The key power amplifier design: the performance of the output matching circuit

For any power amplifier (PA) design, the performance of the output matching circuits are critical. However, in the design process, there is a problem often neglected by people, that is, the output matching circuit of the power loss. The power loss in the matching network capacitors, inductors and other components of energy. Power loss will reduce the power amplifier efficiency and power output capacity.

Because not a 50Ω output matching circuit components, so dissipation losses and sensor gain a big difference. Specific different output matching circuit, losses are not the same. For the designers, even if he did not choose a different technology room, and the dissipation loss in bandwidth between the design can still do a lot of compromise.

Impedance matching network is used to achieve change, as is the power transfer from one system or subsystem to another system or subsystem, RF designers in it under a lot of effort. For power amplifiers, impedance control of power transmission to the output size, its gain, as well as it produces noise. Therefore, the power amplifier matching network design is the key to optimal performance.

Loss have different definitions, but here we are concerned in the matching network, RF power dissipated as heat loss. The loss of power is not out of any use. Function according to different matching circuits, loss of acceptable range is also different. Terms of the power amplifier output matching loss has been an issue of concern, because it involves a lot of power. Low efficiency will not only shorten the talk time, but also in the great heat dissipation and reliability problems.

For example, a GSM power amplifier working at 3.5V voltage, the efficiency is 55%, to 34dBm output power. The maximum output power, the power amplifier's current 1.3A. Match loss is 0.5dB to 1dB of magnitude, which is the output matching of the specific circuit. In the absence of dissipative losses, the power amplifier efficiency of 62% to 69%. Despite the loss can not be completely avoided, but this example tells us that in the PA matching network, the loss is the primary problem.

Dissipative losses

Now we look at a network of a matching network (Figure 1a) in the dissipative losses. Power through the passive matching network to the passive load transmission power. Between the power supply and load impedance without any other restrictions. The matching network and load together to consider the power output of a fixed amount of power Pdel to the network (Figure 1b). Part of output power in the form of heat dissipation in the matching network. The rest are transferred to the load. Pdel is transmitted to the matching network and load (Fig. 1c) on the total power, PL is transferred to that part of the load power.

Read the two volume, we know that, in fact, in the end how much power is a useful part of the power from the power transmission to the load, the ratio is equal to PL / Pdel.

This is the dissipation loss of power amplifier output matching the correct measurement, because it only considered the actual transmission power and dissipation power. Reflected power is not counted.

Can be seen that this ratio is equivalent to the work of the power matching network gain GP. The work of the power gain full expression is:

The key power amplifier design: the performance of the output matching circuit

Here is the load reflection coefficient, is the matching network s parameters, the key to power amplifier design: the performance of the output matching circuit loss is the gain of the countdown. Therefore, the dissipation loss can be defined as:

Ldiss = 1/GP.

For the power amplifier, we design load for it is generally 50Ω. Usually, we used to measure the s parameters of the system impedance is 50Ω. If the system impedance and the load is 50Ω, then 0, then the above expression can be simplified to:

The key power amplifier design: the performance of the output matching circuit

A matching network in the calculation of the dissipation loss, Zhi Zhi need to know the transmission and reflection of the size of Scattering Parameters, such Ke Yi Henrongyide from the s parameters of the calculation process Dedao, because the network analysis instrument will be used Xian Xing's Fang Shi Tong Chang to display the s parameter. In assessing the level between the input and dissipation loss, the load impedance is not 50Ω, but the rules still apply.

Because the reflection and dissipation losses can easily be confused, RF engineers will sometimes use the wrong method to calculate the dissipation losses. The worst way is to use untreated calculated s21. A typical matching network at 1GHz (Figure 2), of the power amplifier, the values for the 4 + j0Ω is the load impedance. Matching network is used to simulate lossless components, so the network does not exist in the matching power dissipation problem. However, s21 is-6dB, because 4Ω 50Ω source impedance and the load between the huge mismatch. As a lossless network, except for some digital noise, the simulation of dissipative losses 0dB.

