John Larkin's LC oscillator
in [Basics]
|
Prev: Twin T circuit wanted
Next: And blocking oscillators
From: Jim Thompson on 14 Jun 2010 20:45 On Mon, 14 Jun 2010 17:30:55 -0700, Jim Thompson <To-Email-Use-The-Envelope-Icon(a)On-My-Web-Site.com> wrote: >On Mon, 14 Jun 2010 16:36:19 -0700, Jim Thompson ><To-Email-Use-The-Envelope-Icon(a)On-My-Web-Site.com> wrote: > >>On Mon, 14 Jun 2010 16:15:09 -0700, Jim Thompson >><To-Email-Use-The-Envelope-Icon(a)On-My-Web-Site.com> wrote: >> >>>On Mon, 14 Jun 2010 16:01:55 -0700, John Larkin >>><jjlarkin(a)highNOTlandTHIStechnologyPART.com> wrote: >>> >>>>On Mon, 14 Jun 2010 15:27:30 -0700 (PDT), dagmargoodboat(a)yahoo.com >>>>wrote: >>>> >>>>>On Jun 14, 4:41�pm, John Larkin >>>>><jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote: >>>>>> On Mon, 14 Jun 2010 12:27:57 -0700 (PDT), dagmargoodb...(a)yahoo.com >>>>>> wrote: >>>>>> >>>>>> >> Q1 doesn't conduct in reverse with your values, but no >>>>>> >> Schottky... the current is always out of the emitter... though it does >>>>>> >> get awfully close to zero: -376uA and -12.5mA peak. >>>>>> >>>>>> >Huh. �LTSpice says Q1 does conduct in reverse, a nasty little 5mA >>>>>> >spike's worth. �The schottky feedback prevents that by cutting the >>>>>> >base bias enough to make sure the collector never gets that low. >>>>>> >>>>>> Right; the schottly is a more pure AGC mode. The reverse emitter >>>>>> conduction depends on the inverse beta of the transistor. If the model >>>>>> includes inverse beta, when the collector dips down to close to >>>>>> ground, and the c-b junction forward biases, it essentially flips >>>>>> ends: collector becomes emitter, emitter becomes collector, emitter >>>>>> current flows upwards. >>>>>> >>>>>> ftp://jjlarkin.lmi.net/LC_YDx.gif >>>>>> >>>>>> Either way, the base cap gets discharged. >>>>> >>>>>It's interesting that Jim's PSpice doesn't show the same spike. >>>> >>>>He blames it on defects in LT Spice. That's crazy. I suspect it's >>>>actually different transistor models. Note that in my sim, the emitter >>>>current reverses exactly at the points in time when Vc swings through >>>>Ve. >>>> >>>>John >>>> >>> >>>I didn't say that at all. I said I don't see it in PSpice, AND it's >>>an advertised feature of LTspice that models are tweaked for speed. >>> >>>However I suspect it's that I assigned a resistance to the feedback >>>winding proportionate to the Q assigned to the primary. You two did >>>not. >>> >>>PSpice does model BR, so it's not that. It could also be that I'm >>>looking way out at 1 second, where the loop is steady, and loop >>>"replenishment current" is very small. It's quite possible that, >>>during loop closure, you have some inverse transistor action. It's >>>certainly not there at 1 second... and the spectral analysis does not >>>show it either. >>> >>>However, it is dead clear, there is no AGC action controlling >>>TRANSCONDUCTANCE :-) >>> >>> ...Jim Thompson >> >>I re-measured: IE is _never_ less than 376uA OUT of the emitter... no >>reverse transistor action. >> >> ...Jim Thompson > >On closer examination I _am_ seeing some kind of burble on the emitter >current. Applying a Schottky _does_ reduce the burble, but not >completely, since you load the bias cap. No real change in spectrum. >Perhaps tie base of Q1 to the juncture of R1/D1? Or a Baker clamp >would certainly provide the isolation. But I doubt the ROI :-) > > ...Jim Thompson Baker clamp does the trick without as much rise in peak IE. But, if your definition of class-A is current not passing thru zero, it doesn't matter... it already was "class-A" :-) And your definition of "is" :-) ...