To make an amplifier from a BJT

Determine VCE and Ic

When you start to learn how to use ADS, the first project might be designing an amplifier. In this blog, I will show you how to design an amplifier using BFP650, step by step. Assuming that it will work at 2GHz.

First things first, let’s take a look at BFP650’s datasheet. Figure 1 shows the variation of transition frequency with collector current. It shows that the transition frequency reaches maximum at Ic=80mA and Vce at 3V-4V.

Figure 1 Transition Frequency

Then let’s start simulation!

Install Design Kit

Figure 2 Manage Design Kits
Figure 3 Ensure the design kit is checked

Now we can add Infineon RF transistor in schematics now.

DC Simulation

DC simulation matters because it helps us determine the operating point. All other simulations are based on DC simulation. Make sure you rename the instance name of IC_probe to IC. Then add a display template of “BJT_curve_tracer” into the schematic, which is presented in figure 4.

Figure 4 DC simulation schematic

After running the DC simulation, we will see the data display page. Figure 5 shows how IC varies with VCE in different values of base current(IBB).

Figure 5 IC varies with VCE in different values of IBB

To amplify, BJT must work in saturation region. It seems m1 is a good operating point, with VCE=3.3V and IC=81mA. So it consumes 268mW, while from datasheet, we know that it can stand 500mW at room temperature(25℃). It may seem this power consumption is a little too much, especially when VCE drops to 2V, the transition frequency does not drop much.(See figure 1) But for now, let’s set it aside.

Design Bias Network

Nobody is willing to use a current source to drive the base of such a little BJT, for which we will design a bias network, making the BJT work at the desired operating point. We will use two lumped resistor to design the bias network. And figure 6 is the schematic how we can get the value of them.

Figure 6 Design bias network

Assuming Vcc=5V, we can get two equations.R_b=\left( 3.3-VBE \right) /IBB and R_c=\left( 5-3.3 \right) /IC.i, then add two other equations calculating their power dissipation to determine the footprints of the two resistors. After defining these equations, we add a list including IC.i, Rb, Rb_Dissipation, Rc and Rc_Dissipation as presented in figure 7.

Figure 7

We want Ic to be about 80mA, so IBB=500uA, Rb=4.9kΩ, Rc=21Ω. Now we create a new schematic and see whether DC voltage of each node is correct. The schematic is shown in figure 8. And from annotations its operating point is as expected. Besides, by sweeping parameter temp, we can see how temperature affects VCE and VBE.

Figure 8 Bias network schematic

AC Simulation

As operating point is determined, it’s time to see how much the gain at 2GHz is. With the schematic in figure 9, we can do a basic AC simulation to see it.

Figure 9 AC simulation schematic

Then we set AC small-signal simulation controller ,open noise tab, enable noise calculation, add node “Vout”, and sort noise contributors by value. At last run the simulations. In the data display page, we add two equations, gain=mag\left( Vout \right) /mag\left( Vin \right) and phase_diff=phase\left( Vout \right) -phase\left( Vin \right). We can see at 2GHz, the gain is around 16dB, and phase_diff is around 70 degrees. In the same schematic ,we can see how gain varies with Vcc. We add a parameter sweep controller, and set sweep parameter “Vcc”, SimInstanceName[1] “AC1” and sweep range. Here I set it from 4.7V to 5.3V, and step of 0.05V. The result is presented in figure 10. We can see that the difference is less than 0.1dB.

Figure 10

S-Parameter Simulation

To optimize performance of the amplifier, we must do impedance matching, and S-Parameter Simulation can help. The schematic is shown in figure 11.

Figure 11 S-Parameter schematic, first time

From S11, we know that Zin1=13.997+j10.482 \varOmega. Instead of SmartComponent, we use Smith chart tool as our assistant. Here we use the most simple match method–LC match. For we don’t want to change the operating point of the BJT, we should use a parallel L and a series C to match. With the help of smith chart, we can get their values–2.4pF and 2.5nH. Then we replace two DC_Feed components with two 120nH inductors(At 2GHz, they can be seen as open-circuit). After that, we run another S-Parameter simulation a second time. The schematic is presented in figure 12.

Figure 12 S-Parameter schematic, second time

In data display page, we can see that S11 at 2GHz is very small(around -32.7dB) but S22 is -13dB(It is ok, but we can do it better). From S22 Smith chart, we know Zin2=46.388-j21.748 \varOmega. As above, we use Smith chart to help matching. Considering their values, I use a 4.4pF series C and a 5.5nH parallel L as well as a 100pF DC block capacitor. Then we simulate a third time.

Figure 12 S-Parameter schematic, third time

Then we found S11 increase to -13.5dB, which means we should optimize the match network at port 2 to decrease S11. Then we enable the tuning options of C7 and L7 ,then tune their values. Generally, we want both S11 and S22 less than 15dB. It’s easy to achieve in this example.

Figure 13 S-Parameters after tuning

Harmonic Balance Simulation

Harmonic balance simulation is a frequency-domain analysis method to simulate non-linear systems. It can help us evaluate non-linearity of a system. For an amplifier, harmonic balance simulation can give general AC characteristics such as IPO3 and P1dB. In this example, we will see how harmonic balance simulation work. Figure 14 is the schematic of the one tone harmonic balance simulation and figure 15 shows the non-linearity.

Figure 14 Harmonic balance simulation, one tone
Figure 15

In practice, input signal is not single tone, so we need to use multiple tones harmonic balance simulation. Assuming the bandwidth is 10MHz, let’s see the non-linearity characteristics. We add a P_nTone source, and set the schematic as the figure 16 below. By adding two IP3out calculation part, we can get IP3 at the data display page. The output IP3 is about 36dBm.

Figure 16 Harmonic balance simulation, two tone

Circuit Envelop Simulation

Circuit Envelop Simulation simulates high-frequency amplifiers, mixers, oscillators, and subsystems that involve transient or modulated RF signals. It is highly efficient in analyzing circuits with digitally modulated signals, because transient simulation takes place only around the carrier and its harmonics. In addition, its calculation are not made where the spectrum is empty(assuming most of the frequency spectrum is empty).

To run circuit envelop simulation for this amplifier, we add a GSM source and a circuit envelop simulation controller, then configure them as below.

Figure 17 Circuit envelop simulation schematic

Run and open data display page, then add a equationbaseband=diff\left( unwrap\left( phase\left( Vout\left[ 1 \right] \right) \right) /360 \right), in which baseband is the demodulated baseband signal at the output port. Figure 15 presents the signal at bits_out and baseband. From it, we can see the amplifier can amplify the GSM signal without much distortion.

Figure 18

Gain Compression Simulation

Gain compression simulation is a large signal simulation. We can get P1dB by running this kind of simulation. The P1dB of this amplifier is 20dBm. Besides, harmonic balance simulation can also help us get P1dB, by sweeping the power of input signal and calculate the gain of fundamental wave, just as figure 19 presents.

Figure 19

One response to “To make an amplifier from a BJT”

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.