this, and saw firsthand on my phone that happening with some test kernels.
Bottom line is benchmarks are of value - but limited value. Android Forums just recently cooperated with another popular web site for some Android benchmarking. I've exchanged with the author and still learned some things about these benchmarks when I thought I was all squared away (I used to write mainframe benchmarks, fwiw).
Those results will published in the coming weeks, and Phandroid.com will be getting first crack at those other-site summaries. I'll try to remember to post that all back here, or a similar thread, when the time comes.
Bottom line is benchmarks are of value - but limited value. Android Forums just recently cooperated with another popular web site for some Android benchmarking. I've exchanged with the author and still learned some things about these benchmarks when I thought I was all squared away (I used to write mainframe benchmarks, fwiw).
Those results will published in the coming weeks, and Phandroid.com will be getting first crack at those other-site summaries. I'll try to remember to post that all back here, or a similar thread, when the time comes.
sremick
Android Expert
I never said benchmarks had no value. I never said the lagfixes don't show real world speed improvements. And it's quite possible I understand how each of the numerous, varied lagfixes work even better than you.
What I did say was that Quadrant benchmarks after a lagfix, especially the OCLF for 2.1 (which it sounds like you used) are meaningless as they aren't proportional/linear or otherwise show any real sense of the actual improvement. The nature of how OCLF in-particular works artificially sends Quadrant scores into the stratosphere. So although your Captivate performs better, going from a 900 Quadrant to a 1800 Quadrant in this case does not mean anything like 2X improvement.
And while OCLF's method might seem "brilliant" to you, it's considered dirty, dangerous and outdated at this point by those in the know. It's a complicated "Rube Goldberg" hack. I'm not sure you'll find anyone left on XDA who has anything good to say about it these days.
I certainly wouldn't trust any data I cared about to live on the house of cards filesystem structure it sets up. Native ext4 is the way to go, without complicated virtualization layers and buffering. It's a blessing that Samsung didn't use the same method as OCLF... and there'd be no reason they would. They would've just built native ext4 support into the stock kernel, making the crazy and fragile tricks OCLF does totally unnecessary. This is effectively how VooDoo works... but unfortunately Voodoo has to use old kernel code... source code that predates even the stock kernel, so the hardware support isn't as good/optimized.
When you flash a new ROM please do not expect the GPS to work 100% with the first try. Often you have to let it sit trying to obtain a lock or delete the GPS data a few times and reboot just to get things working.
That's what I found when I first tried a new ROM, but after a reboot or two it was working better than before.
That's what I found when I first tried a new ROM, but after a reboot or two it was working better than before.
WOW I navigated with a Nexus One last night and the GPS on it blew my Captivate's GPS out of the water big time. It performed extremely solidly. It locked immediately, never fell behind, never put me on side roads and kept up with me around quick turns. I really wish the Captivate's GPS could perform half that well. It's still good enough I suppose. It's just so strange that it is getting completely smoked by older phones.
Sorry I know this isn't news but I couldn't believe the difference.
Sorry I know this isn't news but I couldn't believe the difference.
So I switched to the new Captivate (see my previous posting), immediately ran GPS Restore, and what a difference from my previous experiences! This is how GPS was meant to be. Almost always gets a fix within seconds, stays in a nice tight circle most of the time, hardly loses the fix, and doesn't constantly think I'm on some adjacent road. Now it's not quite as good as my Nuvi as far as consistency goes, but when you couple it with the Google traffic, voice search, up to date maps, etc etc... I'm now using it as my primary navigation. And it's been this way for almost a couple of weeks!
s.m.knipe
Android Expert
I've been saying this for months, its nice to see someone else verify/solidify it. I really wish I could pull the GPS chip from my N1 and put it in the Captivate, then I wish I had the knowledge base to be able to write a relevant firmware and drivers to implement it...
Jeez, I really hope for the sake of the millions affected by this across the series that it won't come down to a hardware fault.
The HTC phones all undoubtedly use variants of a Qualcomm transceiver with GPS.
The best I could find was an assertion that the Captivate uses Broadcom's InConcert BCM20751.
