12/23/2008, 04:57 AM
|
#1 (permalink) |
|
Forum Leader
![]() ![]() Join Date: Jan 2008
Posts: 3,140
Likes Received: 0
Thanks: 24
Thanked 193 Times in 44 Posts
|
The 800w, stock, seems to have pretty typical aGPS Satellite Acquisition times (at least in my experience) of around 30 seconds to 2 minutes, depending on conditions.
I believe I have discovered a way to reduce this time to under ten seconds! Additionally, it appears (from my quick tests) that I am finding more satellites than previously. In my home, during many previous tests I would only pick up 2-3 sats at first, and top out at maybe 4-6. With this change, I am now picking up 8-10 from the same location. This is not an easy hack, so for those who aren't phone savvy, be aware that you will be using some incredibly powerful tools that can literally destroy your phone beyond repair. That being said, if you follow the simple instructions, then you should be just fine. This involves several tools, none of which are mine. (There's a change!) Tools Needed:
That's an intimidating list I know, but it's really much simpler than the list might imply. To apply the hack:
Last edited by Ebag333; 12/23/2008 at 05:02 AM. |
12/23/2008, 10:19 AM
|
#3 (permalink) |
|
Member
![]() Join Date: Oct 2003
Posts: 1,650
Likes Received: 1
Thanks: 186
Thanked 203 Times in 133 Posts
|
Another way to reduce the Acquisition time to under 10 seconds is to just do a quick "Sprint Maps" search from the Today Screen, then open your GPS application and it will already have your position acquired.
The easiest way to get fast aquistion times is to just get the latest version of Garmin XT. It seems to acquire the position instantly without needing the before mentioned priming.
__________________
Pilot 1000 -> Pilot 5000 ->Palm Pilot Professional -> HP 620LX -> TRG Pro -> Palm V -> Palm Vx -> Palm M505 -> Palm i705 -> Palm Tungsten|T -> Samsung i500 -> Treo 600->Treo 650 -> Treo 600-> Treo 700p ->Centro ->Treo 800w + Redfly C8n -> Palm Pre -> HP Touchpad R.I.P Palm 1996-2011 |
12/23/2008, 10:25 AM
|
#4 (permalink) | |
|
Member
![]() Join Date: Jan 2006
Location: Long Island, NY
Posts: 5,890
Likes Received: 0
Thanks: 44
Thanked 109 Times in 80 Posts
|
Quote:
Basically, what you are doing is called a "warm start" where the GPS coordiantes are stored on the device for a certain period of time. Warm starts are always fast. Cold starts, where there is no such info, always takes a longer time. This hack helps to speed that part up. In addition, those extra options enabled may allow more consistent connections to more satellites than before, which could aid in accuracy as well. Ebag: Great work as always! I'm gonna go try to dig up my MSL code and give this a shot today.
|
|
12/23/2008, 11:07 AM
|
#6 (permalink) | |
|
Forum Leader
![]() ![]() Join Date: Jan 2008
Posts: 3,140
Likes Received: 0
Thanks: 24
Thanked 193 Times in 44 Posts
|
Quote:
Mal has it (basically) correct. But since I have a pathological need to ensure that it is clarified beyond clarified..... Default the 800w supports 1 type of aGPS acquisition. This mod allows it to use a second type of aGPS acquisition, as well as utilizing the GPS data stored on the device. Most (all?) modern GPS devices now store ephemeris data, or basically a map of not only where the satellites have been, but where they are going to be in the future. As long as you have an accurate time, you can tell where those sats are. If you're off even a tenth of a second (or a few miles), then that data is essentially useless. Since the 800w stores the ephemeris data, there's no reason to not use this data when finding the sats. But the 800w doesn't use it. It's like downloading the yellow pages to your phone, then calling 411. Calling 411 is a lot faster than dialing random numbers, but it's not as fast as typing in the business name and hitting dial. Typical times for GPS are 15-30 minutes for cold start, 5-15 minutes for warm start (have the info on where the satellites used to be, but don't have information on where they are now), and 30 seconds to 5 minutes for aGPS. Of course all that depends on the exact situation. If you're in a crappy location where you can only view 3 sats, then no amount of aGPS will help, it'll be slow. With using the data stored on your device, you don't have to go to a network server to download that data. This is why you can pick up the sats much faster than even with aGPS. As a side note, all my tests were done with Google Maps. I don't have Garmin installed, and TomTom uses GPSGate which primes the GPS and will throw off any tests I have. I did notice that using Palm's Map program (or the today screen plugin) seemed to start up the GPS faster than Google Maps (back when I used it). I am unable to explain why that happens, however. My only possible explanation is that Google Maps is using the Windows Mobile GPS interface, while Palm's Maps is accessing the GPS more directly. Hope that clarifies what's going on here.
