ESE319 Introduction to Microelectronics Colpitts Oscillator ● ● ● ● ● ● ● Basic positive feedback oscillator The Colpitts LC Oscillator circuit Open-loop analysis Closed-loop analysis Root locus Stability limit Colpitts design 2008 Kenneth R. Laker updated 10Nov09 KRL 1 ESE319 Introduction to Microelectronics Basic Positive Feedback Oscillator closed-loop oscillator Vi = 0 + Vf A(s) + pos. fdbk open-loop: determine loop-gain Vo Acl s= A s 1− A s V o= A sV i V f = As0V o ⇒ V o 1− As=0 Vi + A(s) Vo Vf V f Vo K N s = = As= Vi Vi D s Since: V o ≠ 0 ⇒1−A s=0⇒ A s=1 ⇒ D s−K N s=0 Condition for oscillation at s = jω0: As=1 e j ± 2k Barkhausen criterion for k = 0, 1, 2 .. 2008 Kenneth R. Laker updated 10Nov09 KRL 2 ESE319 Introduction to Microelectronics Colpitts Oscillator Basic Schematic Assumptions: 1. rπ large (compared to 1/ωC2). L R C1 2. Cµ negligible (compared to C1, C2) C2 3. Cπ part of C2 (in closed loop) 4. R represents total resistance in collector circuit, i.e. R∥r o≈ R Emphasizing feedback L Vπ R C1 Vf C2 Emphasizing loop gain A s 2008 Kenneth R. Laker updated 10Nov09 KRL C B V r C C gmV ro E E 3 ESE319 Introduction to Microelectronics Loop-Gain Analysis Vπ Cµ L Vc gmVπ R C1 Vf Vf C2 Vf loop-gain: A s= V Node equation at Vc: ∣ j C ∣≪ g m 1 1 1 s C 1 V c − V f g m −s C V =0 sL R sL 1 1 1 s C 1 V c − V f g m V =0 sL R sL −1 1 at Vf: V c sC 2 V f =0 sL sL 2008 Kenneth R. Laker updated 10Nov09 KRL (1) Note that Vf = Vf (s) Vc = Vc(s) Vπ = Vπ(s) (2) 4 ESE319 Introduction to Microelectronics Open Loop Analysis - cont. from previous slide Rearranging (1) and (2): sC 1 − (1) 1 1 1 V c − V f =−g m V sL R sL 1 1 V c s C 2 V f =0 sL sL (3) (2) Eq. (1) – gm Vπ (4) Eq. (2) Further rearrangement of (3) and (4): s 2 LC 1 1 1 1 V c − V f =−g m V sL R sL s 2 LC 2 1 1 − V c V f =0 sL sL 2008 Kenneth R. Laker updated 10Nov09 KRL (5) (6) 5 ESE319 Introduction to Microelectronics Open Loop Analysis - cont. from previous slide Prepare to add the two equations: Adding (to eliminate Vc terms): 2 2 s L C 11 1 1 − sL sL R 2008 Kenneth R. Laker updated 10Nov09 KRL 2 s L C 21 1 − V c V f =0 sL sL 2 2 −g m 1 s L C 11 1 1 V c− V f= V sL sL R sL sL 2 s 2 LC 1 1 1 1 s L C 1 1 1 − V c sL sL R sL R 2 s L C 11 1 1 V c − V f =−g m V (5) sL R sL s 2 LC 2 1 V f =0 sL s2 L C 2 1 −g m V f= V sL sL Eq 5∗ (6) 1 sL 2 s LC 1 1 1 Eq 6∗ sL R (7) 6 ESE319 Introduction to Microelectronics Open Loop Analysis – cont. From previous slide 2 s2 L C 11 1 1 − sL sL R s2 L C 2 1 −g m V f= V sL sL (7) Multiply (7) by (sL)2: sL 2 −1 s LC 1 1 s LC 21 V f =−g m sL V R 2 Expand and collect terms according to sn: 2 L C2 L 4 2 2 3 −1s C 1 C 2 L s LC 1 LC 2 s s 1 V f =−g m sLV R R 2008 Kenneth R. Laker updated 10Nov09 KRL (8) 7 ESE319 Introduction to Microelectronics Open Loop Analysis - cont. From previous slide 3 2 s L C2 s L 4 2 2 −1s C 1 C 2 L s LC 1 LC 2 1 V f =−g m sLV R R (8) Canceling (-1 by 1) in (8) and dividing by sL: 2 s L C2 1 3 s C 1 C 2 Ls C 1C 2 V f =−g m V R R Multiply by R: s 3 R C 1 C 2 Ls R C 1 C 2 s LC 21 V f =−g m R V 2008 Kenneth R. Laker updated 10Nov09 KRL 2 (9) 8 ESE319 Introduction to Microelectronics Open Loop Analysis - cont. From previous slide s3 R C 1 C 2 Ls R C 1C 2 s 2 LC 21 V f =−g m R V (9) Divide (9) by RC1C2L: C 1C 