On 9/21/2013 9:06 PM, Jim Thompson wrote:> On Sat, 21 Sep 2013 10:51:06 -0700, John Larkin > <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote: > >> On Sat, 21 Sep 2013 10:26:40 -0700, Jim Thompson >> <To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote: >> >>> Here's a way to do capacitor multiplication... >>> >>> >>> <http://www.analog-innovations.com/SED/CapMultiplier_JT_2013-09-21.pdf> >>> >>> limited only by OpAmp GBW product, and no resistive components until >>> the OpAmps run out of steam. >>> >>> Mathematically, this is the same way I did behavioral models for >>> crummy capacitors such as X7R and Y5U, except I had ideal amplifiers, >>> and equations and/or tables to describe the changes versus voltage and >>> temperature. >>> >>> Larkin's "Maybe I'll do this:" >>> >>> >>> <https://dl.dropboxusercontent.com/u/53724080/Parts/Caps/Trim_Cap.JPG> >>> >>> does not fare so well. Do the math. >>> >>> ...Jim Thompson >> >> >> 2.5x the parts (including a GHz opamp), maybe 50x the noise. It would destroy >> the low-noise transimpedance amp that I'm doing. You have just invented the >> world's noisiest capacitor. Well, you probably "invented" it 40 years ago. >> >> And I need a floating capacitor, not a grounded one. And I need about 0.4 pF, >> which would call for a capacitance divider, not a multiplier. >> > > Larkinus Bloviatus Fartus Maximus: > >> Net idiotic. > > I asked, "Do the math." > > Larkin didn't. > > I did. > > Though any kid fresh out of school would look at this... > > <http://www.analog-innovations.com/SED/Larkin_Trim_Cap_Folly.jpg> > > And note, by simple observation, several things... > > It's not a two terminal network. > > The impedance looking into the node on the left (marked arrow-style by > Larkin) looks like that resistor from the arrow to IN- of the OpAmp > (until the OpAmp runs into the GBW stop). > > The impedance looking into the right arrow is, in essence, a voltage > source, zero AC impedance, again until the OpAmp runs into the GBW > stop. > > If that's a _floating_ capacitor, I'm the King of Siam. > > But Larkin's sickophants should be happy, their low-information > mentality, and need for name-calling, has been satisfied. They will > now face toward San Fransicko and chant... >:-} > > ...Jim Thompson >I thought you were trying not to dish that stuff out? Using a voltage divider on a small compensation cap is a pretty common approach in TIAs, though IME it's usually a capacitive divider rather than a resistive one. It works OK up to the point where the summing junction loading is important. You clearly don't want to try synthesizing 1 pF using a 1000:1 voltage divider and a nanofarad in series.(*) Using a pot, you have to worry about the current noise of the low resistances getting into the summing junction. (I posted a TIA fragment a week or two back where I had to short out a 1G resistor with a relay to avoid its Johnson noise current dominating the circuit performance above a few hundred hertz.) However, given that the input capacitance is probably nearly 2 pF to begin with, something like 0.5 pF and a low-value pot isn't necessarily a silly way of making an adjustable 0.3 pF for that sort of use. Cheers Phil Hobbs (*) Back in the early part of my career, I actually managed to make an op amp integrator oscillate all by itself. Fast op amp with high Zout and some phase funnies in its compensation, small R_in/large C_F. -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 USA +1 845 480 2058 hobbs at electrooptical dot net http://electrooptical.net
Cap Multiplier, NOT a "gimmick"
Started by ●September 21, 2013
Reply by ●September 22, 20132013-09-22
Reply by ●September 22, 20132013-09-22
