What XTAL OSC is this?
in [Design]
|
From: Andrew Holme on 1 May 2010 09:46 What is this oscillator configuration called? It seems to have some good and useful features: easy to control XTAL drive level; operates in series resonant mode; differential square-ish wave output with fast rise time; very good amplitude stability versus supply voltage variation; no relaxation mode if crystal removed; no need to tune-out C0 holder capacitance; simple single-point power-supply decoupling. What are the disadvantages? Two active devices = more noise? But the Butler and Driscoll circuits use two active devices. No LC frequency selectivity to force XTAL onto correct harmonic? But that could be added... TIA LTSpice .ASC file: Version 4 SHEET 1 928 744 WIRE -80 -160 -128 -160 WIRE 48 -160 0 -160 WIRE 96 -160 48 -160 WIRE 224 -160 176 -160 WIRE 48 -112 48 -160 WIRE 48 0 48 -32 WIRE -128 48 -128 -160 WIRE -16 48 -128 48 WIRE 224 48 224 -160 WIRE 224 48 112 48 WIRE -128 128 -128 48 WIRE 224 128 224 48 WIRE -16 176 112 48 WIRE -16 176 -64 176 WIRE 112 176 -16 48 WIRE 160 176 112 176 WIRE -128 272 -128 224 WIRE 32 272 -128 272 WIRE 224 272 224 224 WIRE 224 272 96 272 WIRE -128 384 -128 272 WIRE -80 384 -128 384 WIRE 32 384 0 384 WIRE 112 384 96 384 WIRE 224 384 224 272 WIRE 224 384 192 384 WIRE 224 432 224 384 WIRE 352 432 224 432 WIRE -128 480 -128 384 WIRE 224 480 224 432 WIRE 352 480 352 432 WIRE -128 608 -128 560 WIRE 224 608 224 560 WIRE 352 608 352 560 FLAG 48 0 0 FLAG 224 608 0 FLAG 352 608 0 FLAG -128 608 0 SYMBOL voltage 48 -128 R0 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName V2 SYMATTR Value 5 SYMBOL npn 160 128 R0 SYMATTR InstName Q1 SYMATTR Value 2N3904 SYMBOL npn -64 128 M0 SYMATTR InstName Q2 SYMATTR Value 2N3904 SYMBOL current 352 480 R0 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName I2 SYMATTR Value SINE(0 1e-3 10e6 0 0 0 10) SYMBOL ind -96 400 R270 WINDOW 0 32 56 VTop 0 WINDOW 3 5 56 VBottom 0 SYMATTR InstName L1 SYMATTR Value 0.0253303 SYMATTR SpiceLine Rpar=1e300 SYMBOL cap 96 368 R90 WINDOW 0 0 32 VBottom 0 WINDOW 3 32 32 VTop 0 SYMATTR InstName C1 SYMATTR Value 10f SYMBOL res 208 368 R90 WINDOW 0 0 56 VBottom 0 WINDOW 3 32 56 VTop 0 SYMATTR InstName R1 SYMATTR Value 20 SYMBOL res 192 -176 R90 WINDOW 0 0 56 VBottom 0 WINDOW 3 32 56 VTop 0 SYMATTR InstName R2 SYMATTR Value 1k SYMBOL res 16 -176 R90 WINDOW 0 0 56 VBottom 0 WINDOW 3 32 56 VTop 0 SYMATTR InstName R3 SYMATTR Value 1k SYMBOL cap 96 256 R90 WINDOW 0 0 32 VBottom 0 WINDOW 3 32 32 VTop 0 SYMATTR InstName C0 SYMATTR Value 5p SYMBOL res 208 464 R0 SYMATTR InstName R4 SYMATTR Value 3k3 SYMBOL res -144 464 R0 SYMATTR InstName R5 SYMATTR Value 3k3 TEXT 16 600 Left 0 !.tran 1m
