I mean for carriers that don't have 20 MHz of contiguous spectrum, they can aggregate 10 + 10 together and get to that 100 resource block, 150 Mbps rate that 20 MHz would give.
But in that case, the chip is able to aggregate 2x 20 as well, if there was such network, isn't it? If so, it is quite odd that neither Qualcomm nor you in this article emphasize it could do even 300 Mbps.
Also I've just noted: "inter-band and intra-band carrier aggregation for WCDMA, in addition to (..) DC-HSPA+." Wait. DC-HSPA+ is carrier aggregation already, what's the addition?
Not necessarily. If they don't have enough calculation ability on board to perform the 2x as many FFTs required for 2x20MHz throughput, not to mention also calculating the FEC, then they can't do it.
And I would imagine that by now Qualcomm are very good at including on their chips precisely as much calculation as is needed for a particular target throughput and nothing more --- no point in wasting die and power on ability that is never used.
Release 8 W-CDMA supports HSPA+ as dual carrier or MIMO but not both. Furthermore, it supports DC-HSPA+ but only with directly adjacent dual carriers. Release 10 W-CDMA supports all of the above, and dual carriers can be adjacent, non adjacent but within the same band, or even in different bands (e.g. Cellular 850 MHz and PCS 1900 MHz).
In the LTE spec. a resource block is 180Khz, 100 of which make the 20mhz bandwidth, which is not to be confused with the band. With this maximum bandwidth and at 64QAM your antenna can only(!) spit 75mbps at the cell-tower. Thankfully you have two antennae, so you get 150mpbs. Whether you carve this into two 10mhz channels as an aggregate or have one 20mhz channel, you won't get past 150mbps. With Cat5 LTE you will have 4 antennae available, and thus 300mpbs.
Then lets hope we ditch 3G UMTS/WCDMA asap and move on to LTE. The constant switching from LTE to 3G / 2G signals only for voice is seriously killing the battery.
Answering my own question, on the Qualcomm page it seems this new 9x25 will be using the same 28nm process, but with smaller Die Size and lower power usage! Sounds like another candidate for iPhone 5S to me. and that is just a amazing piece of engineering.
The other thing is i have always thought LTE advanced will be more like 5G, as in how we move from 2G to 3G, now that it seems more like a HSDPA to DC-HSDPA thing. The max speed of the purposed 1Gbps will be achieved by aggregate 100Mhz of spectrum . Which gives a efficiency of 200Mbps per 20Mhz spectrum. Compared to Category 4 LTE modems on 20 MHz FDD-LTE see up to 150 Mbps of throughput. We have an 33.3% improvement in bandwidth. Which is still a very good upgrade, and may be we get lower latency number as well, but not anything close to the advertised 1Gbps. I mean which carrier actually own 100Mhz of spectrum? And decided to use it solely for LTE-A?
I suspect that 100mhz blocks are unlikely to show up anywhere outside the lab, possibly excluding rural broadband initiatives. In the bigger cities there isn't any TV spectrum available to be freed up without kicking stations) off the air; so short of either shutting broadcast TV down entirely (or a major restructuring to a distributed transmission setup like phone service) the only places large amounts of TV spectrum are going to be available to be freed are sparsely populated areas in the country. Those areas don't need more bandwidth for phones; but for residential wireless broadband with usage limits comparable to wireline services lots of spectrum would be needed. Combined with cell sizes of up to 100km this could push usable broadband out to everywhere except in the mountains (where LoS issues will also cause major problems for satellite based services).
I work in the embedded PC world, where we use miniPCIe WWAN modules. Years back, there was a Gobi 3000 module from Qualcomm we'd use for data on 2G/3G Networks from all the domestic carriers (AT&T, Verizon, Sprint, ect), switching between networks by changing the firmware.
Does this chipset theoretically provide the same functionality as Gobi3000, but adding LTE to the mix?
Brian Klug 11/18/2012: "We have a lot more information about MDM9x25 and the current state of Qualcomm's upcoming modem and transceiver portfolio that we'll be going over in a larger piece in the coming week."
The World 12/15/2012: still waiting for that piece.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
14 Comments
Back to Article
xdrol - Sunday, November 18, 2012 - link
"aggregate together enough spectrum to get effectively 20MHz. "Does that mean, it cannot aggregate 2x 20 MHz, only e.g. 2x 10 MHz?