On Sun, 22 Sep 2013 15:31:57 -0400, Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:>On 9/21/2013 9:06 PM, Jim Thompson wrote: >> On Sat, 21 Sep 2013 10:51:06 -0700, John Larkin >> <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote: >> >>> On Sat, 21 Sep 2013 10:26:40 -0700, Jim Thompson >>> <To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote: >>> >>>> Here's a way to do capacitor multiplication... >>>> >>>> >>>> <http://www.analog-innovations.com/SED/CapMultiplier_JT_2013-09-21.pdf> >>>> >>>> limited only by OpAmp GBW product, and no resistive components until >>>> the OpAmps run out of steam. >>>> >>>> Mathematically, this is the same way I did behavioral models for >>>> crummy capacitors such as X7R and Y5U, except I had ideal amplifiers, >>>> and equations and/or tables to describe the changes versus voltage and >>>> temperature. >>>> >>>> Larkin's "Maybe I'll do this:" >>>> >>>> >>>> <https://dl.dropboxusercontent.com/u/53724080/Parts/Caps/Trim_Cap.JPG> >>>> >>>> does not fare so well. Do the math. >>>> >>>> ...Jim Thompson >>> >>> >>> 2.5x the parts (including a GHz opamp), maybe 50x the noise. It would destroy >>> the low-noise transimpedance amp that I'm doing. You have just invented the >>> world's noisiest capacitor. Well, you probably "invented" it 40 years ago. >>> >>> And I need a floating capacitor, not a grounded one. And I need about 0.4 pF, >>> which would call for a capacitance divider, not a multiplier. >>> >> >> Larkinus Bloviatus Fartus Maximus: >> >>> Net idiotic. >> >> I asked, "Do the math." >> >> Larkin didn't. >> >> I did. >> >> Though any kid fresh out of school would look at this... >> >> <http://www.analog-innovations.com/SED/Larkin_Trim_Cap_Folly.jpg> >> >> And note, by simple observation, several things... >> >> It's not a two terminal network. >> >> The impedance looking into the node on the left (marked arrow-style by >> Larkin) looks like that resistor from the arrow to IN- of the OpAmp >> (until the OpAmp runs into the GBW stop). >> >> The impedance looking into the right arrow is, in essence, a voltage >> source, zero AC impedance, again until the OpAmp runs into the GBW >> stop. >> >> If that's a _floating_ capacitor, I'm the King of Siam. >> >> But Larkin's sickophants should be happy, their low-information >> mentality, and need for name-calling, has been satisfied. They will >> now face toward San Fransicko and chant... >:-} >> >> ...Jim Thompson >> > >I thought you were trying not to dish that stuff out? > >Using a voltage divider on a small compensation cap is a pretty common >approach in TIAs, though IME it's usually a capacitive divider rather >than a resistive one. It works OK up to the point where the summing >junction loading is important. You clearly don't want to try >synthesizing 1 pF using a 1000:1 voltage divider and a nanofarad in >series.(*) > >Using a pot, you have to worry about the current noise of the low >resistances getting into the summing junction. (I posted a TIA fragment >a week or two back where I had to short out a 1G resistor with a relay >to avoid its Johnson noise current dominating the circuit performance >above a few hundred hertz.) > >However, given that the input capacitance is probably nearly 2 pF to >begin with, something like 0.5 pF and a low-value pot isn't necessarily >a silly way of making an adjustable 0.3 pF for that sort of use.Actually defining the total problem would have been a more useful exercise.> >Cheers > >Phil Hobbs > >(*) Back in the early part of my career, I actually managed to make an >op amp integrator oscillate all by itself. Fast op amp with high Zout >and some phase funnies in its compensation, small R_in/large C_F.Nowhere in Larkin's multiple posts was "capacitor immersed in an integrator" mentioned. When I show a true capacitor multiplier, I get his usual ration of shit... and bloviations about "floating". ...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 ●September 22, 20132013-09-22
On Sun, 22 Sep 2013 15:31:57 -0400, Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:>On 9/21/2013 9:06 PM, Jim Thompson wrote: >> On Sat, 21 Sep 2013 10:51:06 -0700, John Larkin >> <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote: >> >>> On Sat, 21 Sep 2013 10:26:40 -0700, Jim Thompson >>> <To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote: >>> >>>> Here's a way to do capacitor multiplication... >>>> >>>> >>>> <http://www.analog-innovations.com/SED/CapMultiplier_JT_2013-09-21.pdf> >>>> >>>> limited only by OpAmp GBW product, and no resistive components until >>>> the OpAmps run out of steam. >>>> >>>> Mathematically, this is the same way I did behavioral models for >>>> crummy capacitors such as X7R and Y5U, except I had ideal amplifiers, >>>> and equations