Frequently Asked Questions
§ What is Mobile Transmit Diversity (MTD)?
o How does it work?
Mobile Transmit Diversity (MTD) concept is based on sending same replicas of a signal over two different antennas, either via selection or simultaneously; both methods require feedback from the network: while the former needs to know which antenna experiences lower path loss to the serving cell, the latter needs an indication directing towards optimal phase difference between the two antennas, in order to bring about coherent combining of the two signals at the receiver (also referred to as Beam Forming).
The 3GPP has labeled a MTD solution based on specific set of dynamic feedback signaling from the base station as CLTD (Closed Loop Transmit Diversity), and MTD solution based on preexisting Power Control Bit feedback as OLTD (Open Loop Transmit Diversity).
OLTD concept description for the case of simultaneous transmission from both antennas, is described in Fig 1 below, and is based on small gradient seeking up/down phase perturbations, that reach settling point when TPC commands cease to prefer one over the other; as fading has significant impact on this info short term relevance, various filtering and pattern recognition processes are implemented, facilitating beam forming tilt damping. The algorithms constantly learn the ambient channels and adapt the parameters for optimum performance.
Fig. 1: OLTD beam forming using Power Control feedback
§ What are the benefits of MTD?
o For Carriers
MTD is typically and addition on top of MRD (Mobile Receive Diversity), i.e. Antenna Diversity enabled UE, which enjoy DL range extension and capacity boost, will be upgraded to regain range balance between DL and UL, as well as to boost UL capacity. Of particular importance are cell edge UEs, which are typically the soft spot of wireless communications, bringing about the majority of customer dissatisfaction. MTD will address this weakness as follows:
- Match uplink range with down link one
- Support DL enhanced speed at cell edge (compromised by unjustified high DL retransmission rate due to UL failure)
- Boost overall UL capacity by some 30%, and cell edge data rates by 50-100%
o For users
- Indoors cellular service is still a major source of customers' dissatisfaction in both rural and urban environment, perhaps because people spend the lion's share of their time in buildings; voice calls drop as well as data sluggish performance still cause people to wonder around or step towards a window in order to gain few more dBs; the introduction of Home base stations (HNB) will no doubt improve this situation, but is not likely to change the pattern of people moving around to get into these HNBs due to their severe range extension (caused by their Closed Subscribers Group handicap)
- MTD beamforming is essentially a mechanism of increasing the UE antenna directivity which equals to higher antenna gain (about double); given equal path-loss to the base station, which is equivalent to equal EIRP, a higher antenna gain yields lower total Pout requirement (about ½) vis-à-vis a conventional UE; thus the PA current consumption is significantly dropped (by almost ½, save for some current spent on additional required circuitry); note that at either cell edge or while transmitting peak data rate, the PA current budget is close to some 2/3 of the consumption, therefore cutting this part by close to 50% increases battery life by some 30% at these situations
o For UE vendors
The above benefits for both Cellco and subscribers will create a desirable differentiator for MTD capable UEs; note also that while CLTD may be a standard feature starting at REL 11, it will not be downward compatible to previous network releases, however OLTD will apply to some 90% of legacy HSPA network, with the support of a Network Control patch (see details in following sections). Note also the OLTD implementation in a UE that already incorporate CLTD, is software only upgrade, and no hardware augmentation required. Therefore UE vendors will use it as a differentiator.
o For chip makers
As described above, REL 11 UE chip makers will incorporate the CLTD feature, and so OLTD addition will be just a software increment, allowing these baseband and/or Radio chips to differentiate themselves by supporting both REL and legacy network with MTD feature.
§ Does MTD compatible with 2.5 G, 3G and 4G networks and devices?
o OLTD was tested with CDMA 1x, CDMA2000, EVDO REV0, EVDO REV1, UMTS R-99, UMTS HSUP, UMTS HSPA; it is now being developed for LTE REL 8,9 which have no UL MIMO support.
§ Can MTD work with legacy handheld devices?
o No, UEs need enhancement to include the capability, both H/W and S/W (see next sections)
§ Can MTD work with legacy network?
o Yes, OLTD does not require any new signaling from the base station, which is in fact oblivious to the UE OLTD feature (unlike CLTD that requires dynamic control by the network). Legacy will be augmented by a software patch that will facilitate an option to turn the MTD feature on or off.
§ What are the MTD implementation aspects?
o What does it take to implement MTD in the UE?
