Touch Pro and Touch Pro 2 GPS polling tweaks

[RSD]

By the way, I use Garmin Mobile XT not Tom Tom (I could not get Tom Tom working on XV 6800).

The cool thing was that I took the SD out of XV6800 and put it in TP2 and it asked if I wanted to install Garmin and I said yes, and viola it worked out of the box!

Now, if I can get the FM Radio working....

..Ramesh

[BooDaddy]

Awesome! Thanks for the input. I will give it a try on mine as well with the higher polling number. Hows the battery life when using the GPS with that number?

[RSD]

I will try it in few hours when I go to pick up my daughter from school and report back.

..Ramesh

[mswlogo]

Ok here is the scoop guys.

For Grins I tried this poll interval change with iGuidance (via GPSGate).

No change what so ever.

Then I changed iGuidance to NOT use GPSGate (bad idea by the way) and use Windows Mobile Virtual GPS COM Port (i.e. that port you set in "External GPS").

Then iGuidance went nuts and I could only see the clock on the GPS Page to check out update rate and it's resolution was 1 second. But it was "updating" way faster than 2 seconds like it normally does. Just because it updates it does not mean it has new data.

So for folks that are seeing a difference by changing this, it means you have your GPS Setup wrong and you WILL see lag becuase the Windows Mobile Virtual GPS COM port (usually COM4 by default) has been broken for years and still is (it lags like crazy). Boosting the pollinterval may make a bad setup less worse.

What you are boosting is how often the WM Virtual GPS COM port driver calls the Windows Mobile GPS API.

The better solution is to use the Windows Mobile GPS API directly and get the broken Windows Mobile GPS Virtual COM port out of the loop.

You do this by following what I said in the "How to setup your GPS thread" see my signature for how to do this.

If you have your heart set on playing with this pollinterval then use the COM port you have set in "External GPS" (usually COM4 by default on these phones but that might not be the same on all ROMs). And if you want to try a sain PollInterval I would try 500ms. That way you are assured of not missing a change that happens on 1 sec intervals and not wasting too much CPU and Battery for stuff that CAN'T change 10 times a second.

[BooDaddy]

Dude.... you still dont get it.
I have said this time and time again.... none of my GPS apps use the virtual com port...You fix did not work for me at all. GPSGate is not needed.

I can with most certainty confirm I am using the the WinMo gps API, and not its virtual com driver. When my polling settings were default, I had lag. With them set at the higher values, I dont have lag (as bad). My cab worked great for me.

[mswlogo]

Originally Posted by BooDaddy

Dude.... you still dont get it.
I have said this time and time again.... none of my GPS apps use the virtual com port...You fix did not work for me at all. GPSGate is not needed.

I can with most certainty confirm I am using the the WinMo gps API, and not its virtual com driver. When my polling settings were default, I had lag. With them set at the higher values, I dont have lag (as bad). My cab worked great for me.

I tried setting it to 10 secs (10000) and it had absolutely no effect on iGuidance or Google Maps (which uses WinMo GPA API).

I stand corrected. It seems to effect absolutely nothing. I have it set to 10 seconds and both Virtual COM port and WM GPS API behave exactly the same. I thought when I saw iGuidance updating faster than 1 second it was due to pollinterval. But I set it to 10 seconds and it behaves the same.

If you think a parameter does something. Then proove it by setting it to extremes and noting the behavior change.

If setting it smaller goes faster set it big (really big) and see if it goes slower. It doesn't.

[BooDaddy]

You are just outright wrong on this one my friend. The "virtual port" you are referring to is under the registry entry:
HKLM\System\CurrentControlSet\GPS Intermediate Driver\Multiplexer

You need to read up on the WinMo intermediate driver and how it works. First its 2 fold. It firstly creates a simple interface for developers to talk to the GPS chipset through windows. Sure, you can talk to the GPS chipset directly with its hardware port or address, but that ties it up to only one application. Thats why everyone uses the Windows API.
Secondly, it creates the multiplexer that simply forwards the NMEA data to a virtual com port...in most instances its com4.

When we change the polling interval under the key HKLM\System\CurrentControlSet\GPS Intermediate Driver\GPSOneDevice we are actually changing how many times the WinMo gps API actually queries the GPS chip. In that same key, you can see where the port is set to GPS0, which if I am not mistaken is the actual hardware address port for the GPS chip.

What you are referring to is another part....in the mulitplexer. WinMo 5 was the first to have this multiplexer so that more than one app could access the gps data at the same time. It isnt very good, but for legacy apps that needed to access the gps through a direct COM, it worked. I dont know of any developer worth his salt would code an app to talk to the gps this way now.