|
|
12/23/2008, 11:26 AM
|
#8 (permalink) |
|
Member
![]() Join Date: Jan 2006
Location: Long Island, NY
Posts: 5,890
Likes Received: 0
Thanks: 44
Thanked 109 Times in 80 Posts
|
Ebag,
I see AFLT listed on there which is Advanced Forward Link Trilateration aka "tower triangulation". You think enabling that will work with Google Maps "My Location" or is there simply no API layer for WM to do that? Edit: I don't think it works, but I checked that and hybrid and see no negative side effects All in all, this was pretty easy to do and yes, GPS is much faster now with more sats. I mean...this is way fast.
Last edited by Malatesta; 12/23/2008 at 11:38 AM. |
12/23/2008, 11:58 AM
|
#10 (permalink) | |
|
Member
![]() Join Date: May 2006
Posts: 2,435
Likes Received: 0
Thanks: 198
Thanked 114 Times in 72 Posts
|
Quote:
aGPS typcially reduces factory and cold start to warm/hot. On my tomtom if I take it to Europe the first shot is 12.5 minutes. If I shut down completely for a couple of days (cold start) it is 45 seconds, at most two minutes if I am moving in a city with tall buildings. Warm starts from same location are less than half a minute. I get even better warm and cold starts from the 800w then the TomTom. Are you sure you are not getting your reported unusually long reported times because you are reseting during testing and indoors? A post hard reset on my 800w gets me a fix in about 30 seconds! Or perhaps their is some difference in ip access. In other words perhaps for some users in some cases the aGSP is not working fully due to data conenction issues (as opposed to the non ip network time aquisition) |
|
12/23/2008, 12:32 PM
|
#11 (permalink) |
|
Forum Leader
![]() ![]() Join Date: Jan 2008
Posts: 3,140
Likes Received: 0
Thanks: 24
Thanked 193 Times in 44 Posts
|
Mal:
I did enable those and saw no difference. I simple think that WM doesn't use it. Aero: I'm not going to argue with you, because 1) it's a stupid thing to argue about and 2) you will find different times for different devices. My standalone GPS puck gives 15-30 minutes (but says it can take as long as 45 if the stored information is incorrect, such as when it's been used then moved a couple hundred miles) and 5-15 for a warm start. Typically 2-5 is more accurate, which is faster than what is listed in the manual/specs. The 800w always uses aGPS, so 30 seconds (even after a hard reset) is to be expected. I've watched it take up to 2-3 minutes, but that is typically only in poorer conditions (such as testing in my house). You can quibble over the details and whether or not aGPS is 30 seconds or 32.5 seconds. The times I gave above were painted with intentionally broad strokes, there are aGPS devices which take 5-10 minutes and standalone devices which take seconds. The fact of the matter is that with this you can go from 30 seconds (best case scenario) to several minutes (worst case scenario) down to mere seconds. Additionally, and I have only done very limited testing on this so far, this appears to resolve the gps positional "lag" that previously existed. as i've said before, the 800w isn't any worse than many other devices, but after this change the position is MUCH closer to accurate for me. Previously the positioning was about 20-100 feet, now it is 0-50 feet. As I've said, I've only done one test on this (driving to work). This may simply be a side effect of having more sats to synch up to, or it may allow the 800w to use another mode that is faster. I will try and test this further. |
12/23/2008, 01:33 PM
|
#12 (permalink) |
|
Banned
Join Date: Dec 2006
Posts: 1,190
Likes Received: 0
Thanks: 0
Thanked 0 Times in 0 Posts
|
Maybe this is why the GPS lock and accuracy was always better on the HTC devices like the Mogul and Touch Pro. Those settings were already checked. It's a very old hack that the Verizon people were doing a long time ago to boost their GPS performance. http://forum.xda-developers.com/showthread.php?t=400975
I wonder why Sprint didn't configure the 800w correctly before sale? |
12/23/2008, 02:27 PM
|
#13 (permalink) |
|
Member
![]() Join Date: Jan 2006
Location: Long Island, NY
Posts: 5,890
Likes Received: 0
Thanks: 44
Thanked 109 Times in 80 Posts
|
Ebag,
You may want to throw these settings in there too...unless you've determined them to not be necessary. I haven't noticed necessarily any gains yet but also no side effects either. [CODE]PDE IP Address = 68.28.31.49 PDE Port Number = 5017 PDE Transport = IP Position Calculation = PDE[/CODE] Last edited by Malatesta; 12/23/2008 at 02:35 PM. |
12/23/2008, 03:14 PM
|
#15 (permalink) | |
|
Forum Leader
![]() ![]() Join Date: Jan 2008
Posts: 3,140
Likes Received: 0
Thanks: 24