I've never seen much bad about Broadcom, so I'm hoping this really does come down to software for you guys.
FWIW, I have a LOT trouble believing this popular Engadget post and wonder how much of it is rumor-mongering vs. any truth -
Samsung Galaxy S GPS-gate: two problems, not one (and what to do about it) -- Engadget
The HTC phones all undoubtedly use variants of a Qualcomm transceiver with GPS.
The best I could find was an assertion that the Captivate uses Broadcom's InConcert BCM20751.
I've never seen much bad about Broadcom, so I'm hoping this really does come down to software for you guys.
FWIW, I have a LOT trouble believing this popular Engadget post and wonder how much of it is rumor-mongering vs. any truth -
Samsung Galaxy S GPS-gate: two problems, not one (and what to do about it) -- Engadget
What I have been wondering (as have people much smarter about GPS than I) is why do many Captivates seem to work well out of the box but degrade over time? It would seem that if it was data corruption then simply reloading the data, ie GPSRestore, should fix the issue; it didn't for me but then again I never changed my settings in the first place. If it was battery cover coatings and little brass clip then why would flashing a ROM or resetting the phone improve the situation for a bit? If it was poorly implemented drivers that should have been easily detected and fixed with a patch by Samsung or someone at XDA. If it was a bad antenna design then why would it start out okay and fade out over time?
With my last phone the GPS situation was never great and it got worse over the two months I owned it until the phone would barely find any sats and, on many trips along the same route, it would spend up to half the time unsuccessfully searching for GPS. That made me wonder if the GPS chip was going bad, like it got too hot searching and not finding and cooked itself or something.
The replacement I got last night is still in the honeymoon phase and in two long drives it never showed me off track, tracked accurately around sharp corners, showed me stopped at intersections instead of creeping forward and back through them, had either a sharp arrow icon or only a hint of fuzz around it, rerouted accurately and quickly and even achieved lock in under 10 seconds when Navigation was started at highway speeds. I don't expect that to last long and I'm reluctant to use the GPS much in order to try to preserve the 'good stuff'.
In the meantime I'll opt for these prudent rules:
1. the phone should not be exposed to bright lights
2. the phone should not be gotten wet
3. NEVER feed the phone after midnight
With my last phone the GPS situation was never great and it got worse over the two months I owned it until the phone would barely find any sats and, on many trips along the same route, it would spend up to half the time unsuccessfully searching for GPS. That made me wonder if the GPS chip was going bad, like it got too hot searching and not finding and cooked itself or something.
The replacement I got last night is still in the honeymoon phase and in two long drives it never showed me off track, tracked accurately around sharp corners, showed me stopped at intersections instead of creeping forward and back through them, had either a sharp arrow icon or only a hint of fuzz around it, rerouted accurately and quickly and even achieved lock in under 10 seconds when Navigation was started at highway speeds. I don't expect that to last long and I'm reluctant to use the GPS much in order to try to preserve the 'good stuff'.
In the meantime I'll opt for these prudent rules:
1. the phone should not be exposed to bright lights
2. the phone should not be gotten wet
3. NEVER feed the phone after midnight
Happy New Year, compadre!
Maybe next year, we'll know those answers.
I especially agree with not feeding after midnite.
Maybe next year, we'll know those answers.
I especially agree with not feeding after midnite.
s.m.knipe
Android Expert
The Broadcom chip we use (the number and info page are buried back near the beginning of this monstrosity of a thread, you and I probably could have found it easily if all these threads weren't haphazardly thrown together like a book without chapters
) has no cpu to make calculations and algorithm predictions, it relies on the main cpu. It does this to save battery power, and was shown by Broadcom to actually be more sensitive than previous units with their own cpu (like what Apple chose to use for the iPhone 4) IF IMPLEMENTED PROPERLY. I did and still do argue that much of the issue is load and balancing that load on the cpu, and could be at least helped by changing the rfs filesystem, and assigning the gps processes a higher priority level so as to not get pushed out by say notifications or other system polls... That being said, there does seem to be a poor choice on Samsung's part in the connectors used (that hardware fix on xda), and this is also a new chip, it could easily not be performing as well as broadcom proclaims- although I would hope Samsung would have tested it before selecting it...
