After extensive testing, I'm quite positive AT&T is artificially limiting network speed on their version of the Titan by limiting the HSDPA feature set to category 8. The phone hardware of course is able to support HSDPA category 10 features.
To be clear, enabling HSDPA category 10 features doesn't just raise the theoretical maximum throughput to 14.1 Mbps. HSDPA category 10 features should improve performance across the board, even when throughput is below the theoretical maximum of HSDPA category 8, which is 7.2 Mbps.
My theory is based on testing in multiple markets, with and without advanced backhaul, against an LG Thrill and Samsung Galaxy S II. On the two Android phones I am able to adjust and verify the HSDPA category. The LG Thrill set to HSDPA category 10 and the Galaxy S II set to HSDPA category 14 both get about twice the network performance of the HTC Titan in any given location using the same sim card.
If you're skeptical that AT&T would essentially disable HSPA+ on a phone they're advertising as HSPA+ 4G enabled, look no further than the LG Thrill. It has been proven by analyzing code and hidden settings that it was shipped with the HSDPA category set to 8. Setting it to 10 on the LG Thrill roughly doubles network performance for most people. AT&T does this to lighten the load on their network and they figure most people will never notice. Maybe you remember the AT&T HSUPA fiasco that gained wide media coverage early this year. So there are multiple cases of past precedent for this kind of behavior from AT&T.
Anyways, on Android phones it isn't a big deal because we can easily set the proper HSDPA category in our custom ROMs. Windows Phone, however, is closed and very locked down so we have no way (that I know of) to adjust, or even verify, this setting.
The international Titan is undoubtedly not restricted in this way, but apparently it doesn't support AT&T's 1900 MHz frequency so we can't compare performance directly to prove foul play. That said, if anyone has both phones I would like to see some network speed comparisons between the two, or maybe a comparison to other Windows Phones on AT&T's network (though AT&T could be limiting all of them).
If anyone is seeing speeds above 7.2 Mbps on their AT&T Titan it would prove my theory wrong, so please report if you've seen 8-12 Mbps on your AT&T Titan at any time on AT&T's network. These speeds may seem high, but they're fairly common now in HSPA+ areas with advanced backhaul.
The only other possibilities are poorly set TCP buffer sizes or just flat horrible radio hardware, but knowing AT&T the speeds we're seeing are artificially limited.
I used this site in the browser for most testing: http://www.ispgeeks.com/wild/modules.php?name=Bandwidth_Meter_DSL I also used the BandWidth app as well as several other sites to verify my findings.
edit: Forgot to give you an idea of how large the disparity is. At my house outside of Atlanta there is no advanced backhaul and I'm getting 2-3 Mbps on the Titan compared to 3-4 Mbps on my Android phones. In the city where there is advanced backhaul in place I'm maxing out around 4 Mbps on the Titan compared to 8-10 Mbps on my Android phones.
To be clear, enabling HSDPA category 10 features doesn't just raise the theoretical maximum throughput to 14.1 Mbps. HSDPA category 10 features should improve performance across the board, even when throughput is below the theoretical maximum of HSDPA category 8, which is 7.2 Mbps.
My theory is based on testing in multiple markets, with and without advanced backhaul, against an LG Thrill and Samsung Galaxy S II. On the two Android phones I am able to adjust and verify the HSDPA category. The LG Thrill set to HSDPA category 10 and the Galaxy S II set to HSDPA category 14 both get about twice the network performance of the HTC Titan in any given location using the same sim card.
If you're skeptical that AT&T would essentially disable HSPA+ on a phone they're advertising as HSPA+ 4G enabled, look no further than the LG Thrill. It has been proven by analyzing code and hidden settings that it was shipped with the HSDPA category set to 8. Setting it to 10 on the LG Thrill roughly doubles network performance for most people. AT&T does this to lighten the load on their network and they figure most people will never notice. Maybe you remember the AT&T HSUPA fiasco that gained wide media coverage early this year. So there are multiple cases of past precedent for this kind of behavior from AT&T.
Anyways, on Android phones it isn't a big deal because we can easily set the proper HSDPA category in our custom ROMs. Windows Phone, however, is closed and very locked down so we have no way (that I know of) to adjust, or even verify, this setting.
The international Titan is undoubtedly not restricted in this way, but apparently it doesn't support AT&T's 1900 MHz frequency so we can't compare performance directly to prove foul play. That said, if anyone has both phones I would like to see some network speed comparisons between the two, or maybe a comparison to other Windows Phones on AT&T's network (though AT&T could be limiting all of them).
If anyone is seeing speeds above 7.2 Mbps on their AT&T Titan it would prove my theory wrong, so please report if you've seen 8-12 Mbps on your AT&T Titan at any time on AT&T's network. These speeds may seem high, but they're fairly common now in HSPA+ areas with advanced backhaul.
The only other possibilities are poorly set TCP buffer sizes or just flat horrible radio hardware, but knowing AT&T the speeds we're seeing are artificially limited.
I used this site in the browser for most testing: http://www.ispgeeks.com/wild/modules.php?name=Bandwidth_Meter_DSL I also used the BandWidth app as well as several other sites to verify my findings.
edit: Forgot to give you an idea of how large the disparity is. At my house outside of Atlanta there is no advanced backhaul and I'm getting 2-3 Mbps on the Titan compared to 3-4 Mbps on my Android phones. In the city where there is advanced backhaul in place I'm maxing out around 4 Mbps on the Titan compared to 8-10 Mbps on my Android phones.
Last edited: