Reply by March 16, 20152015-03-16
```On Sunday, 15 March 2015 04:36:52 UTC+11, Phil Hobbs  wrote:
> On 3/14/2015 1:07 PM, Jim Thompson wrote:
> > Oscillator Proposal...
> >
> > Suppose I have a series RLC, one end grounded, the other end driven by
> > a chip, how might I make that into an oscillator?
> >
> > All wild ideas accepted... this is for a custom chip.
> >
> >                                          ...Jim Thompson
> >
>
> Series RLC is harder than parallel, because to leading order there's no
> signal on the pin at resonance.  (If it were parallel, you could do a
> single-ended version of the MC1648, which you may have heard of.) ;)

The "leading order" is for zero R, so there's always a voltage to be observe in a real circuit. The catch is that the resonant current all flows through Jim's single pin, and the current flow has to be bi-directional because of the capacitor in the RLC.

> You need to arrange a negative resistance at the pin, which isn't hard
> to do, and use the AC current to drive the next stage.

What you often need to do is to make the negative resistance voltage-controlled, so that you can use it to control the amplitude of the oscillation - hard limiting can do the same job, but introduces higher harmonics.

In my application I found it useful to put a small gain advance into the negative resistance current source - propagation delay will always add some phase retardation. In my application some 4.5 degrees of phase advance gave the minimum sustaining current, and pushing it up to 45 degrees didn't increase it much. Reducing below 4.5 degrees of advance pushed up the sustaining current quite rapidly.

As Baxandall's 1959 paper pointed out Class C oscillators are essentially excited by a string of Dirac spikes, so you've got all the harmonics up to the limit set by the speed of your switch. Baxandall's class-D replaces the string of spikes by a square wave, where the amplitude of the odd harmonic content drops in proportion to the harmonic number.

Wien bridge oscillators don't have a high Q tank, so they do go in for more sophisticated gain control arrangements.

The gain control in my low distortion oscillator looked rather like the gain control in one of Jim William's lower distortion Wien Bridges.

--
Bill Sloman, Sydney
```
Reply by March 16, 20152015-03-16
```See, If he allows use of BOTH VCC and Ground, Jim can make a negative resistance device and use a LC with a Tunnel Diode, Lambda Diode, or PUT structure on chip.

The Lambda Diode or PUT can be made with a NPN/PNP or FET pair.

Steve
```
Reply by March 16, 20152015-03-16
```Can we use a two component tank with 3 nodes, VCC, GND, and a center tapped LC or RC or RLC? The center node is the one remaining pin.

Steve
```
Reply by March 15, 20152015-03-15
```"RobertMacy" <robert.a.macy@gmail.com> wrote in message
news:op.xvjc912c2cx0wh@ajm...
> OpAmp with NOISE into the non-inverting input. Rfdbk very high Z and a
> single external connection to the RLC to GND. Does that produce a
> 'narrowband'waveform that 'looks' like an oscillator's output? Hmmm.
> time  to apply .tranoise and see what THAT simulation shows.
>
> External connections: Vcc, GND, RLC, Output? if interest is only in the
> currents in L, then don't need Output.

It looks just as one would expect: literally worst case phase and
amplitude noise.  I've done it before with an IF strip.  Turn up the gain,
and thermal noise is amplified and selected.  Turn it up further and
inevitable feedback tightens the response into narrower bandwidth, while
clipping/limiting destroys amplitude noise (but leaves phase noise intact,
aside from the regenerative mechanism).

Tim

--
Seven Transistor Labs
Electrical Engineering Consultation
Website: http://seventransistorlabs.com

```
Reply by March 15, 20152015-03-15
```On Sat, 14 Mar 2015 10:07:50 -0700, Jim Thompson
<To-Email-Use-The-Envelope-Icon@on-my-web-site.com> wrote:

> Oscillator Proposal...
>
> Suppose I have a series RLC, one end grounded, the other end driven by
> a chip, how might I make that into an oscillator?
>
> All wild ideas accepted... this is for a custom chip.
>
>                                         ...Jim Thompson

ok, ok, putting noise through the narrowband passband was NOT a good idea
for making a 'clean' oscillator source.

So use the fact that there's a lot of 'logic' potential inside a chip, but
external connections cost a fortune...what's wrong with a set of two
variable, switchable current sources? One source, one sink to GND driving
the RLC node? The idea is to start out with very small current ramp up
until reach voltage near rail, switch OFF and turn on he other current
source, and head back down, near GND [or neg rail] reverse etc. Now
increase the current and keep doing until the voltage turns around on its
own BEFORE it reaches the voltage switch points. Make up an appropriate
rule.

maybe reverse process starrt at high current and come down in current. But
with enough stuff inside the chip, and coming up with appropriate rules,
you should be able to make a nice clean, stable output based upon looking
like a Function Generator output that is a bit independentof the Q of the
external tank.

```
Reply by March 15, 20152015-03-15
```On Sun, 15 Mar 2015 04:09:26 -0700, RobertMacy <robert.a.macy@gmail.com>
wrote:

>> ...snip....
> Didn't say what frequency, assumed audio range. To keep under the
> interest of the world's regulatory agencies. less than 10kHz completely
> ignored and less than 150kHz, somewhat ignored.
>
> OpAmp with NOISE into the non-inverting input. Rfdbk very high Z and a
> single external connection to the RLC to GND. Does that produce a
> 'narrowband'waveform that 'looks' like an oscillator's output? Hmmm.
> time to apply .tranoise and see what THAT simulation shows.
>
> External connections: Vcc, GND, RLC, Output? if interest is only in the
> currents in L, then don't need Output.
>

yecch!!!! as promised I tried a .tranoise analysis and got the worst
looking output of an oscillator I've ever seen!

The output waveform looks ratty beyond belief. However the FFT looks just
like it should - Bandpass for noise.

Watch out for the wrap in the very long line, straighten that out to run
the LTspice .asc [I know i has +/-V connection but I got lazy.]

SimpleOscillator.asc

Version 4
SHEET 1 27404 680
WIRE 1120 -640 1088 -640
WIRE 1136 -640 1120 -640
WIRE 1248 -640 1216 -640
WIRE 1280 -640 1248 -640
WIRE 1408 -640 1360 -640
WIRE 1552 -640 1408 -640
WIRE 1616 -592 1584 -592
WIRE 1712 -592 1680 -592
WIRE 1248 -576 1248 -640
WIRE 1376 -512 1120 -512
WIRE 1584 -512 1584 -592
WIRE 1584 -512 1376 -512
WIRE 1616 -512 1584 -512
WIRE 1712 -512 1712 -592
WIRE 1712 -512 1696 -512
WIRE 1744 -512 1712 -512
WIRE 1552 -480 1552 -640
WIRE 1376 -464 1376 -512
WIRE 1520 -464 1376 -464
WIRE 1744 -448 1744 -512
WIRE 1744 -448 1584 -448
WIRE 1808 -448 1744 -448
WIRE 1824 -448 1808 -448
WIRE 1120 -432 1120 -512
WIRE 1376 -432 1280 -432
WIRE 1520 -432 1376 -432
WIRE 1280 -400 1280 -432
WIRE 1376 -400 1376 -432
WIRE 1552 -368 1552 -416
WIRE 1600 -368 1552 -368
WIRE 1616 -368 1600 -368
WIRE 1744 -352 1744 -448
WIRE 1856 -352 1744 -352
WIRE 1120 -320 1120 -352
WIRE 1744 -320 1744 -352
WIRE 1856 -320 1856 -352
WIRE 1376 -304 1376 -320
WIRE 1120 -208 1120 -240
WIRE 1376 -208 1376 -224
WIRE 1120 -112 1120 -144
WIRE 1280 -112 1280 -336
WIRE 1376 -112 1376 -128
WIRE 1744 -112 1744 -240
WIRE 1856 -112 1856 -256
FLAG 1808 -448 out
FLAG 1248 -576 0
FLAG 1120 -112 0
FLAG 1376 -112 0
FLAG 1744 -112 0
FLAG 1856 -112 0
FLAG 1408 -640 cc
FLAG 1120 -640 ss
FLAG 1600 -368 ss
FLAG 1280 -112 0
SYMBOL Opamps\\LT1115 1552 -512 R0
SYMATTR InstName U1
SYMBOL Misc\\battery 1376 -640 R90
WINDOW 0 -32 56 VBottom 2
WINDOW 3 32 56 VTop 2
WINDOW 123 0 0 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName Vcc
SYMATTR Value 9Vdc
SYMBOL Misc\\battery 1232 -640 R90
WINDOW 0 -32 56 VBottom 2
WINDOW 3 32 56 VTop 2
WINDOW 123 0 0 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName Vss
SYMATTR Value 9Vdc
SYMBOL res 1360 -416 R0
SYMATTR InstName Rbias
SYMATTR Value {R}
SYMBOL res 1712 -528 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName Rfdbk
SYMATTR Value {R}
SYMBOL res 1104 -448 R0
SYMATTR InstName Rcoil
SYMATTR Value 2
SYMBOL ind 1104 -336 R0
SYMATTR InstName Lres
SYMATTR Value 1mH
SYMBOL cap 1104 -208 R0
SYMATTR InstName Cres
SYMATTR Value 0.253&#2013266101;F
SYMBOL res 1728 -336 R0
SYMATTR InstName Rload
SYMATTR Value 10k
SYMBOL cap 1840 -320 R0
SYMATTR InstName Cload
SYMATTR Value 10pF
SYMBOL voltage 1376 -320 R0
WINDOW 123 24 44 Left 2
WINDOW 39 0 0 Left 2
SYMATTR Value2 AC 1mV
SYMATTR InstName Vnoise
SYMATTR Value ""
SYMBOL bv 1376 -224 R0
WINDOW 3 24 96 Invisible 2

SYMATTR Value
V={0.00001*sqrt(12/5/dt)*(rand(time/dt)+rand(tstop/dt+1+2*int(time/dt))
+rand(3*(tstop/dt+1)+3*int(time/dt))+rand(6*(tstop/dt+1)+4*int(time/dt))
+rand(10*(tstop/dt+1)+5*int(time/dt))-2.5) }

SYMATTR InstName Bnoise01
SYMBOL cap 1616 -576 R270
WINDOW 0 32 32 VTop 2
WINDOW 3 0 32 VBottom 2
SYMATTR InstName Cfdbk
SYMATTR Value 1000pF
SYMBOL cap 1264 -400 R0
SYMATTR InstName Cbias
SYMATTR Value 1nF
TEXT 544 -520 Left 2 !.options  plotwinsize=0\n.param  k=1.38e-23
T=300\n.param  N=200000  dt=5uS
TEXT 536 -280 Left 2 !.param  fstart={1/N/dt}  fstop={1/2/dt}\n.param
tstart={0.1*N*dt}  tstop={1.1*N*dt}
TEXT 544 -360 Left 2 ;.tran 0 {tstop} {tstart} {dt/10}\n.save  V(out)
TEXT 544 -416 Left 2 ;.ac dec 100 1kHz 10MEG
TEXT 1496 -328 Left 2 !.param R=20k
TEXT 584 -184 Left 2 !.tran 0 100m0 {dt/10}