While we can't apologize for thread merging, lest the entire page be filled with any one issue, I completely and sincerely thank you for your help in assisting me to understand.
Prior to your clarifying post, as a govt GPS programmer in a previous life and now a contributing member of the semiconductor industry, I was quite at a loss as to how things progressed to this state of affairs.
Given what you say, it seems that with little real-time Linux legerdemain (as you detail), this is update-fixable - I would hope.
As for testing - well, there's testing for success and then there's testing for non-failure - a discipline unto itself, typically lost on engineers until they totally screw one up.
Prior to your clarifying post, as a govt GPS programmer in a previous life and now a contributing member of the semiconductor industry, I was quite at a loss as to how things progressed to this state of affairs.
Given what you say, it seems that with little real-time Linux legerdemain (as you detail), this is update-fixable - I would hope.
As for testing - well, there's testing for success and then there's testing for non-failure - a discipline unto itself, typically lost on engineers until they totally screw one up.
s.m.knipe
Android Expert
I understand about the merging, but it does make it a pain to find anything. A sub-forum could have solved both problems...
Of course! I sincerely thank you for your responsiveness and openness. We are all working to reason out this issue and hope for a solution. I submitted as much info as I could to Samsung through various avenues, so here's to hoping they at least look into it...
Thanks but I am just regurgitating what I have read elsewhere, and smattered in a bit of my own reasoning. So take it with a grain of salt, as much of it comes from the manufacturer's announcements, Broadcom could just be deflecting blame... On the other hand, I remember reading somewhere that tom-tom was adopting the same chip, that would lead me to think that the chip is fine...
You seem very well qualified to help us reason this out, as you have been thus far, just based on your posts I could tell you have been a great asset, but I had no idea you had that in your background...
I really hope so, I wish they would release the driver info so XDA could sort this out once and for all...
True, it certainly seems that Samsung only conducts "non-failure" testing, and it also seems that their dev team is rather small, for it to have taken this long...
And Happy New Year's
SamsungCaptivated
Member
Sorry this post is long......
Mr GPS stated in a earlier post....
"Your GPS receiver needs to re-train again from the intelligent GPS applications to feed local relevant almanac. Currently, the GPS receiver in Galaxy S phone cannot perform an unsupervised learning for automatic collecting the relavant almanac for first fix."
I fully agree with his assessment. When my Captivate isn't used for GPS application for several hours, it doesn't seem to lock onto satellites quickly. I've tried this test many times and it confirms my suspicion that the Captivate phone has programming bug that cause it to forget the last satellite almanac data. It doesn't seem to store that info and deletes it after a few hours or doesn't seem to know how to use it again after a period of time. I started Google Map and let it sit to see how long it would take to lock onto satellites. After 6 minutes, I gave up. Several hours later, I did it again and it failed to lock up for at least 4 minutes. After the initial lock up, I exited Google Map and within a few minutes, I started Google Maps and it locked onto the satellites within 5 sec and it was dead-on accurate. I repeatedly exited and started Google Map with in few minutes of each other and each time it locked up to the satellites within 10-15 secs and was dead-on. I let the phone sit until the next day when I tried Google Map again. It took a LONG time to lock onto satellites. It made me think the phone needed to recalculate all ephemeris and almanac data to figure out where in the world the phone was located so it can look for the appropriate satellite's signal to calculate its position. My Garmin says something to the effect "if you move more than 500 miles from your last ON position, your device needs to recalculate its position by recalculating the ephemeris data." That ephemeris and almanac data is transmitted by the GPS satellites. What the Garmin does is it listens for any and all satellites in orbit. Obviously, only the ones overhead will be the one received by the Garmin device. Garmin GPS will then calculate from its internal database as to which satellites it saw should be overhead in a given area at a given time and date. From there the Garmin can start listening for the appropriate satellites for the the relavant almanac and ephemeris data. And once received, it can calculate its position. Garmins owners manual specifically states that a new unit could take up to 5 minutes to find and display its position. A device moved more than 500 miles since its last ON position could take up to 5 minutes to get its position. That got me thinking....