Both CLTD and OLTD will implement a 2 chain transmitter on top of the pre-existing 2 chain receiver, similarly to LTE-A architecture which will support 2 chain Rx/Tx; the following additions are required:
- A two chain transceiver chip, replacing the single chain current solutions
- A secondary Power Amplifier implemented at one of the following manners:
· Replacing the single conventional PA with a pair of ½ power PAs
· Keeping the conventional full Power PA and adding a secondary ½ power PA
· Using a conventional full Power PA for both primary and secondary antennas, fed by a variable voltage PWR SUP
Note that the first and 3rd alternatives will facilitate the power saving quoted above, while the second alternative will turn current consumption to a wash vis-à-vis a conventional PA
- An additional Duplexer
- Assuming the UE is already served by a secondary antenna for MRD, the specification should be augmented to fit MTD (i.e. correlation factor lower than 0.5, gain difference limited to 1 dB)
- Baseband software augmentation (OLTD algorithms require ~1 MIPS)
o What is the impact on Form-factor?
- As MRD is already implemented in dongles and some Smartphones, a secondary antenna already exists, although it is typically smaller than the primary one; implementing 2 antennas of the same size is currently planned for higher bands (above 1 GHZ) only, and in that case
- Additional components (PA, duplexer) will take some more board space
- Total extra volume of a dongle estimated to grow by some 10%, while Smartphone form-factor impact will be a only couple of percents
o What does it take to implement MTD in the Network?
- As the network is oblivious of OLTD operation, it is essentially unchanged; however a software patch will be available from network vendors, allowing for a turn up / turn down OLTD control at legacy networks; this will not be the case for CLTD which will not be supported by legacy networks, and will require REL 11 and higher network releases
o How does OL co-exist with CL?
- REL 11 networks and UEs will support either one, per the network choice on a UE specific basis
- As stated above, OLTD can be added on top CLTD UEs as a s/w addition
- The expected scenario will be for such a UE will implement Transmit Diversity in both legacy (using OLTD) or in REL11 (Using either OLTD or CLTD)
o How does Beam-Forming (BF) OL compare with Antenna Selection (AS)?
- OL Beam-Forming has superior performance over Antenna-Switching
· BF gain in static environment is higher
· BF gain in vehicular environment is positive with practical antennas versus negative in AS
· BF UE at mixed with conventional UEs do not compromise the latter's performance versus losses to conventional UEs registered in the case of AS
· BF conforms with minimum required Pout (i.e. 24 dBm combines); As requires conformance waver due to the introduction of a RF switch placed between the PA and the antenna
· AS is simpler to implement due to the use is a single PA rather than two
o What are the antennas requirements?
- The GSMA forum has recently unified BF and AS antennas requirements (VER 5)
§ What are the risks for network operators in deploying MTD?
o What are the affects on serving cell base station?
Analysis and simulations show some 0.2-0.3 dB demodulation loss with a generic algorithm; however realistic algorithm tested with a commercial NBs of 2 major vendors, showing ~ 0.1 dB of demodulation loss
o What are the affects on neighboring cells?
One of the main BF contributions, is the reduction of signal level experienced by neighboring cells, positively affecting ROT downwards, thus enhancing their capacity
o What is the cold turn on issue?
- When a UE turns power on and finds a viable network, it goes thru hook up phase using the dedicated PRACH (Physical Random Access Channel); this procedure is carried out prior to the establishment of two way link with the NB, and is composed of increasingly higher level probes that seek acknowledgment.
- Since TPC does not exists during this phase, the OLTD BF algorithm reverts to searching optimal phase via a calculated phase sweep as the probes are progressing
- Tests have shown this procedure to achieve range extension close to 3 dB versus a conventional UE
o What are the affects during hand-over & soft-hand-over?
- During a hard hand-over, the UE reverts to PRACH procedure, so the above cold turn on procedure applies
- A Soft-Hand Over (SHO) situation has been the least understood mechanism as far as beam forming is concerned, due to an apparent lack of degrees of freedoms, i.e. the need to simultaneously optimize received signal in several different locations (base stations), when using only one parameter which is the phase difference between the 2 transmit antenna (geometrically speaking, when the beam is directed towards a given base station, it seems to turn away from the others active participating base stations.
- In order to shade further light in the apparent issue, consider the following example: some two BTSs exhibit similar pathloss towards an MTD UE, so should the UE pick one of them, thus increasing its contribution by say 2 times, it will also be reducing the contribution of the other one by say, ½; so a conventional UE would see 1+1 = 2 while the MTD one will see 1x2 +1+1/2 = 2.5, a far cry from 4 (the factor 4 in this example represents the gain if a single cell will be involved )
- This deterministic example, however, ignores the statistical nature of fading, which constantly changes pathloss and therefore exhibiting somewhat cyclic imbalances between these two base stations; the UE MTD algorithm will in fact be pulled by the dominant feedback, fortifying the stronger and neglecting the weaker, in a rather smooth fading-related pace
- Bottom line is that soft-hand-off gains are slightly lower than single cell ones, but by a relatively small margin (for instance some 0.5 dB of gain reduction were noticed in both field test and simulations
- To summarize, note that soft-hand-over is the default situation in CMDA/HSPA networks (single cell is the exception), and therefore most measurements done in field test conducted with multiple major Cellco, which show some 3-4 db of gain, where done in SHO
o What are the affects on other users in the network?