GPSGate talks to the WinMo gps API the same way 99% of all other apps nowadays does.... so its a waste of a program.

You need to jump on over to MSDN and readup on how it works. Better yet, grab the SDK and play with it.

Now, I will say this, I can see where increasing the polling rate could decrease battery life, as we are essentially increasing how many times windows calls the gps hardware in a given time frame (in the case of my cab, its 10 times per second).

With this, I am done arguing with you. It would be nice if someone more fluent in C++ or C# than us would chime in. The SDKs for both Winmo 5 and Winmo 6 are free. If I am wrong on this one, I am not ashamed to admit it, but I will need someone to show me where I am wrong.

[mswlogo]

Originally Posted by BooDaddy

You are just outright wrong on this one my friend. The "virtual port" you are referring to is under the registry entry:
HKLM\System\CurrentControlSet\GPS Intermediate Driver\Multiplexer

You need to read up on the WinMo intermediate driver and how it works. First its 2 fold. It firstly creates a simple interface for developers to talk to the GPS chipset through windows. Sure, you can talk to the GPS chipset directly with its hardware port or address, but that ties it up to only one application. Thats why everyone uses the Windows API.
Secondly, it creates the multiplexer that simply forwards the NMEA data to a virtual com port...in most instances its com4.

When we change the polling interval under the key HKLM\System\CurrentControlSet\GPS Intermediate Driver\GPSOneDevice we are actually changing how many times the WinMo gps API actually queries the GPS chip. In that same key, you can see where the port is set to GPS0, which if I am not mistaken is the actual hardware address port for the GPS chip.

What you are referring to is another part....in the mulitplexer. WinMo 5 was the first to have this multiplexer so that more than one app could access the gps data at the same time. It isnt very good, but for legacy apps that needed to access the gps through a direct COM, it worked. I dont know of any developer worth his salt would code an app to talk to the gps this way now.

GPSGate talks to the WinMo gps API the same way 99% of all other apps nowadays does.... so its a waste of a program.

You need to jump on over to MSDN and readup on how it works. Better yet, grab the SDK and play with it.

Now, I will say this, I can see where increasing the polling rate could decrease battery life, as we are essentially increasing how many times windows calls the gps hardware in a given time frame (in the case of my cab, its 10 times per second).

With this, I am done arguing with you. It would be nice if someone more fluent in C++ or C# than us would chime in. The SDKs for both Winmo 5 and Winmo 6 are free. If I am wrong on this one, I am not ashamed to admit it, but I will need someone to show me where I am wrong.

The regisrty setting is dormant IMHO (like 100's of others in WM). It's effecting nothing.

If setting polling rate to 10 seconds testing GoogleMaps (on Windows Managed), iGuadance on GPSGate or iGuidance on COM4 (Virtual COM Port) all still update around 1 a second I'm convinced it does nothing.

I live and breath C# every day.

It seems consistent with what most have found that it made no freaken difference.

It seems consistent also with your poll that a majority of people have no issue at all with the GPS in the first place.

[BooDaddy]

All you have to do is google search "Touch Pro gps lag".... TONS of stuff on it. Heck, there is tons of stuff over at XDA Developers about it. The TP2 and TP1 share the same hardware....just go into the TP1 forums and you will see tons of lag threads if you search.

Theres definately something to it.

[mswlogo]

Originally Posted by BooDaddy

All you have to do is google search "Touch Pro gps lag".... TONS of stuff on it. Heck, there is tons of stuff over at XDA Developers about it. The TP2 and TP1 share the same hardware....just go into the TP1 forums and you will see tons of lag threads if you search.

Theres definately something to it.

There may be something about ALL Qualcomm based GPS units that they are not as fast as $2000.00 car units or $700 Garmin units. But those may have other means of getting "smoother" updates.

There are many ways of doing better than 1 update a second.

You can do interpolation (which is what all the chatter is about on XDA).
You can use gyro and compass to esimate between updates.
There are systems that do 1000hz for military.
You can also pan the screen smoother from one update to the next to make it feel continuous (I believe this is what high end Nuvi's and car systems do).

But for the average Joe 1-2 seconds per update works really well for walking or driving. I've been using them since the HTC 6600. The Touch Pro2 is the best GPS Unit I have had.

If you are doing a LOT worse than that and you have set up everything correctly I don't know what else to say.

I'll try to video it walking and driving but my camera that does video is at my lake house.

What we really should test with is an application that has time stamps and positions so we can compare notes. We may be seeing the same performance and some are satisfied and some are not. Or there may be sometime very different.

I know we are both trying to help.