I downloaded GPS Status from the Android Market. This app allows you to delete the cached GPS data and reload the AGPS data manually at will.
This is what I do and it's so far worked for me consistently.
Before I use any GPS based app, I open up GPS Status and use TOOLS to delete the cached GPS data. Again under TOOLS, I then reload the AGPS data from the internet (AGPS data is info sent to your phone telling it approximately where you're located - your cell carrier can determine your general area by which cellsite you're connected to). It's giving the GPS a head start to which satellites to listen for. In simple terms, it's basic data telling the GPS "these satellites should be overhead right now". This headstart allows the GPS to start collecting the ephemeris and almanac data which is needed to calculate postion. Since I cleared the cached GPS data previously, it could take up to 30 secs for the new data to be received by the Captivate to be used as the new cached GPS data. For me, once I follow these steps, my GPS Status apps shows that I'm locked on to satellites within approximately 30-45 sec for the initial lock. I then exit out of GPS Status and start Google Maps. Google Maps consistently so far locks onto satellites within 10-15 sec and its dead-on. And I'm doing this indoors in my bedroom. I've repeated this test at least a dozen times and it's been consistent. It works for my Captivate. Yes, its a workaround and its a pain in the ass, but at least I can get my Captivates GPS up and running in under 1 minute for the initial lock up then I'm good for several hours after with it taking about 10-15 sec to find my position. By the way, my phone is stock, nothing done it to except a few apps download from the market and it was JH7 firmware, 2.1 - update 1 from the factory.
The following info is from Garmin....
GPS satellites transmit two low power radio signals, designated L1 and L2. Civilian GPS uses the L1 frequency of 1575.42 MHz in the UHF band. The signals travel by line of sight, meaning they will pass through clouds, glass and plastic but will not go through most solid objects such as buildings and mountains.
A GPS signal contains three different bits of information - a pseudorandom code, ephemeris data and almanac data. The pseudorandom code is simply an I.D. code that identifies which satellite is transmitting information. You can view this number on your Garmin GPS unit's satellite page, as it identifies which satellites it's receiving.
Ephemeris data, which is constantly transmitted by each satellite, contains important information about the status of the satellite (healthy or unhealthy), current date and time. This part of the signal is essential for determining a position.
The almanac data tells the GPS receiver where each GPS satellite should be at any time throughout the day. Each satellite transmits almanac data showing the orbital information for that satellite and for every other satellite in the system.
Mr GPS stated in a earlier post....
"Your GPS receiver needs to re-train again from the intelligent GPS applications to feed local relevant almanac. Currently, the GPS receiver in Galaxy S phone cannot perform an unsupervised learning for automatic collecting the relavant almanac for first fix."
I fully agree with his assessment. When my Captivate isn't used for GPS application for several hours, it doesn't seem to lock onto satellites quickly. I've tried this test many times and it confirms my suspicion that the Captivate phone has programming bug that cause it to forget the last satellite almanac data. It doesn't seem to store that info and deletes it after a few hours or doesn't seem to know how to use it again after a period of time. I started Google Map and let it sit to see how long it would take to lock onto satellites. After 6 minutes, I gave up. Several hours later, I did it again and it failed to lock up for at least 4 minutes. After the initial lock up, I exited Google Map and within a few minutes, I started Google Maps and it locked onto the satellites within 5 sec and it was dead-on accurate. I repeatedly exited and started Google Map with in few minutes of each other and each time it locked up to the satellites within 10-15 secs and was dead-on. I let the phone sit until the next day when I tried Google Map again. It took a LONG time to lock onto satellites. It made me think the phone needed to recalculate all ephemeris and almanac data to figure out where in the world the phone was located so it can look for the appropriate satellite's signal to calculate its position. My Garmin says something to the effect "if you move more than 500 miles from your last ON position, your device needs to recalculate its position by recalculating the ephemeris data." That ephemeris and almanac data is transmitted by the GPS satellites. What the Garmin does is it listens for any and all satellites in orbit. Obviously, only the ones overhead will be the one received by the Garmin device. Garmin GPS will then calculate from its internal database as to which satellites it saw should be overhead in a given area at a given time and date. From there the Garmin can start listening for the appropriate satellites for the the relavant almanac and ephemeris data. And once received, it can calculate its position. Garmins owners manual specifically states that a new unit could take up to 5 minutes to find and display its position. A device moved more than 500 miles since its last ON position could take up to 5 minutes to get its position. That got me thinking....
