On Tue, 13 May 2014 09:31:15 -0700, George Herold wrote:> On Tuesday, May 13, 2014 12:03:34 PM UTC-4, Tim Wescott wrote: >> On Tue, 13 May 2014 07:36:36 -0700, George Herold wrote: >> > <snip> >> >> How close to true quadrature and amplitude matching do you need? >> You'll be pretty dependent on both capacitor and op-amp characteristics >> for the match and phase shift between channels. >> >> >> >> If you really want precision, run a DDS at a high clock rate so the >> filtering can't add much phase shift or attenuation. >> > Yeah thanks Tim. Right, DDS would be the way to go. > (But there's that whole dang learning curve and then how to do the user > interface.) > Have you done a DDS design? > I guess I was just doing what I know already. > > I don't care too much about the amplitudes. > and I don't think a bit (few degrees) of phase shift will be that bad. > > I was thinking if I could get a cheap signal generator that would output > I and Q then that would solve my problem too. > > OK time to read my Rigol DG1022 manual.Why yes, I have done a DDS design. In 1989. It worked great, but I suspect the parts are completely unobtainable by now. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com
Quadrature Oscillator
Started by ●May 13, 2014
Reply by ●May 13, 20142014-05-13
Reply by ●May 13, 20142014-05-13
On Tuesday, May 13, 2014 1:27:28 PM UTC-4, Jim Thompson wrote:> On Tue, 13 May 2014 08:58:48 -0700 (PDT), George Herold > > <gherold@teachspin.com> wrote: > > > > >On Tuesday, May 13, 2014 11:36:10 AM UTC-4, Jim Thompson wrote: > > >> On Tue, 13 May 2014 07:36:36 -0700 (PDT), George Herold > > >> > > >> <gherold@teachspin.com> wrote: > > >> > > >> > > >> > > >> >Quadrature oscialltor. > > > >> How about this... > >> > >> <http://www.analog-innovations.com/SED/GyratorFilter_A1A_Oscillator.pdf> > >> > >OK I had to redraw it to make it look like the gyrator in AoE. > > >(fig 5.24 Z1=R9, Z2=R7, Z3=r6, Z4=C3, Z5=R8) > > > > > I've never used a gyrator... How much gain is there in any one opamp stage? > > Walk thru the stages... first stage is essentially an integrator,What? You mean I should think about it? :^)> second stage has inverting gain of unity. > > Laplace (Heaviside version) is the easiest way to analyze these > structures. > > > > >And what's going on with U6.. there's a diode clipped gain control, > >but how does anything get through D5 and D6? Capacitive coupling? > > > > > >George H. > > > > > Yep, the Q is so high it takes a bare twiddle to make it go.OK.. From my limited oscillator experience it will depend on how good the cap is. (good, as in low dissipation.) George H.> > I threw this together some 8 years ago to demonstrate the gyrator > > band-pass effect. IRL I used an AGC loop. (An accurate clipper works > > also.) > > > > You can get really low distortion with good OpAmps. I built one for a > > Sperry Flight Systems synchronous detect at 400Hz... 0.001% harmonic > > distortion :-) > > > > ...Jim Thompson > > -- > > | James E.Thompson | mens | > > | Analog Innovations | et | > > | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | > > | San Tan Valley, AZ 85142 Skype: Contacts Only | | > > | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | > > | E-mail Icon at http://www.analog-innovations.com | 1962 | > > > > I love to cook with wine. Sometimes I even put it in the food.
