Design of mobile phone RF power control loop

JasonYoo

Design of mobile phone RF power control loop [Copy link]

In order to ensure the capacity and interoperability of the system, the GSM system specification has strict regulations on the accuracy, flatness, spectrum purity and out-of-band spurious signals of mobile phone transmission power, and puts forward high requirements on the design of the power control loop of the mobile phone RF power amplifier. This paper introduces the power feedback control method and the current detection feedback control method, and gives detailed design steps for the first method.

The GSM system is a time division multiple access (TDMA) system. Different users are separated on the time axis, and each user receives or sends information in a specific time interval (time slot). This feature of the TDMA system greatly improves the spectrum utilization, but also poses a great challenge to the design of the mobile station's RF front end. The GSM system requires the mobile station's transmitter to work in burst mode, that is, it is only turned on to send information in the specified time slot, and is turned off in other time slots. This switching working state will cause a large number of spurious components in the transmission spectrum, which seriously affects other users. In order to ensure system capacity and interoperability, requirements must be put forward for the indicators of mobile station transmitters, which are all stipulated in the ETSI GSM 11.10 series of specifications.

Figure 1 PVT power-time template

In order to meet the regulatory requirements, the rising and falling edges of the mobile station transmitter signal cannot be too steep, but must be a slow rise and fall process, as shown in Figure 1. The top and bottom two curves in the figure are called power-time templates. During the test, the power-time relationship curve of the transmitted signal in each time slot cannot exceed this template, otherwise the transmission spectrum purity will not meet the requirements or the transmission information will be lost. The middle curve is the gain control voltage of the RF power amplifier, which is given by the system control unit to control the RF output power. This requires accurate power control of the RF power amplifier in the transmitter. At the same time, the GSM mobile station transmitter must be able to work at several power levels according to system requirements, which also requires accurate power control. For this purpose, a feedback control loop must be used. There are many methods to achieve power control. The more commonly used one is the output power detection feedback control method, which directly detects the RF output power and realizes closed-loop power control through a feedback loop. Another method is the current detection feedback control method, which detects the current of the final power amplifier tube and then controls the output power through a feedback loop.

Output power detection feedback control method

Figure 2 Mathematical model of the power control loop

To facilitate analysis, the mathematical model of the power control loop is first given, as shown in Figure 2.

The feedback control system consists of five parts:
1. Comparator: This component is responsible for comparing the difference between the control signal SC sent by the system command unit and the feedback signal SF, and multiplying it by the gain Ks to give the error signal SE to the integrator.
2. Integrator: From the following analysis, it can be seen that the purpose of adding the integrator is to make the output voltage Vo only depend on SC and the feedback gain KcKd, and have nothing to do with the amplifier gain Ka, thereby improving the loop control characteristics.
3. Amplifier: It is a radio frequency power amplifier, and the gain can change with the change of the external control voltage, and the gain is Ka. When the external control voltage is lower than a certain value Vthreshold, the amplifier is not turned on and there is no output signal.
4. Coupler: The coupler is a power sampling component that can extract a small amount of radio frequency power. The gain is Kc=10[-CF/20], where CF is called the coupling coefficient.
5. Detector: The detector is responsible for average detection of the radio frequency signal sent by the coupler, and obtains the corresponding DC voltage SF as the feedback signal. The gain of the detector is Kd.

When the control loop is closed, SC is used as an input of the power control loop to set the output power. Vo is the output of the power amplifier. The coupler extracts a portion of the RF energy and converts it into a feedback signal SF through the detector. Then, it is processed with SC through the comparator to obtain the error voltage SE, and then the control voltage of the power amplifier is obtained through the integrator. This process can be expressed as: Eq1 Eq2
The rate of change of Vo over time can be expressed as: Formula 3

In steady state, dVo/dt=0, so Vo=SC/KdKc. This shows that the RF output power is only related to the control voltage and the gain of the feedback branch, but not to Ka. This is the basic characteristic of the feedback control loop with an integrator.

Design of output power detection feedback control circuit

The detailed design steps of the power control loop are explained below using the example shown in FIG3 .

In Figure 3, D1, D2 and R4 form a dual Schottky diode detector circuit. The use of D1 and D2 in pairs can compensate for the influence of the temperature coefficient. In this example, the gain of the detector is 0.45 (-7dB), and the tolerable input signal range is -20dBm--+20dBm.