- When deploying a mixture of MTD and non MTD enabled UE, the slight demodulation noise created by the MTD UE are the base station receiver (a small fraction of a dB), is overwhelmed by the relatively large gain (3-4 dB) granted to the UE by the algorithm; however for the non MTD UEs in that cell, this small base station reduced sensitivity seem to come with no compensating, this actually slightly reduce their performance
- While such a phenomenon will happen if the MTD UEs are completely isolated, a more realistic scenario will be that all neighboring cells will be proliferated with some mix of MTD and non MTD UEs. Simulation of such mixed environments, as well as a field test, have proved that same cell non-MTD UE's fractional loss, is more than compensated by the reduction of interference in the neighboring cells caused by the MTD UEs there
o What is the impact of base station interference cancellation (ICT) on MTD gain?
- As one of the more potent BF contribution is related to interference reduction, one wonders what will be left of this gain contribution once base stations are equipped with ICT.
- Analytical papers prove that when all surrounding base stations are ICT enabled, and the UEs are MTD enabled, the ratio of MTD interference reduction remains unchanged
- IN other words, the ICT gains and the MTD gains are additive
o What will be the impact of Small Cells (Pico/RRH/Femto) on MTD gain?
- The up and coming network topology revolution, splitting cells into PICOs and Femtos, seem to reduce UE range to serving cells dramatically, thus apparently reducing the need for range extension features lie MTD.
- While this is true for noise limited cases , this new topology is going to be dominated by interference limited situation as follows:
- Outdoors Picocells typically deployed under canopy of pre-existing Macrocells, many a time in line of sight to the Macro as well as fir each other, will create multiple & problematic cell edges that will severely reduce their added value; while LTE REL 11 and 12 will address these issue via eICIC and CoMP, HSPA is going to be at least one step behind; therefore MTD will be a useful tool in addressing uplink interference issues
- As for Home BTSs (Femtocells), they were originally designed to enjoy indoors-outdoor sheltering, creating a close to perfect spectrum reuse; however, customers refusal to share access with outdoors random by-passers (CSG), manufactured a pretty sever limitation in the Femtocell's in many cases defeating their purpose as islands of indoors robust coverage; MTD role indoors will be indeed reduced if Femtos penetration will become ubiquitous, but not eliminated
o What is the affect on Specific Absorption Rate (SAR)?
- Conceptually, UE radiation levels will not be affected, as EIRP remains constant; but statistically, the beam will turn more away from the human body rather than towards it, thus reducing average radiation levels by 1-2 dB
- Still in some cases, when the beam is directed towards the human body at max power, SAR will increase due to energy density boost
- For such cases, the BF implementation & algorithms are deterring this increased level via the application of two methods:
· Real time monitoring of such cases and steering the beam slightly away, thus paying with some fraction of gains for the sake of keeping SAR within limits
· A production procedure that uses coupling variations in order to shift peak antenna patterns away from a critical direction
- In summary, average MTD radiation will be go down, while SAR - defined as peak radiation level, will stay unchanged
§ What is the cost associated to deploy MTD on wireless networks?
o UE cost impact
- BOM increase vis-à-vis a conventional UE, assuming transceivers prices will end up equal to today's chips, is represented by one PA + one duplexer
o Network cost impact
§ What frequencies does MTD operate?
o While MTD solution is not frequency dependent, deployment will probably start at frequencies above 1GHZ
§ Comparing BF with MIMO
o For a given serving cell, MIMO will be a preferable solution for cell center and mid cell users
o For cell edge UEs, Signal to Noise & Interference Ratio (SINR) will be insufficient for MIMO and this will be the BF domain; note also that for cell edge for MIMO is twice as large since it will use ½ power PA for each data stream
o As for neighboring cells interference reduction, BF will outperform MIMO, which provides no ROT reduction to the surrounding network
§ What are the alternatives to MTD?
o Improving receiving functionality at the base station (i.e. adding more Rx antenna on cell sites)
o First commercial deployment is anticipated in first half of 2013.
How does one obtain a license to Magnolia's patented technology?