Larkin's Oscillator Explained
in [Design]
|
From: Jan Panteltje on 15 Jun 2010 07:57 On a sunny day (Tue, 15 Jun 2010 12:16:05 +0100) it happened baron <baron.nospam(a)linuxmaniac.nospam.net> wrote in <hv7ndm$ue$1(a)news.eternal-september.org>: >George Herold Inscribed thus: > >> On Jun 14, 3:29 pm, John Larkin >> <jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote: >>> On Mon, 14 Jun 2010 12:12:17 -0700 (PDT), George Herold >>> >>> <gher...(a)teachspin.com> wrote: >>> >>> >Could someone tell me what squegging is? A google search gives >>> >hints of something like motor-boating, but different? >>> >>> We're referring to an oscillator that oscillates in bursts, rather >>> than continuously. A resonator connected to a negative resistance >>> doesn't squegg, but on the other hand its amplitude builds to >>> infinite (for certain values of infinite) at a rate determined by Q. >>> >>> You can think of an LC resonator as having a 1st order response, >>> measuring oscillation envelope versus drive. Oscillators squegg when >>> some amplitude limiting mechanism adds additional time delays or >>> memory, more poles in the control loop, like the capacitor in the >>> base of my circuit. A bigger cap reduces the tendency to squegg. >>> >>> A superregen receiver is a squegging-on-purpose RF oscillator. That's >>> a fascinating circuit. >>> >>> John >> >> Hmm, OK something like motorboating gone to the extreme.. enough >> amplitude modultaion to shut the oscillator off for some time. (?) >> Does the squegg rate stay constant in a sugerregen receiver? I'm not >> real sure what a superregen reciever is, anything like a marginal >> oscillator? >> >> George H. > >Its a technique by which an amplifier is held on the verge of >oscillation by positive feedback, at which point the gain is very high. In my view the gain is not 'very high', but just enough to compensate for any losses to make the overall gain '1', as needed for oscillation. The effect it has, is increasing the Q factor, so also reducing the bandwidth, by reducing the loading on the tuned ciruit. Think negative resistance. That means you get a lot of signal on that tuned ciruit. Think about it, if you really HAD that much gain available then you would not need the super-reg method. >By causing the oscillation to be quenched at the instant it starts is >where the "Super" bit comes from in "Superregenerative". Commonly used >in radio receivers where a single active device is used to both amplify >the incoming RF, demodulate it and amplify the resultant AF, often >driving headphones directly. Often frowned upon because of the >interference that can be caused by radiated RF. Yep. >-- >Best Regards: > Baron. >
From: Jim Thompson on 15 Jun 2010 11:29 On Sun, 13 Jun 2010 16:19:23 -0700, Jim Thompson <To-Email-Use-The-Envelope-Icon(a)On-My-Web-Site.com> wrote: >Larkin's Oscillator Explained > >ASCII drawing courtesy of dagmargoodboat(a)yahoo.com .... > > > Vcc = +5v > --+-------+------+-- > | | | > | | |_ || > | | _)|| > .-. --- L1a _)|| >Rb | | C1 --- 1mH _)|| >100k | | 1uF | _)|| > '-' | _)|| > | | *| || > | .-+------' || > | |/ || > +---| Q1 || > | |>. 2n3904 || > | | * || > C2 --- +---------. || > 1uF --- | L1b _)|| > | | 25uH _)|| > | | | > === | === > | > '-------------> output > > >Quiet lurker kevin93 <kevin(a)whitedigs.com> was close to the actual >solution. Contact me via my web page mailer to claim your bottle of >wine! > >It's really quite simple to understand. > >Most ALL _high_Q_ sinusoidal oscillators behave this way. > >Assume growing amplitude oscillation (not always a good assumption ;-) > >C-B junction of Q1 forward biases pulling voltage at top of C2 lower, >pulling _voltage_ bias point of Q1 base lower. > >Since one end of L1b is tied to ground, there is less >_conduction_angle_ to forward bias the Q1 emitter, > >When energy added (current) by the shrinking _conduction_angle_ at the >emitter equals the losses in the LC (assume finite Q, otherwise you'll >be forever confused) equilibrium is reached. (So the transistor is >acting as a duty cycle _switch_, NOT as AGC.) > >I experimented with the effects of turns ratio here: > >www.analog-innovations.com/SED/Larkins_LC_OSC.pdf > >At high turns ratios some might claim class-A, but clearly the emitter >CURRENT never becomes sinusoidal... it