```
Reply by March 15, 20152015-03-15
```On Sat, 14 Mar 2015 14:08:53 -0700, <meow2222@care2.com> wrote:

> On Saturday, March 14, 2015 at 8:04:30 PM UTC, Robert Macy wrote:
>> On Sat, 14 Mar 2015 10:07:50 -0700, Jim Thompson
>> <To-Email-Use-The-Envelope-Icon@on-my-web-site.com> wrote:
>>
>> > Oscillator Proposal...
>> >
>> > Suppose I have a series RLC, one end grounded, the other end driven by
>> > a chip, how might I make that into an oscillator?
>> >
>> > All wild ideas accepted... this is for a custom chip.
>> >
>> >                                         ...Jim Thompson
>>
>> Hmmm. RLC is a 'short' to GND  How about make the circuitry so noisy
>> that
>> the RLC shorts EVERYTHING BUT the noise and the circuit just 'looks'
>> like
>> it oscillates? I mean narrowband noise is an oscillator output, right?
>
> A short translates to high gain at resonance in a simple opamp with
> feedback. Circuit noise gets it all going.
>
> Jim didn't mention which was wanted, square or sine etc, nor which most
> important properties were wanted. Which make it a very open question.
> Cos there's no shortage of ways.... eg pass mains thru neon, apply to
> your RLC. All depends what you want.
>
>
> NT

Didn't say what frequency, assumed audio range. To keep under the interest
of the world's regulatory agencies. less than 10kHz completely ignored and
less than 150kHz, somewhat ignored.

OpAmp with NOISE into the non-inverting input. Rfdbk very high Z and a
single external connection to the RLC to GND. Does that produce a
'narrowband'waveform that 'looks' like an oscillator's output? Hmmm. time
to apply .tranoise and see what THAT simulation shows.

External connections: Vcc, GND, RLC, Output? if interest is only in the
currents in L, then don't need Output.

```
Reply by March 15, 20152015-03-15
```Jim Thompson wrote:
> Oscillator Proposal...
>
> Suppose I have a series RLC, one end grounded, the other end driven by
> a chip, how might I make that into an oscillator?
>
> All wild ideas accepted... this is for a custom chip.
>
>                                          ...Jim Thompson
1) Ground resistor, feedback is at the top of the resistor, chip drive
at top of LCR; micro does the dirty work to make this stupidity work.
2) Try opamp, positive feedback output to NI via the R, with LC from NI
to gnd; standard resistive FB OP to INV; fiddle on values.

```
Reply by March 14, 20152015-03-14
```On 15/03/2015 4:07 AM, Jim Thompson wrote:
> Oscillator Proposal...
>
> Suppose I have a series RLC, one end grounded, the other end driven by
> a chip, how might I make that into an oscillator?
>
> All wild ideas accepted... this is for a custom chip.

It could be the serial version of the Baxandall Class-D oscillator.

http://sophia-elektronica.com/0344_001_Baxandal.pdf

Tony Williams and I discussed this here some years ago - before Tony
died. Driving a tank circuit with a square wave via an inductor implies
peak voltage that go outside the rails by pi/2.

You can cut the amplitude (and reduce the third harmonic content of the
excitation current to zero with a perfect drive) by contriving to make
the drive voltage a "modified sine wave" which is to say a three-level
rectangular wave that sits at 0V for a third of the time - the drive is
low for one third of the period, centred of one sixth of the period,
high for one third of the period and again centred for the last sixth of
the period.

Don Lancaster's "magic sines" take the idea further.

--
Bill Sloman, Sydney

```
Reply by March 14, 20152015-03-14
```>"I'm pondering.  I need a high circulating current in the inductor"

Why ?
```