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: JosephKK on 15 Jun 2010 01:06 On Mon, 14 Jun 2010 11:15:00 -0700, Jim Thompson <To-Email-Use-The-Envelope-Icon(a)On-My-Web-Site.com> wrote: >On Mon, 14 Jun 2010 10:50:58 -0700 (PDT), dagmargoodboat(a)yahoo.com >wrote: > >>On Jun 14, 12:08 pm, Jim Thompson <To-Email-Use-The-Envelope-I...(a)On- >>My-Web-Site.com> wrote: >>> On Mon, 14 Jun 2010 09:49:09 -0700 (PDT), dagmargoodb...(a)yahoo.com >>> wrote: >[snip] >>> >>> > Vcc = +5v >>> > --+--------------+------+-- >>> > | | | >>> > | | .-. >>> > | | | | e.s.r. = 1 ohm >>> > | | | | >>> > | | '-' >>> > | | | >>> > | | |_ || >>> > | | _)|| >>> > .-. --- L1a _)|| >>> > Rb | | C1 --- 1mH _)|| >>> > 47k | | 1uF | _)|| >>> > '-' | _)|| >>> > | | *| || >>> > | '---+--' || >>> > | R1 D1 | || >>> > | 220r schottky | || >>> > +--/\/\/----|>]----+ || >>> > | | || >>> > | .------------' || >>> > | |/ || >>> > +---| Q1 || >>> > | |>. 2n3904 || >>> > | | * || >>> > C2 --- +----------------. || >>> > 1uF --- | L1b _)|| >>> > | | 100nH _)|| >>> > | | | >>> > === | === >>> > | >>> > '--------------------> 5KHz output >>> >>> >Cheers, >>> >James Arthur >>> >>> [snip LTspice Schematic] >>> >>> James, How can you call it "class-A" when the emitter current is >>> _not_linear_, _not_linear_, _not_linear_, _not_linear_ !! >> >>True, the emitter current isn't linear, but it is continuous and non- >>zero, so I call that "class-A." >> >> >>> All you've done is use a Schottky bypassing the C-B junction... >>> otherwise it's identical. >> >>That makes a big difference. It prevents Q1 saturating and from >>conducting b-c. And, without it Q1 conducts in reverse mode during >>negative peaks, making the i(c) not just discontinuous, but reversing >>in direction. That loads the tank, obviously. > >"Obviously"? Q1 doesn't conduct in reverse with your values, but no >Schottky... the current is always out of the emitter... though it does >get awfully close to zero: -376uA and -12.5mA peak. > >> >>> I liken such oscillators to how you push your kid on a swing set. >>> Giving a "nudge" every cycle. >> >>That's exactly how I think of them too. A "nudge" each cycle (short >>conduction cycle) is what I did first. I call that "class-C." >> >>> To be "class-A" you'd need to sit upon the top bars and _continuously_ >>> push and pull the ropes. >> >>Or you can pull continuously, harder at some times, not as hard at >>others. That gives the swing position a d.c. bias, but is otherwise >>the same, yes? > >Class-A implies _linear_, does it not? Or do we have a Larkin >definition for today ?:-) > The definition i have always heard is that it is conducting through the whole cycle. Nothing about linearity. And that is one fougly non-linear circuit.
From: John Larkin on 15 Jun 2010 01:18 On Mon, 14 Jun 2010 22:06:53 -0700, "JosephKK"<quiettechblue(a)yahoo.com> wrote: >On Mon, 14 Jun 2010 11:15:00 -0700, Jim Thompson ><To-Email-Use-The-Envelope-Icon(a)On-My-Web-Site.com> wrote: > >>On Mon, 14 Jun 2010 10:50:58 -0700 (PDT), dagmargoodboat(a)yahoo.com >>wrote: >> >>>On Jun 14, 12:08�pm, Jim Thompson <To-Email-Use-The-Envelope-I...(a)On- >>>My-Web-Site.com> wrote: >>>> On Mon, 14 Jun 2010 09:49:09 -0700 (PDT), dagmargoodb...(a)yahoo.com >>>> wrote: >>[snip] >>>> >>>> > � � � � Vcc = +5v >>>> > � �--+--------------+------+-- >>>> > � � �| � � � � � � �| � � �| >>>> > � � �| � � � � � � �| � � .-. >>>> > � � �| � � � � � � �| � � | | e.s.r. = 1 ohm >>>> > � � �| � � � � � � �| � � | | >>>> > � � �| � � � � � � �| � � '-' >>>> > � � �| � � � � � � �| � � �| >>>> > � � �| � � � � � � �| � � �|_ || >>>> > � � �| � � � � � � �| � � � _)|| >>>> > � � .-. � � � � � �--- �L1a _)|| >>>> > Rb �| | � � � � C1 --- �1mH _)|| >>>> > 47k | | � � � � 1uF | � � � _)|| >>>> > � � '-' � � � � � � | � � � _)|| >>>> > � � �| � � � � � � �| � � *| �|| >>>> > � � �| � � � � � � �'---+--' �|| >>>> > � � �| � �R1 � � D1 � � | � � || >>>> > � � �| � 220r �schottky | � � || >>>> > � � �+--/\/\/----|>]----+ � � || >>>> > � � �| � � � � � � � � �| � � || >>>> > � � �| � � .