I downloaded GPS Status from the Android Market. This app allows you to delete the cached GPS data and reload the AGPS data manually at will.
This is what I do and it's so far worked for me consistently.
Before I use any GPS based app, I open up GPS Status and use TOOLS to delete the cached GPS data. Again under TOOLS, I then reload the AGPS data from the internet (AGPS data is info sent to your phone telling it approximately where you're located - your cell carrier can determine your general area by which cellsite you're connected to). It's giving the GPS a head start to which satellites to listen for. In simple terms, it's basic data telling the GPS "these satellites should be overhead right now". This headstart allows the GPS to start collecting the ephemeris and almanac data which is needed to calculate postion. Since I cleared the cached GPS data previously, it could take up to 30 secs for the new data to be received by the Captivate to be used as the new cached GPS data. For me, once I follow these steps, my GPS Status apps shows that I'm locked on to satellites within approximately 30-45 sec for the initial lock. I then exit out of GPS Status and start Google Maps. Google Maps consistently so far locks onto satellites within 10-15 sec and its dead-on. And I'm doing this indoors in my bedroom. I've repeated this test at least a dozen times and it's been consistent. It works for my Captivate. Yes, its a workaround and its a pain in the ass, but at least I can get my Captivates GPS up and running in under 1 minute for the initial lock up then I'm good for several hours after with it taking about 10-15 sec to find my position. By the way, my phone is stock, nothing done it to except a few apps download from the market and it was JH7 firmware, 2.1 - update 1 from the factory.
The following info is from Garmin....
GPS satellites transmit two low power radio signals, designated L1 and L2. Civilian GPS uses the L1 frequency of 1575.42 MHz in the UHF band. The signals travel by line of sight, meaning they will pass through clouds, glass and plastic but will not go through most solid objects such as buildings and mountains.
A GPS signal contains three different bits of information - a pseudorandom code, ephemeris data and almanac data. The pseudorandom code is simply an I.D. code that identifies which satellite is transmitting information. You can view this number on your Garmin GPS unit's satellite page, as it identifies which satellites it's receiving.
Ephemeris data, which is constantly transmitted by each satellite, contains important information about the status of the satellite (healthy or unhealthy), current date and time. This part of the signal is essential for determining a position.
The almanac data tells the GPS receiver where each GPS satellite should be at any time throughout the day. Each satellite transmits almanac data showing the orbital information for that satellite and for every other satellite in the system.
SamsungCaptivated
Member
Another thought on GPS....
It was mentioned that Samsung used Broadcom's GPS chip. Isn't SiRF III chips better at low signal processing? Their website says it can detect and process signals as low as -160dBm. That's a low signal strength. I can attest to the effectiveness of SiRF chips. I own 5 mobile GPS devices. Two of them have the SiRF GPS processor in it (I made sure it had them when I bought them). The two with the SiRF will still receive and process GPS signals when I'm sitting under a bridge or in an area with tall high-rise buildings (like in downtown LA) without much problem. The other 3 (one which is an older Garmin and 2 Magellan) do not have the SiRF chips. They lose the signal when I'm under the same bridge and they frequently lose track of the satellites when I'm roaming around the tall high rises. That made me a believer in the efficiency of the SiRF III chips and processors.