can't with the simple-minded >low-impedance drive winding driving a non-linear bipolar junction >characteristic. > >At low turns ratios (<5) I observe squegging due to over-dragging C2. > >I'd add at least one additional biasing resistor so that the feedback >coil can't drive the emitter so hard (in series with the feedback >coil). Perhaps an R in parallel with the bias cap so that node can't >rise so high, possibly speeding up start-up. > >For a truly-AGC'd sinusoidal oscillator see the venerable MC1648. As in... http://analog-innovations.com/SED/MC1648-DataSheet.pdf >In >the past year I was commissioned to design a similar architecture on a >modern-day process. (With a fancier AGC so that tank losses could be >accurately measured :-) > But I can't show that... can't give away _real_ circuit design :-) ...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: Phil Hobbs on 15 Jun 2010 14:54 baron wrote: > George Herold Inscribed thus: > >> On Jun 14, 3:29 pm, John Larkin >> <jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote: >>> On Mon, 14 Jun 2010 12:12:17 -0700 (PDT), George Herold >>> >>> <gher...(a)teachspin.com> wrote: >>> >>>> Could someone tell me what squegging is? A google search gives >>>> hints of something like motor-boating, but different? >>> We're referring to an oscillator that oscillates in bursts, rather >>> than continuously. A resonator connected to a negative resistance >>> doesn't squegg, but on the other hand its amplitude builds to >>> infinite (for certain values of infinite) at a rate determined by Q. >>> >>> You can think of an LC resonator as having a 1st order response, >>> measuring oscillation envelope versus drive. Oscillators squegg when >>> some amplitude limiting mechanism adds additional time delays or >>> memory, more poles in the control loop, like the capacitor in the >>> base of my circuit. A bigger cap reduces the tendency to squegg. >>> >>> A superregen receiver is a squegging-on-purpose RF oscillator. That's >>> a fascinating circuit. >>> >>> John >> Hmm, OK something like motorboating gone to the extreme.. enough >> amplitude modultaion to shut the oscillator off for some time. (?) >> Does the squegg rate stay constant in a sugerregen receiver? I'm not >> real sure what a superregen reciever is, anything like a marginal >> oscillator? >> >> George H. > > Its a technique by which an amplifier is held on the verge of > oscillation by positive feedback, at which point the gain is very high. > By causing the oscillation to be quenched at the instant it starts is > where the "Super" bit comes from in "Superregenerative". Commonly used > in radio receivers where a single active device is used to both amplify > the incoming RF, demodulate it and amplify the resultant AF, often > driving headphones directly. Often frowned upon because of the > interference that can be caused by radiated RF. > I think you're confusing it with a Q-multiplier, which is a marginally-stable positive feedback gizmo used to sharpen up crummy RF tank circuits in HF radios. (They're noisy as can be, but you don't care at HF.) A superregen is an oscillator that gets quenched (i.e. turned on and off) at some more or less fixed ultrasonic frequency like 50 kHz. Quench can be internal (due to squegging or blocking), or can be applied externally. Since oscillation has to build up from noise, even a small input signal changes the average output level dramatically. The build-up is exponential, so an input signal e times the thermal noise speeds up the build-up by one whole time constant. If you make the excess gain big enough that the oscillation achieves its full amplitude before the next quench, the average output level goes as the logarithm of the input level. Another way to see this is that making the input signal e times bigger chops off one TC worth of infinitesimal output and adds one TC of full amplitude output, i.e. the duty cycle increases by TC*f_quench per factor of e at the input. With a class-C oscillator, the cathode current demodulates the envelope, which is where the detector aspect comes in. There are a lot of other interesting aspects too--for instance, the carrier of each tone burst is at the free-running frequency of the oscillator, but if you do the Fourier transform, it consists only of the input frequency plus sidebands at multiples of f_quench, and the input frequency component is phase coherent across quench cycles. There are also lots of interesting subtleties depending on whether the quench is sinusoidal or square-wave. Really a pretty circuit. I've been wanting to do an optical version for years, but so far it's never been quite the right hammer for the nail. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal ElectroOptical Innovations 55 Orchard Rd Briarcliff Manor NY 10510 845-480-2058 hobbs at electrooptical dot net http://electrooptical.net
