Emerging RF Technologies for Smartphones and Connected Devices
By Ben Thomas, Director of Marketing, Advanced RF Platforms, RFMD®

As we approach the new year, we have a much better idea of what market dynamics and resulting technologies will be driving handset RF developments in 2013. From the market side of the equation, mobile data usage continues to grow as rapidly as expected and mobile data creation is clearly on the rise. Additionally, cloud-based services will be larger than ever, providing access to streaming content, along with the security and convenience of not having files resident on an individual mobile device.

This insatiable focus to consume, create, and communicate data has a profound effect on RF component developments and is leading to the most impactful cellular technology transitions recorded over the last decade. The most prominent RF-related technologies will be carrier aggregation (CA), envelope tracking (ET), and antenna control solutions (ACS). From an active RF view, ET is certainly front and center, and so it is with this background that we peer into this complex topic to understand why ET is on the rise and the impact it is predicted to have.

LTE’s Network Efficiency Carries RF Challenges
The consumer’s demand for data consumption is driving the rapid adoption of the 3rd Generation Partnership Project’s (3GPP) Long Term Evolution (LTE) wireless communication standard at mobile operators (MOs) across the globe. In its September 11, 2012 update, the Global Mobile Suppliers Association (GSA) reported 96 commercial LTE networks in 46 countries, and has forecasted 56 more by the end of 2012. Despite the decision to use a common 3GPP communication standard across the globe, the frequency bands for LTE deployment are heavily fragmented. The downstream effect of supporting LTE across fragmented frequency bands in mobile devices is tremendous RF complexity. And, while our industry thrives as suppliers create value by solving these complex RF problems, it is consumers who are set to pay the ultimate price of LTE’s deployment—shorter battery life for their LTE-enabled smartphones.

Before those who architected the LTE standards label this idea heresy, let’s examine the premise behind this shorter battery life. There is no argument that the concept driving the LTE standards creation remains steadfast: increase data bandwidth while decreasing the watts-per-bit consumed when transmitting the data. In other words, for a given 100kb of data, LTE is a more efficient transmit modulation. Of this there is no argument. It is with the practical application of LTE that we find the negative impact on current consumption, stemming from two main instigators—frequency band effects and output power requirements.

Recall that while LTE will be overlaid on existing GSM/EDGE/WCDMA frequency bands, most implementations will include new frequency bands, specifically allocated to LTE. With the need for backward compatible to existing MO infrastructure, LTE’s additional frequency bands mean increased post-PA loss derived from a higher number of antenna switch ports, more antenna tuning elements, or simply a less efficient antenna design that comes from the need to cover a broader frequency range. Next, the duplex filtering requirements of many of these new LTE bands are very demanding, often driving up duplexer insertion loss, over and above the average for a traditional WCDMA frequency band. The results being higher PA output power (POUT) in order to transmit the same output power from the mobile device (PANT). Whether higher POUT or higher PANT, PA current consumption rises.

LTE’s output power requirement itself could be a topic of lengthy discussion, but we can stay with the basics to drive the next point home. The 3GPP standards body made an interesting decision with regard to LTE output power: transmission power can be 1dB or more reduced in comparison to WCDMA output power even in the baseline LTE modulation case (WCDMA nominal, maximum PANT is +24dBm +1/-3dB, whereas LTE is +23dBm +2/-2dB with multiple further power back-off conditions as an LTE modulation’s peak-to-average (PAR) increases). It is a conventionally held wisdom that this PANT reduction is derived from the simple fact that trying to hold POUT at the same level as WCDMA will require a dramatic increase in PA power consumption. A demonstration of this point is the PA efficiency comparison of RF7245 designed for WCDMA with that of RF7305 designed for LTE (Figure 1).

While this seems a reasonable reprieve for PA designers, in practical application this causes problems. Each successive communications technology has decreased transmit power from handsets, from GSM to EDGE to WCDMA, and now to LTE. It quickly becomes a matter of MO infrastructure reutilization versus new construction. If MOs wish to reutilize existing infrastructure, and a 23dBm signal does not propagate as far as a 24dBm signal, network coverage is diminished. In short, LTE is subject to the same basic signal propagation dynamics as any other communications standard—you can’t avoid physics. Now, think about this in terms of all the “widest coverage available” ads we see on television—not good. To overcome this, more base stations are needed, which drives more capital investment to roll out a technology—again, not good. It is RFMD’s prediction that we will see average LTE PA power outputs rise and current consumption along with it, in order to allow MOs to deliver the best network coverage possible. Additionally, as soon as a technology is available that can reasonably raise the output power capabilities of handsets without a dramatic increase in current consumption, MOs will drive OEMs to utilize this technology and drive LTE PANT higher, returning it to the levels desired for WCDMA, at a minimum.