Reply by ●May 13, 20142014-05-13
On Tue, 13 May 2014 12:50:50 -0700 (PDT), George Herold <gherold@teachspin.com> wrote:>On Tuesday, May 13, 2014 1:27:28 PM UTC-4, Jim Thompson wrote: >> On Tue, 13 May 2014 08:58:48 -0700 (PDT), George Herold >> >> <gherold@teachspin.com> wrote: >> >> >> >> >On Tuesday, May 13, 2014 11:36:10 AM UTC-4, Jim Thompson wrote: >> >> >> On Tue, 13 May 2014 07:36:36 -0700 (PDT), George Herold >> >> >> >> >> >> <gherold@teachspin.com> wrote: >> >> >> >> >> >> >> >> >> >> >> >> >Quadrature oscialltor. >> >> >> >> How about this... >> >> >> >> <http://www.analog-innovations.com/SED/GyratorFilter_A1A_Oscillator.pdf> >> >> >> >OK I had to redraw it to make it look like the gyrator in AoE. >> >> >(fig 5.24 Z1=R9, Z2=R7, Z3=r6, Z4=C3, Z5=R8) >> >> > >> > I've never used a gyrator... How much gain is there in any one opamp stage? >> >> Walk thru the stages... first stage is essentially an integrator, >What? You mean I should think about it? :^) >> second stage has inverting gain of unity. >> >> Laplace (Heaviside version) is the easiest way to analyze these >> structures. >> >> >> >> >And what's going on with U6.. there's a diode clipped gain control, >> >but how does anything get through D5 and D6? Capacitive coupling? >> >> > >> >> >George H. >> >> > >> >> Yep, the Q is so high it takes a bare twiddle to make it go. >OK.. From my limited oscillator experience it will depend on how good >the cap is. (good, as in low dissipation.) > >George H. >> >> I threw this together some 8 years ago to demonstrate the gyrator >> >> band-pass effect. IRL I used an AGC loop. (An accurate clipper works >> >> also.) >> >> >> >> You can get really low distortion with good OpAmps. I built one for a >> >> Sperry Flight Systems synchronous detect at 400Hz... 0.001% harmonic >> >> distortion :-) >> >> >> >> ...Jim Thompson[snip] If you run the numbers you'll find that dissipation factor (in active filters) is only a problem at low frequencies. See... <http://www.analog-innovations.com/SED/StateVariableFilter(P+1).pdf> for how I cured (in the early '70's) the dissipation factor problem in telephone-signaling low-frequency filters. ...Jim Thompson -- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | San Tan Valley, AZ 85142 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | I love to cook with wine. Sometimes I even put it in the food.
Reply by ●May 13, 20142014-05-13
On Tuesday, May 13, 2014 2:37:00 PM UTC-4, Tim Wescott wrote:> On Tue, 13 May 2014 13:03:28 -0400, Phil Hobbs wrote: > > > > > On 05/13/2014 12:03 PM, Tim Wescott wrote: > > >> On Tue, 13 May 2014 07:36:36 -0700, George Herold wrote: > > >> > > >>> Quadrature oscialltor. > > >>> > > >>> Hi all, As part of the CV measurements (from a previous post) > > >>> I'm thinking about a 1 MHz lockin. > > >>> (Maybe a different thread to talk about how to do the switching.) > > >>> I was wondering if I could make a quadrature oscillator from opamps. > > >>> (I've got some AD825's I thought I'd try.) > > >>> Anyway spice seems to think it's possbile. > > >>> Comments or ideas welcome. > > >>> (Oh I had to add the sine wave input to get the thing to start.) You > > >>> can play with C1 R7 C2 R8 to change the frequency. > > >>> > > >>> Here's a circuit scribble if you don't like the spice file > > >>> https://www.dropbox.com/s/sb5azyl69fy142t/DSCF0012.JPG > > >> > > >> How close to true quadrature and amplitude matching do you need? > > >> You'll be pretty dependent on both capacitor and op-amp characteristics > > >> for the match and phase shift between channels. > > >> > > >> If you really want precision, run a DDS at a high clock rate so the > > >> filtering can't add much phase shift or attenuation. > > >> > > >> > > > One method is to make one section a true integrator. That'll guarantee > > > quadrature. Then servo the frequency to make the amplitudes equal. > > > You'll need to measure the frequency to compute the capacitance, but > > > that's not so bad. > > > > That works if a "true" integrator is true enough at 1MHz, but leaves you > with the problem of balancing the amplitudes. > > > I'd be tempted to integrate-integrate-invert. It would take three op- > amps, but you may have a better chance to get each section working > correctly. Of course you'd need the obligatory teeny bit of lag > somewhere, to make the oscillation start.Hmm OK, I always figured it just needed the right bit of gain somewhere. I've never had that much problem getting oscillators to start...They always seem to get going and bump into the rail right away. (I love the way LTspice goes up to several tera volts, before it pukes.... Can you even make a tera volt without ripping apart the fabric of space-time?) George H.> > > > -- > > > > Tim Wescott > > Wescott Design Services > > http://www.wescottdesign.com
Reply by ●May 13, 20142014-05-13