------------' � � || >>>> > � � �| � |/ � � � � � � � � � || >>>> > � � �+---| � �Q1 � � � � � � �|| >>>> > � � �| � |>. 2n3904 � � � � � || >>>> > � � �| � � | � � � � � � � * �|| >>>> > �C2 --- � �+----------------. || >>>> > 1uF --- � �| � � � � � L1b �_)|| >>>> > � � �| � � | � � � � 100nH �_)|| >>>> > � � �| � � | � � � � � � � | >>>> > � � === � �| � � � � � � �=== >>>> > � � � � � �| >>>> > � � � � � �'--------------------> 5KHz output >>>> >>>> >Cheers, >>>> >James Arthur >>>> >>>> [snip LTspice Schematic] >>>> >>>> James, �How can you call it "class-A" when the emitter current is >>>> _not_linear_, _not_linear_, _not_linear_, _not_linear_ !! >>> >>>True, the emitter current isn't linear, but it is continuous and non- >>>zero, so I call that "class-A." >>> >>> >>>> All you've done is use a Schottky bypassing the C-B junction... >>>> otherwise it's identical. >>> >>>That makes a big difference. It prevents Q1 saturating and from >>>conducting b-c. And, without it Q1 conducts in reverse mode during >>>negative peaks, making the i(c) not just discontinuous, but reversing >>>in direction. That loads the tank, obviously. >> >>"Obviously"? Q1 doesn't conduct in reverse with your values, but no >>Schottky... the current is always out of the emitter... though it does >>get awfully close to zero: -376uA and -12.5mA peak. >> >>> >>>> I liken such oscillators to how you push your kid on a swing set. >>>> Giving a "nudge" every cycle. >>> >>>That's exactly how I think of them too. A "nudge" each cycle (short >>>conduction cycle) is what I did first. I call that "class-C." >>> >>>> To be "class-A" you'd need to sit upon the top bars and _continuously_ >>>> push and pull the ropes. >>> >>>Or you can pull continuously, harder at some times, not as hard at >>>others. That gives the swing position a d.c. bias, but is otherwise >>>the same, yes? >> >>Class-A implies _linear_, does it not? Or do we have a Larkin >>definition for today ?:-) >> > >The definition i have always heard is that it is conducting through the >whole cycle. Nothing about linearity. And that is one fougly non-linear >circuit. Only if you want it to be. It can be designed to be a nice smooth class A oscillator with precise automatic gain control that servoes oscillation amplitude to almost exactly 2*Vcc, with a low TC. Not bad for one transistor. John
From: Jim Thompson on 15 Jun 2010 10:21 On Mon, 14 Jun 2010 22:18:22 -0700, John Larkin <jjlarkin(a)highNOTlandTHIStechnologyPART.com> wrote: >On Mon, 14 Jun 2010 22:06:53 -0700, >"JosephKK"<quiettechblue(a)yahoo.com> wrote: > >>On Mon, 14 Jun 2010 11:15:00 -0700, Jim Thompson >><To-Email-Use-The-Envelope-Icon(a)On-My-Web-Site.com> wrote: >> >>>On Mon, 14 Jun 2010 10:50:58 -0700 (PDT), dagmargoodboat(a)yahoo.com >>>wrote: >>> >>>>On Jun 14, 12:08�pm, Jim Thompson <To-Email-Use-The-Envelope-I...(a)On- >>>>My-Web-Site.com> wrote: >>>>> On Mon, 14 Jun 2010 09:49:09 -0700 (PDT), dagmargoodb...(a)yahoo.com >>>>> wrote: >>>[snip] >>>>> >>>>> > � � � � Vcc = +5v >>>>> > � �--+--------------+------+-- >>>>> > � � �| � � � � � � �| � � �| >>>>> > � � �| � � � � � � �| � � .-. >>>>> > � � �| � � � � � � �| � � | | e.s.r. = 1 ohm >>>>> > � � �| � � � � � � �| � � | | >>>>> > � � �| � � � � � � �| � � '-' >>>>> > � � �| � � � � � � �| � � �| >>>>> > � � �| � � � � � � �| � � �|_ || >>>>> > � � �| � � � � � � �| � � � _)|| >>>>> > � � .