From: Andrew Holme on 1 May 2010 10:59 "Andrew Holme" <ah(a)nospam.com> wrote in message news:_gWCn.13199$fO7.7783(a)newsfe22.ams2... > What is this oscillator configuration called? > > It seems to have some good and useful features: easy to control XTAL drive > level; operates in series resonant mode; differential square-ish wave > output with fast rise time; very good amplitude stability versus supply > voltage variation; no relaxation mode if crystal removed; no need to > tune-out C0 holder capacitance; simple single-point power-supply > decoupling. > > What are the disadvantages? Two active devices = more noise? But the > Butler and Driscoll circuits use two active devices. No LC frequency > selectivity to force XTAL onto correct harmonic? But that could be > added... > > TIA > > LTSpice .ASC file: > > Version 4 > SHEET 1 928 744 > WIRE -80 -160 -128 -160 > WIRE 48 -160 0 -160 > WIRE 96 -160 48 -160 > WIRE 224 -160 176 -160 > WIRE 48 -112 48 -160 > WIRE 48 0 48 -32 > WIRE -128 48 -128 -160 > WIRE -16 48 -128 48 > WIRE 224 48 224 -160 > WIRE 224 48 112 48 > WIRE -128 128 -128 48 > WIRE 224 128 224 48 > WIRE -16 176 112 48 > WIRE -16 176 -64 176 > WIRE 112 176 -16 48 > WIRE 160 176 112 176 > WIRE -128 272 -128 224 > WIRE 32 272 -128 272 > WIRE 224 272 224 224 > WIRE 224 272 96 272 > WIRE -128 384 -128 272 > WIRE -80 384 -128 384 > WIRE 32 384 0 384 > WIRE 112 384 96 384 > WIRE 224 384 224 272 > WIRE 224 384 192 384 > WIRE 224 432 224 384 > WIRE 352 432 224 432 > WIRE -128 480 -128 384 > WIRE 224 480 224 432 > WIRE 352 480 352 432 > WIRE -128 608 -128 560 > WIRE 224 608 224 560 > WIRE 352 608 352 560 > FLAG 48 0 0 > FLAG 224 608 0 > FLAG 352 608 0 > FLAG -128 608 0 > SYMBOL voltage 48 -128 R0 > WINDOW 123 0 0 Left 0 > WINDOW 39 0 0 Left 0 > SYMATTR InstName V2 > SYMATTR Value 5 > SYMBOL npn 160 128 R0 > SYMATTR InstName Q1 > SYMATTR Value 2N3904 > SYMBOL npn -64 128 M0 > SYMATTR InstName Q2 > SYMATTR Value 2N3904 > SYMBOL current 352 480 R0 > WINDOW 123 0 0 Left 0 > WINDOW 39 0 0 Left 0 > SYMATTR InstName I2 > SYMATTR Value SINE(0 1e-3 10e6 0 0 0 10) > SYMBOL ind -96 400 R270 > WINDOW 0 32 56 VTop 0 > WINDOW 3 5 56 VBottom 0 > SYMATTR InstName L1 > SYMATTR Value 0.0253303 > SYMATTR SpiceLine Rpar=1e300 > SYMBOL cap 96 368 R90 > WINDOW 0 0 32 VBottom 0 > WINDOW 3 32 32 VTop 0 > SYMATTR InstName C1 > SYMATTR Value 10f > SYMBOL res 208 368 R90 > WINDOW 0 0 56 VBottom 0 > WINDOW 3 32 56 VTop 0 > SYMATTR InstName R1 > SYMATTR Value 20 > SYMBOL res 192 -176 R90 > WINDOW 0 0 56 VBottom 0 > WINDOW 3 32 56 VTop 0 > SYMATTR InstName R2 > SYMATTR Value 1k > SYMBOL res 16 -176 R90 > WINDOW 0 0 56 VBottom 0 > WINDOW 3 32 56 VTop 0 > SYMATTR InstName R3 > SYMATTR Value 1k > SYMBOL cap 96 256 R90 > WINDOW 0 0 32 VBottom 0 > WINDOW 3 32 32 VTop 0 > SYMATTR InstName C0 > SYMATTR Value 5p > SYMBOL res 208 464 R0 > SYMATTR InstName R4 > SYMATTR Value 3k3 > SYMBOL res -144 464 R0 > SYMATTR InstName R5 > SYMATTR Value 3k3 > TEXT 16 600 Left 0 !.tran 1m > > > I think there is a trade-off between working-Q and drive level. The crystal is in series with 2*re. The lower re, the higher Qw. Unfortunately, reducing re means increasing Ie which increases drive. Nevertheless, quite a reasonable compromise may be possible.