On Tue, 13 May 2014 12:50:50 -0700 (PDT), George Herold <gherold@teachspin.com> wrote:>On Tuesday, May 13, 2014 1:27:28 PM UTC-4, Jim Thompson wrote: >> On Tue, 13 May 2014 08:58:48 -0700 (PDT), George Herold >> >> <gherold@teachspin.com> wrote: >> >> >> >> >On Tuesday, May 13, 2014 11:36:10 AM UTC-4, Jim Thompson wrote: >> >> >> On Tue, 13 May 2014 07:36:36 -0700 (PDT), George Herold >> >> >> >> >> >> <gherold@teachspin.com> wrote: >> >> >> >> >> >> >> >> >> >> >> >> >Quadrature oscialltor. >> >> >> >> How about this... >> >> >> >> <http://www.analog-innovations.com/SED/GyratorFilter_A1A_Oscillator.pdf> >> >> >> >OK I had to redraw it to make it look like the gyrator in AoE. >> >> >(fig 5.24 Z1=R9, Z2=R7, Z3=r6, Z4=C3, Z5=R8) >> >> > >> > I've never used a gyrator... How much gain is there in any one opamp stage? >> >> Walk thru the stages... first stage is essentially an integrator, >What? You mean I should think about it? :^) >> second stage has inverting gain of unity. >> >> Laplace (Heaviside version) is the easiest way to analyze these >> structures. >> >> >> >> >And what's going on with U6.. there's a diode clipped gain control, >> >but how does anything get through D5 and D6? Capacitive coupling? >> >> > >> >> >George H. >> >> > >> >> Yep, the Q is so high it takes a bare twiddle to make it go. >OK.. From my limited oscillator experience it will depend on how good >the cap is. (good, as in low dissipation.) > >George H. >> >> I threw this together some 8 years ago to demonstrate the gyrator >> >> band-pass effect. IRL I used an AGC loop. (An accurate clipper works >> >> also.) >> >> >> >> You can get really low distortion with good OpAmps. I built one for a >> >> Sperry Flight Systems synchronous detect at 400Hz... 0.001% harmonic >> >> distortion :-) >> >>[snip] Here's a link to the BP filter math... <http://www.analog-innovations.com/SED/UglyXferVsGyrator.pdf> ...Jim Thompson -- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | San Tan Valley, AZ 85142 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | I love to cook with wine. Sometimes I even put it in the food.
Reply by ●May 13, 20142014-05-13
On Tue, 13 May 2014 13:00:00 -0700 (PDT), George Herold <gherold@teachspin.com> wrote:>On Tuesday, May 13, 2014 2:37:00 PM UTC-4, Tim Wescott wrote: >> On Tue, 13 May 2014 13:03:28 -0400, Phil Hobbs wrote: >> >> >> >> > On 05/13/2014 12:03 PM, Tim Wescott wrote: >> >> >> On Tue, 13 May 2014 07:36:36 -0700, George Herold wrote: >> >> >> >> >> >>> Quadrature oscialltor. >> >> >>> >> >> >>> Hi all, As part of the CV measurements (from a previous post) >> >> >>> I'm thinking about a 1 MHz lockin. >> >> >>> (Maybe a different thread to talk about how to do the switching.) >> >> >>> I was wondering if I could make a quadrature oscillator from opamps. >> >> >>> (I've got some AD825's I thought I'd try.) >> >> >>> Anyway spice seems to think it's possbile. >> >> >>> Comments or ideas welcome. >> >> >>> (Oh I had to add the sine wave input to get the thing to start.) You >> >> >>> can play with C1 R7 C2 R8 to change the frequency. >> >> >>> >> >> >>> Here's a circuit scribble if you don't like the spice file >> >> >>> https://www.dropbox.com/s/sb5azyl69fy142t/DSCF0012.JPG >> >> >> >> >> >> How close to true quadrature and amplitude matching do you need? >> >> >> You'll be pretty dependent on both capacitor and op-amp characteristics >> >> >> for the match and phase shift between channels. >> >> >> >> >> >> If you really want precision, run a DDS at a high clock rate so the >> >> >> filtering can't add much phase shift or attenuation. >> >> >> >> >> >> >> >> > One method is to make one section a true integrator. That'll guarantee >> >> > quadrature. Then servo the frequency to make the amplitudes equal. >> >> > You'll need to measure the frequency to compute the capacitance, but >> >> > that's not so bad. >> >> >> >> That works if a "true" integrator is true enough at 1MHz, but leaves you >> with the problem of balancing the amplitudes. >> >> >> I'd be tempted to integrate-integrate-invert. It would take three op- >> amps, but you may have a better chance to get each section working >> correctly. Of course you'd need the obligatory teeny bit of lag >> somewhere, to make the oscillation start. > >Hmm OK, I always figured it just needed the right bit of gain somewhere. >I've never had that much problem getting oscillators to start...They always seem to get going and bump into the rail right away. (I love the way LTspice goes up to several tera volts, before it pukes.... >Can you even make a tera volt without ripping apart the fabric of space-time?) > >George H. >>[snip] I don't have that problem... I devised parts that know to stop at some bounds, like the rails... <http://www.analog-innovations.com/CompliantParts.zip> Also makes for faster convergence. ...Jim Thompson -- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | San Tan Valley, AZ 85142 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | I love to cook with wine. Sometimes I even put it in the food.