2 −g m R 1 1 s s s V f= V C1C 2 L R C1 R C1C 2 L RC1 C2 L 3 2 The loop-gain transfer function: −g m Vf C1 C 2 L K N s = As= = V D s C 1C 2 1 3 2 1 s s s R C1 C1 C 2 L R C1C 2 L 2008 Kenneth R. Laker updated 10Nov09 KRL (10) 9 ESE319 Introduction to Microelectronics Closed Loop Analysis - cont. The closed loop equation: (10) A s K N s K N s Acl s= = = ' 1− A s D s−K N s D s where: K N s As= D s −g m K= C1 C 2 L N s=1 C 1 C 2 1 1 D s=s s s RC1 C1 C2 L RC1 C2 L 3 then: 2 (11) KN(s) −g m C 1 C 2 1g m R 3 2 1 D s=D s− =s s s C1 C 2 L RC1 C1 C2 L R C1 C2 L ' 2008 Kenneth R. Laker updated 10Nov09 KRL (12) 10 ESE319 Introduction to Microelectronics Closed Loop Analysis - cont. We know that the open loop system A(s) is stable. It has poles in the left half plane, since it is a passive RLC circuit. We also know that it has 3 poles. One is negative-real, the other 2 can be negative-real or LHP complex conjugates. C 1C 2 1 1 D s=s s s R C1 C1C2 L R C1C 2 L 3 2 (11) So, let's do a rough sketch of the root locus for a feedback system with a 3 pole A(s). 2008 Kenneth R. Laker updated 10Nov09 KRL 11 ESE319 Introduction to Microelectronics Root Locus Characteristic Open loop poles somewhere in here F = Fx j j x X ● Closed loop pole locus ● F = Fx X − j x ● The loop will become unstable for any value of F Fx . Rather than sketch the root locus in more exacting detail – it has served its purpose by verifying that oscillation is possible. Let's solve for the required Fx . C 1C 2 1 1 D s =s s s R C1 C 1 C 2 L RC 1 C 2 L (11) 3 2 C 1C 2 1 g m R (12) 1 D s= D s − K N s =s s s RC 1 C 1 C 2 L RC 1 C 2 L ' 2008 Kenneth R. Laker updated 10Nov09 KRL 3 2 12 ESE319 Introduction to Microelectronics Stability Limit Calculation If the closed loop system is at the stability limit point: D' s=DsK N s=sa s2 2x to oscillate at ω = ωx Multiplying terms: =0 =0 ' 3 2 2 2 ' 2 2 2 3 D s=s a s x sa x ⇒ D j =a x −a j x − Match term by term with: 2 x a 2 ax C 1C 2 1g m R 1 D s=s s s RC 1 C1 C 2 L R C1C 2 L ' 1 a= R C1 3 2 x = 2 C 1 C 2 C1 C2 L 2008 Kenneth R. Laker updated 10Nov09 KRL 1g m R 2 2 a = s.t. a∗ x =a x R C1C 2 L 2 x 13 ESE319 Introduction to Microelectronics Stability Limit C 1C 2 = C1C 2 L 1 a= R C1 2 x 1g m R a = R C1C 2 L 2 x a∗ 2x =a 2x 2 2 To find the “gain” requirement for oscillation, equate a∗ x =a x 1g m R 1 C 1C 2 C 1 C 2 = = 2 R C1 C1C 2 L R C1 C 2 L R C1 C2 L a 2x a 2x C 1 C 2 1g m R C 1 C 2 C2 = ⇒1g m R= =1 2 C1 C1 R C1 C2 L R C1C 2 L C2 g m R= C1 2008 Kenneth R. Laker updated 10Nov09 KRL C 1 C 2 x= C1 C2 L 14 ESE319 Introduction to Microelectronics Oscillator Design Summary The oscillation frequency: The required feedback gain ΓF = gmR: To insure the start-up of oscillation: Comments: 1. Unfortunately we don't control “R”. C 1 C 2 x= C1 C2 L C2 poles at ± jωx Fx = g m x R= C1 C2 g m x R C1 2. Set C2 & C1 to convenient values, say C2 = C1 = C, and adjust gm using IC s.t. gmR > C2/C1; then adjust “L” to maintain ωx at the desired oscillation frequency. 2008 Kenneth R. Laker updated 10Nov09 KRL 15 ESE319 Introduction to Microelectronics Practical Colpitts Oscillator Circuit VCC RFC = RF choke -> RB1 RFC C1 2008 Kenneth R. Laker updated 10Nov09 KRL RE low resistance at dc L large RB2 large reactance at ωx C2 large 16