From: George Herold on 15 Jun 2010 15:57 On Jun 15, 2:54 pm, Phil Hobbs <pcdhSpamMeSensel...(a)electrooptical.net> wrote: > baron wrote: > > George Herold Inscribed thus: > > >> On Jun 14, 3:29 pm, John Larkin > >> <jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote: > >>> On Mon, 14 Jun 2010 12:12:17 -0700 (PDT), George Herold > > >>> <gher...(a)teachspin.com> wrote: > > >>>> Could someone tell me what squegging is? A google search gives > >>>> hints of something like motor-boating, but different? > >>> We're referring to an oscillator that oscillates in bursts, rather > >>> than continuously. A resonator connected to a negative resistance > >>> doesn't squegg, but on the other hand its amplitude builds to > >>> infinite (for certain values of infinite) at a rate determined by Q. > > >>> You can think of an LC resonator as having a 1st order response, > >>> measuring oscillation envelope versus drive. Oscillators squegg when > >>> some amplitude limiting mechanism adds additional time delays or > >>> memory, more poles in the control loop, like the capacitor in the > >>> base of my circuit. A bigger cap reduces the tendency to squegg. > > >>> A superregen receiver is a squegging-on-purpose RF oscillator. That's > >>> a fascinating circuit. > > >>> John > >> Hmm, OK something like motorboating gone to the extreme.. enough > >> amplitude modultaion to shut the oscillator off for some time. (?) > >> Does the squegg rate stay constant in a sugerregen receiver? I'm not > >> real sure what a superregen reciever is, anything like a marginal > >> oscillator? > > >> George H. > > > Its a technique by which an amplifier is held on the verge of > > oscillation by positive feedback, at which point the gain is very high. > > By causing the oscillation to be quenched at the instant it starts is > > where the "Super" bit comes from in "Superregenerative". Commonly used > > in radio receivers where a single active device is used to both amplify > > the incoming RF, demodulate it and amplify the resultant AF, often > > driving headphones directly. Often frowned upon because of the > > interference that can be caused by radiated RF. > > I think you're confusing it with a Q-multiplier, which is a > marginally-stable positive feedback gizmo used to sharpen up crummy RF > tank circuits in HF radios. (They're noisy as can be, but you don't > care at HF.) > > A superregen is an oscillator that gets quenched (i.e. turned on and > off) at some more or less fixed ultrasonic frequency like 50 kHz. > Quench can be internal (due to squegging or blocking), or can be applied > externally. > > Since oscillation has to build up from noise, even a small input signal > changes the average output level dramatically. The build-up is > exponential, so an input signal e times the thermal noise speeds up the > build-up by one whole time constant. Excellent! I was reading a few web things about it and this is the critical point that I did not see mentioned. Hmm, seems like you might be able to measure changes in noise with a super-regen too. (I'm not sure how you would calibrate it.) George H. > If you make the excess gain big > enough that the oscillation achieves its full amplitude before the next > quench, the average output level goes as the logarithm of the input level.. > > Another way to see this is that making the input signal e times bigger > chops off one TC worth of infinitesimal output and adds one TC of full > amplitude output, i.e. the duty cycle increases by TC*f_quench per > factor of e at the input. With a class-C oscillator, the cathode > current demodulates the envelope, which is where the detector aspect > comes in. > > There are a lot of other interesting aspects too--for instance, the > carrier of each tone burst is at the free-running frequency of the > oscillator, but if you do the Fourier transform, it consists only of the > input frequency plus sidebands at multiples of f_quench, and the input > frequency component is phase coherent across quench cycles. > > There are also lots of interesting subtleties depending on whether the > quench is sinusoidal or square-wave. > > Really a pretty circuit. I've been wanting to do an optical version for > years, but so far it's never been quite the right hammer for the nail. > > Cheers > > Phil Hobbs > > -- > Dr Philip C D Hobbs > Principal > ElectroOptical Innovations > 55 Orchard Rd > Briarcliff Manor NY 10510 > 845-480-2058 > hobbs at electrooptical dot nethttp://electrooptical.net- Hide quoted text - > > - Show quoted text -
From: Phil Hobbs on 15 Jun 2010 17:32 George Herold wrote: > On Jun 15, 2:54 pm, Phil Hobbs > <pcdhSpamMeSensel...(a)electrooptical.net> wrote: >> baron wrote: >>> George Herold Inscribed thus: >>>> On Jun 14, 3:29 pm, John Larkin >>>> <jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote: >>>>> On Mon, 14 Jun 2010 12:12:17 -0700 (PDT), George Herold >>>>> <gher...(a)teachspin.com> wrote: >>>>>> Could someone tell me what squegging is? A google search gives >>>>>> hints of something like motor-boating, but different? >>>>> We're referring to an oscillator that oscillates in bursts, rather >>>>> than continuously. A resonator connected to a negative resistance >>>>> doesn't squegg, but on the other hand its amplitude builds to >>>>> infinite (for certain values of infinite) at a rate determined by Q. >>>>> You can think of an LC resonator as having a 1st order response, >>>>> measuring oscillation envelope versus drive. Oscillators squegg when >>>>> some amplitude limiting mechanism adds additional time delays or >>>>> memory, more poles in the control loop, like the capacitor in the >>>>> base of my circuit. A bigger cap reduces the tendency to squegg. >>>>> A superregen receiver is a squegging-on-purpose RF oscillator. That's >>>>> a fascinating circuit. >>>>> John >>>> Hmm, OK something like motorboating gone to the extreme.. enough >>>> amplitude modultaion to shut the oscillator off for some time. (?) >>>> Does the squegg rate stay constant in a sugerregen receiver? I'm not >>>> real sure what a superregen reciever is, anything like a marginal >>>> oscillator? >>>> George H. >>> Its a technique by which an amplifier is held on the verge of >>> oscillation by positive feedback, at which point the gain is very high. >>> By causing the oscillation to be quenched at the instant it starts is >>> where the "Super" bit comes from in "Superregenerative". Commonly used >>> in radio receivers where a single active device is used to both amplify >>> the incoming RF, demodulate it and amplify the resultant AF, often >>> driving headphones directly. Often frowned upon because of the >>> interference that can be caused by radiated RF. >> I think you're confusing it with a Q-multiplier, which is a >> marginally-stable positive feedback gizmo used to sharpen up crummy RF >> tank circuits in HF radios. (They're noisy as can be, but you don't >> care at HF.) >> >> A superregen is an oscillator that gets quenched (i.e. turned on and >> off) at some more or less fixed ultrasonic frequency like 50 kHz. >> Quench can be internal (due to squegging or blocking), or can be applied >> externally. >> >> Since oscillation has to build up from noise, even a small input signal >> changes the average output level dramatically. The build-up is >> exponential, so an input signal e times the thermal noise speeds up the >> build-up by one whole time constant. > > Excellent! I was reading a few web things about it and this is the > critical point that I did not see mentioned. > > Hmm, seems like you might be able to measure changes in noise with a > super-regen too. (I'm not sure how you would calibrate it.) > > George H. > The rushing sound from the speaker is amplified thermal noise. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal ElectroOptical Innovations 55 Orchard Rd Briarcliff Manor NY 10510 845-480-2058 hobbs at electrooptical dot net http://electrooptical.net
First
|
Prev
|
Next
|
Last
Pages: 1 2 3 4 5 Prev: harnessing lightning, or not Next: One mouse click, 2 PC's |