From: Jim Thompson on 1 May 2010 11:56 On Sat, 1 May 2010 14:46:37 +0100, "Andrew Holme" <ah(a)nospam.com> wrote: >What is this oscillator configuration called? Not a very good crystal oscillator... it's not linear. If you want quality, do it this way... http://analog-innovations.com/SED/XtalSeriesOsc.pdf > >It seems to have some good and useful features: easy to control XTAL drive >level; operates in series resonant mode; differential square-ish wave output >with fast rise time; very good amplitude stability versus supply voltage >variation; no relaxation mode if crystal removed; no need to tune-out C0 >holder capacitance; simple single-point power-supply decoupling. > >What are the disadvantages? Two active devices = more noise? But the >Butler and Driscoll circuits use two active devices. No LC frequency >selectivity to force XTAL onto correct harmonic? But that could be added... > >TIA > >LTSpice .ASC file: > >Version 4 >SHEET 1 928 744 >WIRE -80 -160 -128 -160 >WIRE 48 -160 0 -160 >WIRE 96 -160 48 -160 >WIRE 224 -160 176 -160 >WIRE 48 -112 48 -160 >WIRE 48 0 48 -32 >WIRE -128 48 -128 -160 >WIRE -16 48 -128 48 >WIRE 224 48 224 -160 >WIRE 224 48 112 48 >WIRE -128 128 -128 48 >WIRE 224 128 224 48 >WIRE -16 176 112 48 >WIRE -16 176 -64 176 >WIRE 112 176 -16 48 >WIRE 160 176 112 176 >WIRE -128 272 -128 224 >WIRE 32 272 -128 272 >WIRE 224 272 224 224 >WIRE 224 272 96 272 >WIRE -128 384 -128 272 >WIRE -80 384 -128 384 >WIRE 32 384 0 384 >WIRE 112 384 96 384 >WIRE 224 384 224 272 >WIRE 224 384 192 384 >WIRE 224 432 224 384 >WIRE 352 432 224 432 >WIRE -128 480 -128 384 >WIRE 224 480 224 432 >WIRE 352 480 352 432 >WIRE -128 608 -128 560 >WIRE 224 608 224 560 >WIRE 352 608 352 560 >FLAG 48 0 0 >FLAG 224 608 0 >FLAG 352 608 0 >FLAG -128 608 0 >SYMBOL voltage 48 -128 R0 >WINDOW 123 0 0 Left 0 >WINDOW 39 0 0 Left 0 >SYMATTR InstName V2 >SYMATTR Value 5 >SYMBOL npn 160 128 R0 >SYMATTR InstName Q1 >SYMATTR Value 2N3904 >SYMBOL npn -64 128 M0 >SYMATTR InstName Q2 >SYMATTR Value 2N3904 >SYMBOL current 352 480 R0 >WINDOW 123 0 0 Left 0 >WINDOW 39 0 0 Left 0 >SYMATTR InstName I2 >SYMATTR Value SINE(0 1e-3 10e6 0 0 0 10) >SYMBOL ind -96 400 R270 >WINDOW 0 32 56 VTop 0 >WINDOW 3 5 56 VBottom 0 >SYMATTR InstName L1 >SYMATTR Value 0.0253303 >SYMATTR SpiceLine Rpar=1e300 >SYMBOL cap 96 368 R90 >WINDOW 0 0 32 VBottom 0 >WINDOW 3 32 32 VTop 0 >SYMATTR InstName C1 >SYMATTR Value 10f >SYMBOL res 208 368 R90 >WINDOW 0 0 56 VBottom 0 >WINDOW 3 32 56 VTop 0 >SYMATTR InstName R1 >SYMATTR Value 20 >SYMBOL res 192 -176 R90 >WINDOW 0 0 56 VBottom 0 >WINDOW 3 32 56 VTop 0 >SYMATTR InstName R2 >SYMATTR Value 1k >SYMBOL res 16 -176 R90 >WINDOW 0 0 56 VBottom 0 >WINDOW 3 32 56 VTop 0 >SYMATTR InstName R3 >SYMATTR Value 1k >SYMBOL cap 96 256 R90 >WINDOW 0 0 32 VBottom 0 >WINDOW 3 32 32 VTop 0 >SYMATTR InstName C0 >SYMATTR Value 5p >SYMBOL res 208 464 R0 >SYMATTR InstName R4 >SYMATTR Value 3k3 >SYMBOL res -144 464 R0 >SYMATTR InstName R5 >SYMATTR Value 3k3 >TEXT 16 600 Left 0 !.tran 1m > > > ...Jim Thompson -- | James E.Thompson, CTO | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | The only thing bipartisan in this country is hypocrisy