Now, let’s turn our attention back to the point that backward compatibility is required and new LTE bands will be added, thereby increasing the total number of frequency bands covered in any given handset. For example, in a regional-focused, North American (NA) handset, quad-band GSM/EDGE plus bands 2 and 5 for WCDMA are implemented, and LTE then adds bands 4 and 17. In Europe (EU), the same quad-band GSM/EDGE is implemented along with band 1 and, often, band 8 for WCDMA with LTE driving the addition of bands 3, 7, and 20. These additional bands drive additional RF content, leading to increased cost and solution size. For the global coverage case, the content and size problem is mind-boggling. As one could imagine, given the consumer product nature of today’s handsets, cost reduction is the aim, so an increase in cost is untenable, long term. Equally impactful is the LTE solution size growth, because handset form factors are stabilizing around screen sizes, while digital hardware content is growing—all resulting in a decreased area for RF implementation. A solution, growing in popularity, to these two concerns is the utilization of a multi-band (MB) PA architecture. An MB PA is a single power amplifier chain that is designed to cover a wider frequency range, followed by a post-PA switch to route the amplified signal to the frequency dependent duplexer function (Figure 2).

Unfortunately, the post PA band-select switch comes with insertion loss, and equally impactful is the broadbanding of the PA to maintaining band-edge linearity requirements. In combination, we can see as much as 6 points of power-added-efficiency (PAE) drop compared with a PA designed for a single frequency band. So, while MB dramatically helps solve the LTE size and cost adders, we are again faced with increased PA current consumption. As much as handset designers may not want this increase in current consumption, there is little option given the cost and size targets of today’s smartphones.

Enter ET
As depicted in Figure 3, ET describes an approach to RF system design in which the power supply voltage applied to the collector of a power amplifier follows the RF envelope, maintaining the PA in a saturated state (operating in compression) at every point, thereby maximizing its efficiency.

This is in contrast to today’s LTE PAs, which operate with a voltage applied to the collector based on the linearity requirements needed for a given power output and modulation, termed average power tracking (APT). Although these APT techniques do improve efficiency as compared to the fixed supply voltage application depicted in Figure 3, there is still a great deal of energy dissipated as heat simply because the PA operates in a less efficiency linear region and the voltage does not dynamically move with the envelope.

Now, most approaches targeted today for mobile cellular applications, although coined ET, are not able to operate the PA in full compression (PSAT) as described above, but rather utilize a region of amplifier operation slightly below in order to comply with the stringent system requirements (particularly noise) that accompany LTE’s highly complex modulation schemes. The relative efficiencies of these three power amplifier operating points are approximately 42%, 56%, 61% for PAPT, PET, and PSAT, respectively.

From this rudimentary description of such an advanced technology as ET, one can quickly see the overarching benefit and how this translates directly to improved current consumption. Thus far, however, we have only considered the impact from PA efficiency. To be complete we must evaluate system efficiency, which includes the performance of the DC-DC converter that is critical to achieving either power reduction implementation. This comparison can be quite complex when considering ET DC-DC converters, because the DC-DC efficiency varies with operational conditions. If we normalize to a maximum power PA load condition and a given LTE modulation, however, we can more easily compare solutions at the system efficiency level—a term that has been coined “system lift.” Under these conditions, we can weigh current consumption savings alongside the additional cost and solution size involved in ET versus APT—all culminating in a cost-benefit analysis for implementing ET.

As the number of ET solutions increases, and the adoption of MB for LTE becomes more mandatory, we find the emergence of a common target justifying the cost and size of ET implementation. If ET can enable an MB LTE PA to achieve the same current consumption as an SB APT PA, e.g. ≥0% system lift, we have a solution that gains the cost and size savings of MB without paying a current consumption penalty. When this level of system efficiency is achieved, we will have viable MMMB solutions for widespread OEM adoption. (Figure 4).

Once we have achieved the primary current consumption goals, additional ET benefits emerge, such as increased LTE PANT. As shown in Figure 5, we can see that when ET is implemented, the POUT of the PA can increase while improving its efficiency as the PA operates close to compression. Linearity margin decreases, but today’s measurements show that linearity is not a limitation as it typically is in high power LTE APT transmit conditions. Even with margin decreasing as much as 4dBc for ACLR, we are well within 3GPP specifications and still provide OEMs manufacturing margin—all while increasing POUT; yet another benefit of ET shining through.

Aside from improved current consumption at nominal LTE power levels, the ability to increase POUT could also facilitate MO demands for higher PANT and thus better network coverage. In fact, early measurements show that with an MB ET PA operating 1dB higher in POUT, aligning with WCDMA power targets, still achieves lower current consumption than early 2013 handsets that have, due to size and cost constraints, implemented MB APT PAs.

Whether the results are longer battery life or better network coverage, there is clear data emerging that indicates ET will benefit consumers. As we implement this first generation of handset ET and begin work on the second, RFMD is confident that we will continue to see benefit from this emerging handset technology. For now, we can rest assured there are RF engineers working to ensure that the smartphones we can’t live without will give us the battery life we need, while LTE helps us consume data, create experiences, and communicate our life everywhere we want to be, globally.

RFMD®
www.rfmd.com
Email this article to a friend!