It was mentioned that Samsung used Broadcom's GPS chip. Isn't SiRF III chips better at low signal processing? Their website says it can detect and process signals as low as -160dBm. That's a low signal strength. I can attest to the effectiveness of SiRF chips. I own 5 mobile GPS devices. Two of them have the SiRF GPS processor in it (I made sure it had them when I bought them). The two with the SiRF will still receive and process GPS signals when I'm sitting under a bridge or in an area with tall high-rise buildings (like in downtown LA) without much problem. The other 3 (one which is an older Garmin and 2 Magellan) do not have the SiRF chips. They lose the signal when I'm under the same bridge and they frequently lose track of the satellites when I'm roaming around the tall high rises. That made me a believer in the efficiency of the SiRF III chips and processors.
SamsungCaptivated
Member
Ok, just did another GPS test with my Captivate.... I last did the GPS Status "delete the cached GPS data and reload the AGPS data" about hour ago. I then went to Google Maps to look around. My Captivate locked onto the satellites in 4 seconds (I was outside in a parking lot with clear view of the sky) and was dead-on accurate. In fact, it was actually kinda funny.... I turned on the satellite layer and zoomed in on the map to where the dot was showing and it was on the parking stall that I was standing in! How's that for accuracy!! I then drove home about 5 minutes away and checked Google Map again. The GPS locked on to the satellites in 7 seconds and it showed me in my driveway! I have not tested my Captivate GPS in a moving car. I have a Garmin GPS for that.
A comparison of your Garmin and Captivate tuned up with that app would be most welcome here, I'm sure.
reinbeau
Klaatu barada nikto
That is exactly what kind of pinpoint accuracy you should expect from a GPS unit. As I've mentioned in the past, my best with my iQue was 5.84'. The only time Betty took awhile to lock was when I had moved over a distance and had not used her to navigate there - I'm talking over 90 miles away from the last lock. Having to go through all these gyrations to get a good lock is ridiculous, but if it works, it's what we've got to do for the moment.
I haven't navigated with the Captivate mainly because I can't figure out how to make it happen, and there are only so many hours in a day. Navigation and GPS is still in its infancy in Android, and I really never cared for newborns
Just use the LBSTESTMODE to delete the GPS data and back out to home then try getting a location in Google Maps. I do this and while its not locked in seconds everytime, it'll obtain a lock quickly. Even while driving. Running cog 2.3b8. I even bent the little metal clip to the right of the SIM card upward a very very tiny amount.
I have an external bluetooth gps receiver, model BT-338 which uses that chipset.
Bought an app on the marketplace called bluetooth gps provider which allows me to use my external gps reciver with the captivate. See the app here: Bluetooth GPS Provider - Android app on AppBrain
The external receiver was left over from my treo 650 gps days and I can only recall once or twice the receiver would not lock. I had to battery pull it and try again.
However, in my case I can use the captivate's GPS without major issue and I always have my external gps reciever with me should I need it.
I purchased the phone back in September 2010. My GPS was way off but of course AT&T claimed there will be an update to fix it. While waiting I found a GPS hack which did prove to be a little more accurate but it was still off. Sure enough the update came and it was not a fix at all. At this point I took it upon myself to update to Firmware 2.2 and guess what? My GPS is still off.
I am sure this is still a known issue. Has anyone fixed this yet?
I am sure this is still a known issue. Has anyone fixed this yet?
I replaced my Captivate under the warranty for a separate issue and WOW the GPS is showing a MASSIVE improvement in speed and accuracy compared to my old one. I never knew that a cellphone GPS could be this good.
It tracks my movement around corners immediately and it isn't putting me on side streets or in the wrong direction like my old Captivate. Even when my previous phone was at its best it took the GPS 30 seconds or more to catch up around corners.
Also, it doesn't need to be right under the windshield to work.
I'm crossing my fingers that this phone doesn't have the GPS issue. And before anyone asks yes I have 'Use wireless networks' disabled.
It tracks my movement around corners immediately and it isn't putting me on side streets or in the wrong direction like my old Captivate. Even when my previous phone was at its best it took the GPS 30 seconds or more to catch up around corners.
Also, it doesn't need to be right under the windshield to work.
I'm crossing my fingers that this phone doesn't have the GPS issue. And before anyone asks yes I have 'Use wireless networks' disabled.
