SchematicsSummaryAppendix: DC FeedbackECE 304: Feedback Amplifier Comparison Schematics Q_NQ110.18mA0+OUT{R_L}OUT2+100k0VOUT10Q_PQ220.27mA15.00VVPVPVP+{R_T}713.9mV!_DC{I_B}89.62uA0VP14.27V.model Q_P PNP (Is=10fA Bf=100 Vaf=10V)10mA+10_F+10_F0711.8mV+-15V{I_E1}{I_E2}0+{R_S}Transient AnalysisI_SIN{I_S}1kHzACSweepI_AC1AQ_NQ310.10mAPARAMETERS:R_S = 100kR_L = 100kR_T = 10kR_B = 1kI_B = 89.618591uAI_E2 = 20mAI_E1 = 10mAI_S = 0.4mA+10_F0OUT4.818V.model Q_N NPN (Is=10fA Bf=100 Vaf=10V) FIGURE 1 A transresistance amplifier Transient AnalysisI_SIN{I_S}1kHzOUTQ_NQ310.10mA0+-15V+10_F+10_F+100k014.27V10mA{I_E1}711.8mVPARAMETERS:R_S = 100kR_L = 1kR_T = 10kR_B = 1kI_B = 89.618591uAI_E2 = 20mAI_E1 = 10mAI_S = 200uA.model Q_N NPN (Is=10fA Bf=100 Vaf=10V)0{I_E2}VP+10_F+{R_B}.model Q_P PNP (Is=10fA Bf=100 Vaf=10V)I_DC{I_B}89.62uAQ_PQ220.27mA015.00V0OUT20ACSweepI_AC1A+OUT{R_L}VP+{R_T}VP+{R_S}OUT14.818V713.9mVVPQ_NQ110.18mA FIGURE 2 A current amplifier Figure 1 and Figure 2 show the same amplifier with different feedback arrangements. In Figure 1 the feedback is voltage and the intention is to make the output voltage the same regardless of RL (same output voltage, regardless of output current). In Figure 2 the feedback is current and the intention is to make the output current the same regardless of RL (same output current, regardless of output voltage). J R Brews Page 1 3/31/2006The success of the feedback can be judged by looking at the output behavior. Time0s 0.4ms 0.8ms 1.2ms 1.6ms 2.0msV(OUT)-5.0V0V5.0V FIGURE 3 Output voltage in transR amplifier for RL = 430 Ω, 1 kΩ and 10 kΩ; no difference in voltage despite a large difference in current, as shown below in Figure 4 Time0s 0.4ms 0.8ms 1.2ms 1.6ms 2.0msI(OUT)-10mA0A10mARL=10kRL=1kRL=430 FIGURE 4 Current variation in load resistor of transR amplifier for three values of load resistance RL Time0s 0.4ms 0.8ms 1.2ms 1.6ms 2.0msIC(Q3)0A10mA20mARL=10kRL=1kRL=430 FIGURE 5 Current in output transistor of transR amplifier showing for small RL cutoff of Q3 is immanent Figure 3 - Figure 5 show the behavior of the transR amplifier. The feedback succeeds in keeping the voltage output the same regardless of the current output, until the current drawn by the load is so large that the output transistor cuts off. J R Brews Page 2 3/31/2006Time0s 0.4ms 0.8ms 1.2ms 1.6ms 2.0msI(OUT)-4.0mA0A4.0mA FIGURE 6 Output current in load RL for the current amplifier of Figure 2 for RL = 100 Ω, 500 Ω and 1.2 kΩ; feedback holds the current constant even though the output voltage varies a lot with RL as shown below in Figure 7 Time0s 0.4ms 0.8ms 1.2ms 1.6ms 2.0msV(OUT)-5.0V0V5.0VRL=1.2kRL=500RL=100 FIGURE 7 Voltage variation across load resistor RL for several resistor values Time0s 0.4ms 0.8ms 1.2ms 1.6ms 2.0msV(Q3:c)- V(Q3:b) 00510RL=1.2kRL=500RL=100 FIGURE 8 VCB for output transistor Q3 showing that at large RL saturation of Q3 is immanent Figure 6 - Figure 8 show the behavior of the current amplifier. Feedback maintains the current in RL the same regardless of the output voltage. Summary The introduction of feedback causes a big difference in how the feedback amplifier behaves. Starting with the same initial three-stage amplifier, a Shunt input/Shunt output feedback arrangement shown in Figure 1 results in a transR amplifier that can be viewed as an almost ideal voltage source driving the load. In contrast, a Shunt input/Series output feedback as shown in Figure 2 results in a current amplifier that can be viewed as an almost ideal current source driving the load RL. Because the TransR amplifier doesn’t control current, it is limited at small RL by cutoff of the output transistor Q3. Because the current amplifier doesn’t control voltage, it is limited at large RL by saturation of the output transistor Q3. J R Brews Page 3 3/31/2006Appendix: DC Feedback A shortcoming of the circuits in Figure 1 and Figure 2 is that the DC bias point is set using a base current I_B to Q1. Because the gain of the amplifier is very high, the selection of I_B is very delicate: a small change will cause the amplifier output to pin at the upper or lower supply voltage. Instead, we would like to have a feedback loop set the Q-point. However, we don’t want this feedback to upset the AC feedback loop we already have. So, we set up a DC feedback loop as shown in Figure 9. 0{I_E1}PARAMETERS:R_S = 100kR_L = 1kR_T = 10kR_B = 1kI_E2 = 20mAI_E1 = 10mAI_S = 200uAR_FB = 3kR_E = 3k+10_F+{R_FB}89.62uAQ_NQ49.808uA1.080mAVP1mA+{R_E}1.090mAVP{I_E2}Q_PQ220.27mAOUT2VP0.model Q_N NPN (Is=10fA Bf=100 Vaf=10V)+-15V+10_F00+10_FQ_NQ110.18mA+100k101.0uA.model Q_P PNP (Is=10fA Bf=100 Vaf=10V)VP+{R_T}Transient AnalysisI_SIN{I_S}1kHz4.904V980.7mV10mA+{R_S}+10_F0VPOUTV+OUT{R_L}Q_NQ310.09mA0OUT1711.8mV4.250V014.27V15.00V0ACSweepI_AC1A AC Bypass Cap FIGURE 9 The transR amplifier with a DC feedback loop made up of Q4, R_E and R_FB to set the Q-point; a bypass capacitor shorts any AC feedback through this loop to ground With the DC feedback loop, the voltage at the collector of Q3 is easily set by choosing the value of resistor RE because RE sets the emitter voltage of Q4, and VC(Q3) = VE(Q4) + VBE(Q4).1 This little DC feedback trick is used in the commercial amplifier MC 1553. AC operation of the amplifier is unchanged from Figure 1. The same trick works for the current amplifier of Figure 2. 1 The voltage of the collector of Q3 is significant because it sets the maximum output swing of the amplifier. J R Brews Page 4