HomeBrew RF Siganl Generator Emitter Follower

Transcription

1 V1 is 9.0 Volts dc V2 is input signal at 30 Mhz 4Vp-p Purpose and Function The drives the load and isolates the Adjustable Gain Amplifier from the load. Theory and Design This is simply a classic Emitter follower which gives a gain just under 1 but has high input impedance and low output impedance. C6 is used as a dc blocking capacitor but gives us two undesired effects High pass filter Distortion William R. Robinson Jr. p1of 10

2 Calculated Frequency HomeBrew RF Siganl Generator Cutoff Frequency o C Bias Vc = 9.0 Vb = 4.3 Ve = F db 3 1 2RC RFc C10 = 1/(2*pi*50*500kHz) C10 = 6,368 pf or larger 1 Distortion C6 charges on the positive going cycle getting its current from Q2, but on the negative going cycle it must discharge through R10. If the combination is to large then Ve cannot follow Vb properly. This reduces Vbe and puts the transistor near or into it s cutoff region. After the input signal at Vb rises sufficiently to turn on Q2, Ve again follows Vb 2 One possible fix is to reduce the blocking Capacitor, thus less current must flow out of the blocking capacitor for the same voltage change. o This has the undesired effect of raising the cutoff frequency of the high pass filter created by the blocking capacitor and the load. In this specific case we are at the cutoff frequency limit. Another possible fix is to reduce Re. This allows more current to flow out of the blocking capacitor. o This has the undesired effect of raising Icdc which uses more power and is also limited by Ic max of the transitor In this specific case Rc had to be lowered to 50 ohms to stop distortion. The dc current is (Vcc/2)/50 = 90 ma which burns up a standard 9v battery in about 5 hours 5 forcing me to use an external supply voltage. Yet another fix is to raise Vbdc, this again raises the current through Re and allows the blocking capacitor to discharge faster. o This has the undesired effect of reducing the available swing, and raises the Dc power consumption. In this specific case is that Vbdc is determined by the previous stage (unless we add biasing resistors and another blocking capacitor.) It is obvious that some compromise must be made. I choose Re = 50 ohms and took the hit on battery life. o I do not know how to calculate the value which will work and choose 50 ohms by experimentation William R. Robinson Jr. p2of 10

3 o Ic max = 9V/50 ohms = 180 ma Icmax from data sheet is 600 ma 3 Simulation Rsim1 and Rsim2 are required to put Vb at the dc bias point, normally this is the output of the Adjustable Gain Amplifier Csim is also required to isolate V2 from the circuit; the large Value of 10 uf only affects frequencies below 100 hz. Rload has been selected at its lowest allowable value of 50 ohms Frequency VDB(vbuf1) VDB(vload) Emitter_Follower-Small Signal AC-2-Graph k 100k 1000k 1.000M 100M 1000M 1.000G 100 Frequency Note High pass cutoff from Rsim1, Rsim2 and Csim into the circuit is below the area of interest at 10 Hz Note Fc lower caused by C6 and Rload is 780 Khz o This is larger than desired but picked as a compromise with distortion Bias Vc = 9.0 Vb = 4.3 Ve = 3.5 Distortion The output will distort if R10 is to large as shown below o For R10 = 50 to 200 ohms Bottom is 50 ohms William R. Robinson Jr. p3of 10

4 Emitter_Follower-Transient-4-Sweep-Graph v(vload)[0] v(vload)[1] v(vload)[2] v(vload)[3] v(vload)[4] m -5000m n 1100n 1200n 1300n 1400n 1500n Time 1600n 1700n 1800n 1900n 200 o Same plot as above at Vc Emitter_Follower-Transient-5-Sweep-Graph v(vbuf1)[0] v(vbuf1)[1] v(vbuf1)[2] v(vbuf1)[3] v(vbuf1)[4] n 1100n 1200n 1300n 1400n 1500n Time 1600n 1700n 1800n 1900n 200 I am unable to explain the ramp up of the output when Vbe is reversed biased. o Does not occur if disconnect and Vb is held at ground (at 90 nsec below) (File is Emitter_Follower_stop.cpr) v(vbuf1) v(vload) v(vbase) Emitter_Follower_stop q-transient-0-graph m -5000m n 400n 600n 800n 1000n Time 1200n 1400n 1600n 1800n 200 o Does not occur if disconnect Vout form the cutoff Q2 (at 90 nsec below) William R. Robinson Jr. p4of 10

5 william-transient-3-graph v(vbuf1) v(vload) v(vbase) m -5000m n 400n 600n 800n 1000n Time 1200n 1400n 1600n 1800n 200 o I was unable to cause distortion with larger caps in simulation William R. Robinson Jr. p5of 10

6 Real Circuit Below is the input and output of the at 10 Mhz o Channel 1 is the from Q1c Vc = 0.68 Vp-p Vdc at 1 o Channel 2 is the from Rload Ve = 0.64 Vp-p + Vdc 1 Mhz o Note scope probe detuned the oscillator frequency from the 10 Mhz which was measured at the RF Generator output before the scope probe was attached. Frequency Fc lower = 40 Khz with available C6 = 10,000pF 1 1 o Cutoff Frequency F 3db 2RC 1/(2 *pi * 50 * 10,000pF) = 318 K o This is significantly less than the calculated value Bias Vc = 9.0 Vb = 4.6 Ve = 4.0 Distortion The screen shots below demostate the distortion with three different Re values. The screen shots show the input and output of the at 10 Mhz o Channel 1 is the from Q1b/Q2b o Channel 2 is the from Q2e Re = 50 Ohms o Note Ve follows Vb William R. Robinson Jr. p6of 10

7 Re = 220 ohms o Note Ve not following Vb completely especially at the bottom of the wave forms Re = 2K ohms o Note Ramping distortion similar to that seen in simulations William R. Robinson Jr. p7of 10

8 William R. Robinson Jr. p8of 10

9 Comparision Comparison is reasonable o Fc lower is lower than calculated o Distortion starts at lower values of Re in Simulation than in the real circuit Real-Measured Simulation Calculated Fc lower in Khz (@ c6 10,000pf) Biases Volts Vc Vb Ve William R. Robinson Jr. p9of 10

10 References 1. Horowitz, Paul and Hill, Winfield, The Art Of Electronics Second Edition, (Cambridge University Press 1989) Section 1.19 RC Filters, p Hayward, Wes, W7ZOI, et al, Experimental Methods in RF Design First Edition, Section 2.3 Large Signal Amplifiers ( Coupling Capacitor) p UNKNOWN, 2N2219A 2N2222A High Speed Switches (STMicroelectronics 2003), ( accessed William R. Robinson Jr. p10of 10