R5, C3 and U1A form a comparator and integrator, which is responsible for comparing the output of the detector and the control signal SC, deriving the error voltage SE and integrating it.

In the figure, the gain Kc=10[-CF/20], where CF is the coupling coefficient. In the design of the entire loop, the selection of the coupler and the determination of the integrator time constant are critical. Improper selection of the former will cause the amplitude of the coupled signal to exceed the dynamic range of the detector, while the latter determines whether the loop can complete the power-on lock within the specified time. The GSM specification requires that the minimum power level of the mobile station is 5dBm and the maximum is 33dBm (the above values are all measured at the antenna). In this example circuit, the minimum power that the detector can detect is -20dBm and the maximum power is 20dBm. In the initial stage of the power control loop, the system control unit must first give a smaller power control signal to enable the loop to complete the lock and enter the tracking state. This initial power control signal is called Vpedestal. Vpedestal cannot be too large. The GSM specification states that the value should be 1-6dB lower than the minimum power level. Here, 4dB is selected for calculation:

Vpedestal=(Pmin+Loss)-Pmargin=(5dBm+1dB)-4dB=2dBm

Where Loss is the insertion loss of the device connected to the power amplifier. In order to prevent the feedback RF signal from being lower than the minimum detectable power of the detector, the coupling coefficient of the coupler should have a margin. Here, the margin safety factor (Safety Factor) is taken as 3dB. Considering the above factors and calculating in the worst case, it can be known that:
CF≤Ppedestal-Pmindet-Safety Factor
= 2dBm-(-20dBm)-3dBm
= 19dBm

At the same time, in order to avoid overloading the detector:
CF≥(Pmax+Loss)-Pmaxdet+Safety Factor
=(33+1)dBm-20dBm+3dB
=17dB

Where Pmax is the maximum transmit power level of the mobile station (33dBm), Pmaxdet and Pmindet are the maximum and minimum tolerable powers of the detector respectively.

The GSM specification also places requirements on the lock time of the power control loop, as shown in Figure 2.

When the loop is just powered on, the RF power amplifier has no power output and the loop is not closed because the voltage at its gain control terminal has not reached Vthreshold. In this way, the input of the integrator is only SC, which needs a certain amount of time to initialize in order to reach Vthreshold and close the control loop. In the first few microseconds, the system instruction unit outputs a very small voltage Vpedestal, and the integrator continuously integrates this constant voltage until it reaches Vthreshold. The power amplifier has an output signal to close the loop. At this time, SC can follow the step-shaped curve shown in the figure until a stable power output is achieved.

As can be seen from the figure, this time is actually the duration of the Vpeddstal state, which is specified as 8 microseconds in the specification. During this time, the loop must complete the locking using the given initial control signal Vpedestal, which actually puts forward requirements for the selection of the integrator time constant. According to the characteristics of the first-order loop, the locking time can be approximated by the following formula:

Tlock=Vthreshold×C×R/Vpedestal

To speed up the locking of the loop, a "coarse adjustment" voltage Voffset can be added to the output of the integrator, which together with the output of the integrator forms the control voltage of the power amplifier. This is achieved through U2A in Figure 3. At this time, the loop locking time becomes: Tlock = (Vthreshold-Voffseet) × C × R / Vpedestal

Current sensing feedback control

The power control method is current feedback control type, which realizes power control by detecting the current of the final power amplifier tube, as shown in Figure 4.

Corresponding to different output powers, the RF power amplifier requires different currents from the power supply. As can be seen from the figure, the current sampling resistor detects this change in current, compares it with the SC as feedback information, and integrates it to obtain the power amplifier control voltage, thereby realizing closed-loop control of the output power.

The advantage of this method is that it can save components (couplers, detectors and related peripheral devices) and simplify system design. However, since this method does not directly detect the output power, the relationship between the current and output power of the RF power amplifier is relatively complex and is related to many time-varying factors, so the control accuracy is not as high as the power detection method.

Conclusion

GSM specification 11.10 places high demands on the accuracy, tracking speed and stability of the mobile station transmitter power control loop. At present, the power detection method using coupler-detector is the most commonly used power control method with the best performance and the widest application range. In order to ensure the performance of the loop, it is necessary to carefully consider the dynamic range and thermal stability of the detector, the choice of coupler, the choice of integrator time constant, and the addition of "coarse adjustment" voltage.