In the circuit of the simulation, we may be able to use s21 to find the correct dissipation losses. This process involves the use of complex simulation of the conjugate impedance of the load line as the source impedance. As the source impedance and dissipation losses and no ties, so this is a correct way, but not easy to use.

Another common method is to use the circuit simulator to calculate the maximum gain. Since this measurement using ADS, so it was easier to use. However, it might get the wrong answer. Only 50Ω series resistance in a simple circuit, obviously, the load is 50Ω, 50Ω series resistance of dissipative losses is 3dB, because the transmission power is divided equally among the series resistance and load (Table 1). In this example, the simulator can choose 1GΩ load impedance. When the 50Ω load resistor and 1GΩ series together, it is above the voltage drop is very low, while the power dissipation is also very small.

The correct method of calculation should be the work of power gain using. Other methods may also be able to get the same results, but can not guarantee that the results can be. When the load is 50Ω, it is necessary to work and power gain, is very simple, we have no reason not to use it.

Output matching circuit

Specific different output matching circuit, the final loss is different. The low end of the spectrum in the microwave transmission line takes up so much space, so using the method of lumped elements. In a typical power amplifier output matching circuit module, the use of large volumes of blocking capacitor to prevent DC current from the power amplifier in the power flow to the load (Figure 2). With surface mount capacitors and inductors and surface mount printed inductors formed two low-pass matching network can be the name of the load impedance of 50Ω into the appropriate load line. The load-line setting is based on the specified output power and available power amplifier supply voltage. Phone amplifier load line ranges from 1Ω to 5Ω.

We can use standard or high-Q capacitors. Another popular approach is gradually integrating capacitor is used. In many process technologies (including GaAs and CMOS), the high-quality metal - high dielectric - metal structure of the storage capacitors can all be used. There is a supply of providers do not use any surface mount components to complete the GSM power amplifier module, all the matching network are using the framework of alignment pins and the integration capacitor. In addition to reduced size, the use of integrated capacitors in cost has its advantages, this point can be prevented by better production line, reducing the complexity of assembly, logistics savings, as well as shorten delivery time to achieve.

To minimize the loss

Even if the designer can select a different technology, bandwidth and dissipative losses in between, they can still have great room for compromise in design. To understand the loss mechanism of an output matching, there is a way, that is, without loss of components used to simulate the match, then each component is introduced at a loss mechanism (Table 2).

Capacitor's quality factor and its capacitance is inversely proportional. To match the dissipation of the output losses to the minimum, then the output matching, Cl values must be as small as possible. Bandwidth and compromise is made between the dissipative losses.

For a power amplifier efficiency, the dissipation loss is critical. Dissipative loss of the value of the matching network is equivalent to the reciprocal of the work of power gain, and any characteristics of the source impedance does not matter. When the load impedance of 50Ω, the dissipation loss of the formula very simple and easy application design.

There are other ways to measure the output matching loss, but they sometimes get the wrong measurement results. In the output matching circuits, different capacitor technologies will bring a different loss. Integration capacitor is suitable for use in low-loss output matching. Even if selected capacitor technology, bandwidth and dissipation still exists between the loss of a lot of space in the design tradeoffs.

Table 1 50Ω series resistance of the dissipative losses

The results -3.5 dB

Maximum gain of 0.0 dB

Gp-3.0 dB

Table 2 Mechanical loss output matching

Components are lost when the dissipative losses in the 1GHz

L10.17 dB

C10.66 dB

L20.15 dB

C20.11 dB

Cout0.03 dB

Total 1.11 dB

The key power amplifier design: the performance of the output matching circuits

Figure 1 shows the calculation of the dissipation seek matching network structure of network losses (a). The matching network and load considered together, the power output of a certain value to this complex network of power (b),. When the power output Pdel to the matching network and load the composite network, PL is transferred to the load that part of the power (c).

The key power amplifier design: the performance of the output matching circuit

Figure 2 A typical matching network at 1GHz, the terms of the power amplifier is a value for the 4 + j0Ω of the load impedance. Matching network is used to simulate lossless components, so the matching network, there is no power dissipation occurs