Thanked 193 Times in 44 Posts
|
Quote:
As for configuration, it's appples to oranges. Many older phones such as the Mogul didn't have a way of storing GPS data natively and required 3rd party apps from HTC to do so. (Nor did they have gpsONE.) The 800w is one of the newer devices with the beginnings of gpsONE on them, where this information is handled directly by the GPS drivers. That's not to say that the 800w has true gpsONE ala the Treo Pro, but then again there have been several reports where when in network service (and should be able to use aGPS) the Treo Pro has been taking an extended time to synch to the sats. In those situations gpsONE failed, and aGPS failed, and the phone is falling back on standalone lookups. As with all new technology, there are bugs to iron out. The GPS works wonderfully on the Treo Pro, and the vast majority of the people are happy with it. And of course there is the GPS problems with the Diamond and (non-Sprint) Touch Pro's that HTC has yet to iron out. As they have resolved it for the Sprint Touch Pro, it somewhat surprises me that a fix for their older phones has not been yet released. I have no doubt that they WILL resolve it, but given they have the fix in hand it's somewhat odd it hasn't been released (or hacked off the Sprint Touch Pro). Additionally since the 800w was originally intended for Verizon, the options may not have been enabled in the NV to lock it down to aGPS only. While that is not an excuse for Palm or Sprint who clearly missed some of this, it is a posible explanation. Additionally ther may have been licensing/feature costs here that could have come into play. Qualcomm could have told Palm/Sprint "pay us $X more per device and we will enable Sat lookup on the device." Palm/Sprint may have turned them down, figuring that saving 20 seconds wasn't worth the extra cost. You see this happen quite often. Toyota often does this, where they manufacture their vehicles with all the features but not the way of using them. Such as with ATRAC on the FJC, where it's something like a $2500 "upgrade", or you can buy a $50 button and have it. Or like Intel does with their processors, where they only manufacture a few and lock them down. Take your pick of either explanation, both are just as reasonable. But since none of us work for Palm, Qualcomm, or Sprint, we are all just guessing. Anyway, all that being said, the 800w didn't take any longer to acquire sats than many phones, including HTC phones. With this enabled, it's faster. Pretty sure that's not a dig against HTC who (usually) has quite good acquisition times, and users should be happy with only 30 seconds compared to the bad old days of 15 minutes or more. I had looked at porting gpsONE to the 800w (and even played with it some). You can install it and it will run, but it does not function properly. Now that we have this, we have the equivelent of gpsONE and other similar applications (such as HTCs), which simply puts us at a level playing field. ![]() Mal: I entered those settings, but thy did not stick. Did you go back and check to se if they stuck in there, or were removed? I als played around with some other stuff such as the volume settings. Didn't see a differnce unfortunatly. ![]() Lots of interesting settings though. I'm to chicken to play with some of the settings in QPST...some dangerous stuff in there. Last edited by Ebag333; 12/23/2008 at 03:20 PM. |
|
12/23/2008, 03:25 PM
|
#16 (permalink) | |
|
Member
![]() Join Date: Jan 2006
Location: Long Island, NY
Posts: 5,890
Likes Received: 0
Thanks: 44
Thanked 109 Times in 80 Posts
|
Quote:
And I have a neat trick coming up in a few mins separate from the GPS one... PS You mean "gpsOneExtra" in your above post ;-) |
|
12/23/2008, 03:30 PM
|
#17 (permalink) | |
|
Forum Leader
![]() ![]() Join Date: Jan 2008
Posts: 3,140
Likes Received: 0
Thanks: 24
Thanked 193 Times in 44 Posts
|
Quote:
The 800w is a gpsONE chipset, but it does not have the gpsONE program or driver on it. I can get both to install and seemingly function. It's a selectable option as a driver for the GPS. But it doesn't actually work. There's probably a way to get it going, but it's going to be a big pain in the ****. And now that we have basically instant GPS, for zero gain. ![]() And don't tease me with your hint dropping!
|
|
12/23/2008, 03:38 PM
|
#18 (permalink) |
|
Member
![]() Join Date: Jan 2006
Location: Long Island, NY
Posts: 5,890
Likes Received: 0
Thanks: 44
Thanked 109 Times in 80 Posts
|
ah..duly noted!
You know, we can update our PRL now directly. Of course there is no newer version that I know of... And my trick will be up in...5 mins, lol. |
![]() |
|
| Thread Tools | |
| Display Modes | |
|
|



