Qualcomm X65: 10 Gbps peak data rate

With the announcement of the Snapdragon X65 and X62, Qualcomm boosts 5G data rates, lowers power consumption, and supports 3Gpp Release 16.

As telecom carriers keep building out 5G networks, phones must keep up. To that end, Qualcomm has announced the Snapdragon X65 and Snapdragon X62 modem-RF systems for mobile phone, fixed wireless access (FWA), and industrial IoT applications.

With the X65, Qualcomm claims peak download speeds of 10 Gbps. The modem supports bandwidth up to 1 GHz for mmWave and 300 MHz for sub-6 GHz frequencies. The X62 has a peak data rate spec of 4.4 Gbps with 400 MHz bandwidth for mmWave and 120 MHz bandwidth for sub-6 GHz. Both the X65 and X62 support carrier aggregation, FDD and TDD communications, and dynamic spectrum sharing in accordance with 3GPP Release 16.

With the modem being a part of an RF system, Qualcomm also announced the QTM547 mmWave antenna, which supports worldwide mmWave frequencies including the new n259 (42 GHz) band. The antenna also supports up to 1 GHz channel bandwidth and 2×2 MIMO. The company also announced the QET7100 Wideband Envelope Tracker, part of an RF front end.

There’s always more to a product companies provide in press materials. To learn more, we asked Qualcomm’s Ignacio Contreras, Sr. director of 5G marketing, some questions that engineers might ask.

5GTW: Qualcomm claims a peak download speed of 10 Gbps. What speeds can users generally expect to see?
Contreras: The 10 Gbps download spec is just that, a spec showing what the modem can do. In a real wireless environment, it’s practically impossible to reach that speed. For sub-6 GHz frequencies, you can expect to get about 20% to 25% of the peak specified speed. For example, suppose your device has a peak data rate of 2 Gbps, you can expect to get about 500 Mbps.

With mmWave, however, we see users getting speeds that are about 50% of the peak specified rate. That’s because of the focused signal beam. The Qualcomm campus has a mmWave base station. With an 800 MHz channel bandwidth, we’ve measured peak speeds of 3.3 Gbps when the modem we tested had a peak speed of 7.3 Gbps. With 10 Gbps peak speed, you can expect to see about half of that in regular use. Those speeds depend on conditions such as channel quality and thermal management. In our tests with 400 MHz channel bandwidth, we achieved data rates of about 2 Gbps.

5GTW: The announcement mentioned that the X65 has 30% better power efficiency. Compared to what?
Contreras: Power efficiency depends on many factors such as transmit chains and receive chains. In this case, the improved power efficiency comes from the QET7100 Wideband Envelope Tracker, our 7th generation. We designed the modem-RF system for all-day battery life. The 30% better efficiency we specify is compared to competitors.

5GTW: How does the envelope tracking 7th generation improve power efficiency over previous generations? For example, does it track the envelope tighter? Faster? How does it do that?
Contreras: It has to do with covering the wider bandwidth of the channels. Our previous generation envelope tracker could handle 100-MHz wide channels and before that, we had LTE. Today, the envelope tracker supports up to 1 GHz channel bandwidth and it can track the outputs of multiple power amplifiers. The envelope tracking can cover both 4G and 5G, resulting in a reduced size. The improved power efficiency comes from having one envelope tracker handling multiple power amplifiers. [Contreras did not specify how many amplifiers, but the product brief states that the modems support 2×2 MIMO for mmWave and 4×4 MIMO for sub-6 GHz frequencies.]

5GTW: Qualcomm also announced an “AI-Enhanced Signal Boost” feature that can dynamically adjust the antenna’s aperture and impedance based on actual conditions. What data does the AI software use and how does it process the data?
Contreras: The AI Signal Boost can improve connectivity by up to 3 dB in the transmit chain and 4 dB in the receive chain. It does that by collecting signal data from the antennas. From the data, the AI-trained model tries to guess the phone’s environment such as the phone on a table, you’re holding it while talking, or you’re using it to watch a video in a landscape orientation. It can even detune and antenna when, say, you’re using the phone for a game while your thumb is covering an antenna. It can then enhance the operation of an uncovered antenna to compensate. Previously, we predicted the environment using fixed algorithms. The trained AI improved accuracy by about 30% over the fixed algorithms.

5GTW: Qualcomm has moved beyond mobile and into addressing the “digital divide” with FWA. The Fixed Wireless Access Reference Design (gen2) looks like a finished product as opposed to a development board. How can engineers use the reference design?
Contreras: The reference design is a finished product, from which was can share our system architecture with customers looking to design our ICs into customer premises equipment. We share schematics and software. We also share the details of the third-party components in the reference design. Engineers use the reference design as a starting point for their designs.