and/or tables to describe the changes versus voltage and >>>> temperature. >>>> >>>> Larkin's "Maybe I'll do this:" >>>> >>>> >>>> <https://dl.dropboxusercontent.com/u/53724080/Parts/Caps/Trim_Cap.JPG> >>>> >>>> does not fare so well. Do the math. >>>> >>>> ...Jim Thompson >>> >>> >>> 2.5x the parts (including a GHz opamp), maybe 50x the noise. It would destroy >>> the low-noise transimpedance amp that I'm doing. You have just invented the >>> world's noisiest capacitor. Well, you probably "invented" it 40 years ago. >>> >>> And I need a floating capacitor, not a grounded one. And I need about 0.4 pF, >>> which would call for a capacitance divider, not a multiplier. >>> >> >> Larkinus Bloviatus Fartus Maximus: >> >>> Net idiotic. >> >> I asked, "Do the math." >> >> Larkin didn't. >> >> I did. >> >> Though any kid fresh out of school would look at this... >> >> <http://www.analog-innovations.com/SED/Larkin_Trim_Cap_Folly.jpg> >> >> And note, by simple observation, several things... >> >> It's not a two terminal network. >> >> The impedance looking into the node on the left (marked arrow-style by >> Larkin) looks like that resistor from the arrow to IN- of the OpAmp >> (until the OpAmp runs into the GBW stop). >> >> The impedance looking into the right arrow is, in essence, a voltage >> source, zero AC impedance, again until the OpAmp runs into the GBW >> stop. >> >> If that's a _floating_ capacitor, I'm the King of Siam. >> >> But Larkin's sickophants should be happy, their low-information >> mentality, and need for name-calling, has been satisfied. They will >> now face toward San Fransicko and chant... >:-} >> >> ...Jim Thompson >> > >I thought you were trying not to dish that stuff out? > >Using a voltage divider on a small compensation cap is a pretty common >approach in TIAs, though IME it's usually a capacitive divider rather >than a resistive one. It works OK up to the point where the summing >junction loading is important. You clearly don't want to try >synthesizing 1 pF using a 1000:1 voltage divider and a nanofarad in >series.(*) > >Using a pot, you have to worry about the current noise of the low >resistances getting into the summing junction. (I posted a TIA fragment >a week or two back where I had to short out a 1G resistor with a relay >to avoid its Johnson noise current dominating the circuit performance >above a few hundred hertz.) > >However, given that the input capacitance is probably nearly 2 pF to >begin with, something like 0.5 pF and a low-value pot isn't necessarily >a silly way of making an adjustable 0.3 pF for that sort of use. > >Cheers > >Phil Hobbs > >(*) Back in the early part of my career, I actually managed to make an >op amp integrator oscillate all by itself. Fast op amp with high Zout >and some phase funnies in its compensation, small R_in/large C_F.Oscillation isn't an unusual event; stability is! I sim'd the pot thing (yes, you prefer algebra) and it had an inconsequential effect on noise. I'd only do this on the breadboard. I like being able to calculate the effective capacitance after I twiddle the pot; I can't think of any other trimmer that easily allows that. -- John Larkin Highland Technology Inc www.highlandtechnology.com jlarkin at highlandtechnology dot com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom timing and laser controllers Photonics and fiberoptic TTL data links VME analog, thermocouple, LVDT, synchro, tachometer Multichannel arbitrary waveform generators
Reply by ●September 22, 20132013-09-22
On Sun, 22 Sep 2013 14:37:18 -0700, Jim Thompson <To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote:>On Sun, 22 Sep 2013 15:31:57 -0400, Phil Hobbs ><pcdhSpamMeSenseless@electrooptical.net> wrote: > >>On 9/21/2013 9:06 PM, Jim Thompson wrote: >>> On Sat, 21 Sep 2013 10:51:06 -0700, John Larkin >>> <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote: >>> >>>> On Sat, 21 Sep 2013 10:26:40 -0700, Jim Thompson >>>> <To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote: >>>> >>>>> Here's a way to do capacitor multiplication... >>>>> >>>>> >>>>> <http://www.analog-innovations.com/SED/CapMultiplier_JT_2013-09-21.pdf> >>>>> >>>>> limited only by OpAmp GBW product, and no resistive components