From: John Larkin on 1 May 2010 13:06 On Sat, 1 May 2010 14:46:37 +0100, "Andrew Holme" <ah(a)nospam.com> wrote: >What is this oscillator configuration called? > >It seems to have some good and useful features: easy to control XTAL drive >level; operates in series resonant mode; differential square-ish wave output >with fast rise time; very good amplitude stability versus supply voltage >variation; no relaxation mode if crystal removed; no need to tune-out C0 >holder capacitance; simple single-point power-supply decoupling. > >What are the disadvantages? Two active devices = more noise? But the >Butler and Driscoll circuits use two active devices. No LC frequency >selectivity to force XTAL onto correct harmonic? But that could be added... > >TIA The crystal is working in series mode. I think it sees a lot of equivalent circuit series impedance, effectively in series with the internal 20 ohm thing, so Q will be low. The small-signal emitter impedances are around 30 ohms, so you have roughly 60 right there, but in fact only one transistor is on at a time, so it's really a lot higher. It may be oscillating a bit off the actual series resonance point. I'd expect relatively bad stability and phase noise compared to a circuit that puts the series-mode xtal into a lower impedance loop. Cute circuit. It has all the virtues you mentioned. Spice is rotten for analyzing xo's. You can't easily measure frequency to PPM resolution, and even if you could you'd need to run such absurdly small time steps to get usable accuracy, the sim would run for days. You can do good XO stuff using frequency domain analysis, but not for nonlinear circuits like this one. I mostly buy packaged oscillators these days. Simplifies life. John
From: MooseFET on 1 May 2010 13:41
On May 1, 6:46 am, "Andrew Holme" <a...(a)nospam.com> wrote: > What is this oscillator configuration called? > > It seems to have some good and useful features: easy to control XTAL drive > level; operates in series resonant mode; differential square-ish wave output > with fast rise time; very good amplitude stability versus supply voltage > variation; no relaxation mode if crystal removed; no need to tune-out C0 > holder capacitance; simple single-point power-supply decoupling. > > What are the disadvantages? Two active devices = more noise? But the > Butler and Driscoll circuits use two active devices. No LC frequency > selectivity to force XTAL onto correct harmonic? But that could be added... The basic topology is that of the emitter timed multivibrator. If you do something to reduce the gain to the point where the transistors remain nearly linear, you would have a type of Buttler oscillator. At the lower gain value, the oscillator would be one of the better designs from the point of view of low harmonic content in the output and also be quite stable. When you have too much gain, this sort is really quite crappy. The frequency of oscillation come unstuck from the series resonance point of the crystal. Normally, the collector load would have a tuned circuit to select the operating frequency. A couple of small Schottky diodes would do a limiting circuit. |