Reply by ●May 13, 20142014-05-13
Tim Wescott <tim@seemywebsite.really> wrote:> On Tue, 13 May 2014 13:03:28 -0400, Phil Hobbs wrote: > >> On 05/13/2014 12:03 PM, Tim Wescott wrote: >>> On Tue, 13 May 2014 07:36:36 -0700, George Herold wrote: >>> >>>> Quadrature oscialltor. >>>> >>>> Hi all, As part of the CV measurements (from a previous post) >>>> I'm thinking about a 1 MHz lockin. >>>> (Maybe a different thread to talk about how to do the switching.) >>>> I was wondering if I could make a quadrature oscillator from >>>> opamps. (I've got some AD825's I thought I'd try.) >>>> Anyway spice seems to think it's possbile. >>>> Comments or ideas welcome. >>>> (Oh I had to add the sine wave input to get the thing to start.) >>>> You can play with C1 R7 C2 R8 to change the frequency. >>>> >>>> Here's a circuit scribble if you don't like the spice file >>>> https://www.dropbox.com/s/sb5azyl69fy142t/DSCF0012.JPG >>> >>> How close to true quadrature and amplitude matching do you need? >>> You'll be pretty dependent on both capacitor and op-amp >>> characteristics for the match and phase shift between channels. >>> >>> If you really want precision, run a DDS at a high clock rate so the >>> filtering can't add much phase shift or attenuation. >>> >>> >> One method is to make one section a true integrator. That'll >> guarantee quadrature. Then servo the frequency to make the >> amplitudes equal. You'll need to measure the frequency to compute the >> capacitance, but that's not so bad. > > That works if a "true" integrator is true enough at 1MHz, but leaves > you with the problem of balancing the amplitudes. > > I'd be tempted to integrate-integrate-invert. It would take three op- > amps, but you may have a better chance to get each section working > correctly. Of course you'd need the obligatory teeny bit of lag > somewhere, to make the oscillation start. >This is getting way too complicated. Analog circuits drift and can never guarantee quadrature. A simple 74HC74 will. See http://www.wb5rvz.com/sdr/ensemble_rx_ii_vhf/04_div.htm
Reply by ●May 13, 20142014-05-13
On Tue, 13 May 2014 13:00:00 -0700, George Herold wrote:> On Tuesday, May 13, 2014 2:37:00 PM UTC-4, Tim Wescott wrote: >> On Tue, 13 May 2014 13:03:28 -0400, Phil Hobbs wrote: >> >> >> >> > On 05/13/2014 12:03 PM, Tim Wescott wrote: >> >> >> On Tue, 13 May 2014 07:36:36 -0700, George Herold wrote: >> >> >> >> >> >>> Quadrature oscialltor. >> >> >> >>> >> >>> Hi all, As part of the CV measurements (from a previous post) >> >> >>> I'm thinking about a 1 MHz lockin. >> >> >>> (Maybe a different thread to talk about how to do the switching.) >> >> >>> I was wondering if I could make a quadrature oscillator from >> >>> opamps. >> >> >>> (I've got some AD825's I thought I'd try.) >> >> >>> Anyway spice seems to think it's possbile. >> >> >>> Comments or ideas welcome. >> >> >>> (Oh I had to add the sine wave input to get the thing to start.) >> >>> You >> >> >>> can play with C1 R7 C2 R8 to change the frequency. >> >> >> >>> >> >>> Here's a circuit scribble if you don't like the spice file >> >> >>> https://www.dropbox.com/s/sb5azyl69fy142t/DSCF0012.JPG >> >> >> >> >> >> How close to true quadrature and amplitude matching do you need? >> >> >> You'll be pretty dependent on both capacitor and op-amp >> >> characteristics >> >> >> for the match and phase shift between channels. >> >> >> >> >> >> If you really want precision, run a DDS at a high clock rate so the >> >> >> filtering can't add much phase shift or attenuation. >> >> >> >> >> >> >> >> > One method is to make one section a true integrator. That'll >> > guarantee >> >> > quadrature. Then servo the frequency to make the amplitudes equal. >> >> > You'll need to measure the frequency to compute the capacitance, but >> >> > that's not so bad. >> >> >> >> That works if a "true" integrator is true enough at 1MHz, but leaves >> you with the problem of balancing the amplitudes. >> >> >> I'd be tempted to integrate-integrate-invert. It would take three op- >> amps, but you may have a better chance to get each section working >> correctly. Of course you'd need the obligatory teeny bit of lag >> somewhere, to make the oscillation