until >>>>> the OpAmps run out of steam. >>>>> >>>>> Mathematically, this is the same way I did behavioral models for >>>>> crummy capacitors such as X7R and Y5U, except I had ideal amplifiers, >>>>> and equations and/or tables to describe the changes versus voltage and >>>>> temperature. >>>>> >>>>> Larkin's "Maybe I'll do this:" >>>>> >>>>> >>>>> <https://dl.dropboxusercontent.com/u/53724080/Parts/Caps/Trim_Cap.JPG> >>>>> >>>>> does not fare so well. Do the math. >>>>> >>>>> ...Jim Thompson >>>> >>>> >>>> 2.5x the parts (including a GHz opamp), maybe 50x the noise. It would destroy >>>> the low-noise transimpedance amp that I'm doing. You have just invented the >>>> world's noisiest capacitor. Well, you probably "invented" it 40 years ago. >>>> >>>> And I need a floating capacitor, not a grounded one. And I need about 0.4 pF, >>>> which would call for a capacitance divider, not a multiplier. >>>> >>> >>> Larkinus Bloviatus Fartus Maximus: >>> >>>> Net idiotic. >>> >>> I asked, "Do the math." >>> >>> Larkin didn't. >>> >>> I did. >>> >>> Though any kid fresh out of school would look at this... >>> >>> <http://www.analog-innovations.com/SED/Larkin_Trim_Cap_Folly.jpg> >>> >>> And note, by simple observation, several things... >>> >>> It's not a two terminal network. >>> >>> The impedance looking into the node on the left (marked arrow-style by >>> Larkin) looks like that resistor from the arrow to IN- of the OpAmp >>> (until the OpAmp runs into the GBW stop). >>> >>> The impedance looking into the right arrow is, in essence, a voltage >>> source, zero AC impedance, again until the OpAmp runs into the GBW >>> stop. >>> >>> If that's a _floating_ capacitor, I'm the King of Siam. >>> >>> But Larkin's sickophants should be happy, their low-information >>> mentality, and need for name-calling, has been satisfied. They will >>> now face toward San Fransicko and chant... >:-} >>> >>> ...Jim Thompson >>> >> >>I thought you were trying not to dish that stuff out? >> >>Using a voltage divider on a small compensation cap is a pretty common >>approach in TIAs, though IME it's usually a capacitive divider rather >>than a resistive one. It works OK up to the point where the summing >>junction loading is important. You clearly don't want to try >>synthesizing 1 pF using a 1000:1 voltage divider and a nanofarad in >>series.(*) >> >>Using a pot, you have to worry about the current noise of the low >>resistances getting into the summing junction. (I posted a TIA fragment >>a week or two back where I had to short out a 1G resistor with a relay >>to avoid its Johnson noise current dominating the circuit performance >>above a few hundred hertz.) >> >>However, given that the input capacitance is probably nearly 2 pF to >>begin with, something like 0.5 pF and a low-value pot isn't necessarily >>a silly way of making an adjustable 0.3 pF for that sort of use. > >Actually defining the total problem would have been a more useful >exercise. > >> >>Cheers >> >>Phil Hobbs >> >>(*) Back in the early part of my career, I actually managed to make an >>op amp integrator oscillate all by itself. Fast op amp with high Zout >>and some phase funnies in its compensation, small R_in/large C_F. > >Nowhere in Larkin's multiple posts was "capacitor immersed in an >integrator" mentioned. > >When I show a true capacitor multiplier, I get his usual ration of >shit... and bloviations about "floating". > > ...Jim ThompsonOther people understood what I meant; I want a sub-pF variable cap to trim a circuit, and neither end is grounded. I posted the actual circuit. If you had posted "your" cap multiplier with some neutral comment, like "take a look at this cap multiplier", you'd look somewhat less stupid. Or if the context wasn't clear, you could have asked. But you misunderstood the situation and posted a challenge. How did you get to be so rash and insecure? You're sure not aging gracefully. -- John Larkin Highland Technology Inc www.highlandtechnology.com jlarkin at highlandtechnology dot com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom timing and laser controllers Photonics and fiberoptic TTL data links VME analog, thermocouple, LVDT, synchro, tachometer Multichannel arbitrary waveform generators