Brian Klug - Sunday, November 18, 2012 - link
I mean for carriers that don't have 20 MHz of contiguous spectrum, they can aggregate 10 + 10 together and get to that 100 resource block, 150 Mbps rate that 20 MHz would give.-Brian
xdrol - Sunday, November 18, 2012 - link
But in that case, the chip is able to aggregate 2x 20 as well, if there was such network, isn't it? If so, it is quite odd that neither Qualcomm nor you in this article emphasize it could do even 300 Mbps.Also I've just noted: "inter-band and intra-band carrier aggregation for WCDMA, in addition to (..) DC-HSPA+." Wait. DC-HSPA+ is carrier aggregation already, what's the addition?
name99 - Sunday, November 18, 2012 - link
Not necessarily. If they don't have enough calculation ability on board to perform the 2x as many FFTs required for 2x20MHz throughput, not to mention also calculating the FEC, then they can't do it.And I would imagine that by now Qualcomm are very good at including on their chips precisely as much calculation as is needed for a particular target throughput and nothing more --- no point in wasting die and power on ability that is never used.
WiWavelength - Monday, November 19, 2012 - link
Release 8 W-CDMA supports HSPA+ as dual carrier or MIMO but not both. Furthermore, it supports DC-HSPA+ but only with directly adjacent dual carriers. Release 10 W-CDMA supports all of the above, and dual carriers can be adjacent, non adjacent but within the same band, or even in different bands (e.g. Cellular 850 MHz and PCS 1900 MHz).AJ
Kogies - Monday, November 19, 2012 - link
In the LTE spec. a resource block is 180Khz, 100 of which make the 20mhz bandwidth, which is not to be confused with the band. With this maximum bandwidth and at 64QAM your antenna can only(!) spit 75mbps at the cell-tower. Thankfully you have two antennae, so you get 150mpbs. Whether you carve this into two 10mhz channels as an aggregate or have one 20mhz channel, you won't get past 150mbps. With Cat5 LTE you will have 4 antennae available, and thus 300mpbs.iwod - Sunday, November 18, 2012 - link
What is happening to VoLTE? Does it needs Chipset support to work? If yes i wont be buying a LTE devices soon.Would this be lower power then 9x15?
A5 - Sunday, November 18, 2012 - link
As far as I know, VoLTE doesn't require anything beyond an LTE radio.Since the whole point of VoLTE is that it treats voice the same as data, it just requires carrier and software support once you have an LTE device.
iwod - Sunday, November 18, 2012 - link
Then lets hope we ditch 3G UMTS/WCDMA asap and move on to LTE. The constant switching from LTE to 3G / 2G signals only for voice is seriously killing the battery.Answering my own question, on the Qualcomm page it seems this new 9x25 will be using the same 28nm process, but with smaller Die Size and lower power usage! Sounds like another candidate for iPhone 5S to me. and that is just a amazing piece of engineering.
The other thing is i have always thought LTE advanced will be more like 5G, as in how we move from 2G to 3G, now that it seems more like a HSDPA to DC-HSDPA thing. The max speed of the purposed 1Gbps will be achieved by aggregate 100Mhz of spectrum . Which gives a efficiency of 200Mbps per 20Mhz spectrum. Compared to Category 4 LTE modems on 20 MHz FDD-LTE see up to 150 Mbps of throughput. We have an 33.3% improvement in bandwidth. Which is still a very good upgrade, and may be we get lower latency number as well, but not anything close to the advertised 1Gbps. I mean which carrier actually own 100Mhz of spectrum? And decided to use it solely for LTE-A?
DanNeely - Monday, November 19, 2012 - link
I suspect that 100mhz blocks are unlikely to show up anywhere outside the lab, possibly excluding rural broadband initiatives. In the bigger cities there isn't any TV spectrum available to be freed up without kicking stations) off the air; so short of either shutting broadcast TV down entirely (or a major restructuring to a distributed transmission setup like phone service) the only places large amounts of TV spectrum are going to be available to be freed are sparsely populated areas in the country. Those areas don't need more bandwidth for phones; but for residential wireless broadband with usage limits comparable to wireline services lots of spectrum would be needed. Combined with cell sizes of up to 100km this could push usable broadband out to everywhere except in the mountains (where LoS issues will also cause major problems for satellite based services).Peanutsrevenge - Monday, November 19, 2012 - link
Until the inevitable happens and we finally switch to IPTV.That would free up a nice chunk of spectrum.
tuxRoller - Monday, November 19, 2012 - link
Is this what TMo is supposed to be targeting with their LTE rollout next year?RU482 - Wednesday, November 21, 2012 - link
I work in the embedded PC world, where we use miniPCIe WWAN modules. Years back, there was a Gobi 3000 module from Qualcomm we'd use for data on 2G/3G Networks from all the domestic carriers (AT&T, Verizon, Sprint, ect), switching between networks by changing the firmware.Does this chipset theoretically provide the same functionality as Gobi3000, but adding LTE to the mix?
Klug4Pres - Saturday, December 15, 2012 - link
Brian Klug 11/18/2012: "We have a lot more information about MDM9x25 and the current state of Qualcomm's upcoming modem and transceiver portfolio that we'll be going over in a larger piece in the coming week."The World 12/15/2012: still waiting for that piece.