start. > > Hmm OK, I always figured it just needed the right bit of gain somewhere. > I've never had that much problem getting oscillators to start...They > always seem to get going and bump into the rail right away. (I love the > way LTspice goes up to several tera volts, before it pukes.... > Can you even make a tera volt without ripping apart the fabric of > space-time?)It's field strength, not voltage, that would potentially (pun not intended) rip apart the fabric of space-time. Assuming that you could achieve a high enough field strength without ripping electrons off of your negative electrode, you would eventually generate real positron/electron pairs from the quantum foam. I think -- I'm not a quantum mechanic. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com
Reply by ●May 13, 20142014-05-13
On Tuesday, May 13, 2014 1:15:54 PM UTC-7, John Silverman wrote:> This is getting way too complicated. Analog circuits drift and can never > guarantee quadrature. A simple 74HC74 will. See > > > > http://www.wb5rvz.com/sdr/ensemble_rx_ii_vhf/04_div.htmThere's ways to make an analog signal lock-in to a digital quadrature clock, of course. There's also a switched-capacitor solution, see figure 15 <http://www.ti.com/lit/an/snoa572c/snoa572c.pdf> And, if one wishes to go old-school, an LC oscillator puts out sin(wt) voltage on the capacitor, and cos(wt) current through the inductor; you can make a sense coil lightly coupled to the inductor to pick off the second phase. If Q is high, it needs no correction; if Q is low, you can make a weighted sum of the two signals that is tunable to get closer to 90.00 degrees phase, and a simple multiplicative mixer makes a DC signal of the error, to guide the tuning.
Reply by ●May 13, 20142014-05-13
On Tue, 13 May 2014 20:15:54 +0000, John Silverman wrote:> Tim Wescott <tim@seemywebsite.really> wrote: > >> On Tue, 13 May 2014 13:03:28 -0400, Phil Hobbs wrote: >> >>> On 05/13/2014 12:03 PM, Tim Wescott wrote: >>>> On Tue, 13 May 2014 07:36:36 -0700, George Herold wrote: >>>> >>>>> Quadrature oscialltor. >>>>> >>>>> Hi all, As part of the CV measurements (from a previous post) I'm >>>>> thinking about a 1 MHz lockin. >>>>> (Maybe a different thread to talk about how to do the switching.) >>>>> I was wondering if I could make a quadrature oscillator from opamps. >>>>> (I've got some AD825's I thought I'd try.) >>>>> Anyway spice seems to think it's possbile. >>>>> Comments or ideas welcome. >>>>> (Oh I had to add the sine wave input to get the thing to start.) You >>>>> can play with C1 R7 C2 R8 to change the frequency. >>>>> >>>>> Here's a circuit scribble if you don't like the spice file >>>>> https://www.dropbox.com/s/sb5azyl69fy142t/DSCF0012.JPG >>>> >>>> How close to true quadrature and amplitude matching do you need? >>>> You'll be pretty dependent on both capacitor and op-amp >>>> characteristics for the match and phase shift between channels. >>>> >>>> If you really want precision, run a DDS at a high clock rate so the >>>> filtering can't add much phase shift or attenuation. >>>> >>>> >>> One method is to make one section a true integrator. That'll >>> guarantee quadrature. Then servo the frequency to make the amplitudes >>> equal. You'll need to measure the frequency to compute the >>> capacitance, but that's not so bad. >> >> That works if a "true" integrator is true enough at 1MHz, but leaves >> you with the problem of balancing the amplitudes. >> >> I'd be tempted to integrate-integrate-invert. It would take three op- >> amps, but you may have a better chance to get each section working >> correctly. Of course you'd need the obligatory teeny bit of lag >> somewhere, to make the oscillation start. >> >> > This is getting way too complicated. Analog circuits drift and can never > guarantee quadrature. A simple 74HC74 will. See > > http://www.wb5rvz.com/sdr/ensemble_rx_ii_vhf/04_div.htmIf George can deal with square waves, yes, that's going to work way better than an analog solution. If he needs sine waves, by the time he's gotten the harmonics beaten out of that he'll be back to depending on analog components for his precision. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com