-. � � � � � �--- �L1a _)|| >>>>> > Rb �| | � � � � C1 --- �1mH _)|| >>>>> > 47k | | � � � � 1uF | � � � _)|| >>>>> > � � '-' � � � � � � | � � � _)|| >>>>> > � � �| � � � � � � �| � � *| �|| >>>>> > � � �| � � � � � � �'---+--' �|| >>>>> > � � �| � �R1 � � D1 � � | � � || >>>>> > � � �| � 220r �schottky | � � || >>>>> > � � �+--/\/\/----|>]----+ � � || >>>>> > � � �| � � � � � � � � �| � � || >>>>> > � � �| � � .------------' � � || >>>>> > � � �| � |/ � � � � � � � � � || >>>>> > � � �+---| � �Q1 � � � � � � �|| >>>>> > � � �| � |>. 2n3904 � � � � � || >>>>> > � � �| � � | � � � � � � � * �|| >>>>> > �C2 --- � �+----------------. || >>>>> > 1uF --- � �| � � � � � L1b �_)|| >>>>> > � � �| � � | � � � � 100nH �_)|| >>>>> > � � �| � � | � � � � � � � | >>>>> > � � === � �| � � � � � � �=== >>>>> > � � � � � �| >>>>> > � � � � � �'--------------------> 5KHz output >>>>> >>>>> >Cheers, >>>>> >James Arthur >>>>> >>>>> [snip LTspice Schematic] >>>>> >>>>> James, �How can you call it "class-A" when the emitter current is >>>>> _not_linear_, _not_linear_, _not_linear_, _not_linear_ !! >>>> >>>>True, the emitter current isn't linear, but it is continuous and non- >>>>zero, so I call that "class-A." >>>> >>>> >>>>> All you've done is use a Schottky bypassing the C-B junction... >>>>> otherwise it's identical. >>>> >>>>That makes a big difference. It prevents Q1 saturating and from >>>>conducting b-c. And, without it Q1 conducts in reverse mode during >>>>negative peaks, making the i(c) not just discontinuous, but reversing >>>>in direction. That loads the tank, obviously. >>> >>>"Obviously"? Q1 doesn't conduct in reverse with your values, but no >>>Schottky... the current is always out of the emitter... though it does >>>get awfully close to zero: -376uA and -12.5mA peak. >>> >>>> >>>>> I liken such oscillators to how you push your kid on a swing set. >>>>> Giving a "nudge" every cycle. >>>> >>>>That's exactly how I think of them too. A "nudge" each cycle (short >>>>conduction cycle) is what I did first. I call that "class-C." >>>> >>>>> To be "class-A" you'd need to sit upon the top bars and _continuously_ >>>>> push and pull the ropes. >>>> >>>>Or you can pull continuously, harder at some times, not as hard at >>>>others. That gives the swing position a d.c. bias, but is otherwise >>>>the same, yes? >>> >>>Class-A implies _linear_, does it not? Or do we have a Larkin >>>definition for today ?:-) >>> >> >>The definition i have always heard is that it is conducting through the >>whole cycle. Nothing about linearity. And that is one fougly non-linear >>circuit. > >Only if you want it to be. It can be designed to be a nice smooth >class A oscillator with precise automatic gain control that servoes >oscillation amplitude to almost exactly 2*Vcc, with a low TC. Not bad >for one transistor. > >John So show us. I maintain there is no variable _linear_gain_ element there. As Win says, "Discuss it" :-) ...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: JosephKK on 15 Jun 2010 23:36
On Mon, 14 Jun 2010 22:18:22 -0700, John Larkin <jjlarkin(a)highNOTlandTHIStechnologyPART.com> wrote: >On Mon, 14 Jun 2010 22:06:53 -0700, >"JosephKK"<quiettechblue(a)yahoo.com> wrote: > >>On Mon, 14 Jun 2010 11:15:00 -0700, Jim Thompson >><To-Email-Use-The-Envelope-Icon(a)On-My-Web-Site.com> wrote: >> >>>On Mon, 14 Jun 2010 10:50:58 -0700 (PDT), dagmargoodboat(a)yahoo.com >>>wrote: >>> <snip> >>>Class-A implies _linear_, does it not? Or do we have a Larkin >>>definition for today ?:-) >>> >> >>The definition i have always heard is that it is conducting through the >>whole cycle. Nothing about linearity. And that is one fougly non-linear >>circuit. > >Only if you want it to be. It can be designed to be a nice smooth >class A oscillator with precise automatic gain control that servoes >oscillation amplitude to almost exactly 2*Vcc, with a low TC. Not bad >for one transistor. > >John So post that version instead. |