Thanks for letting me know. Note I also supply the PLT files so you can see the actual results. You might try downloading LTspice to see the actual waveforms and compare them to what you get in HSpice. That would be an interesting comparison.
You might also be interested in my Delay Discriminator article at
This gives new insights into the operation of the double balanced mixer as well as various quadrature phase detectors.
I have used your design rules in a simple C program to calculate the component values for a 150 MHz common emitter Colpitts oscillator. This C program can be used in interactive or command line argument mode, and the input and output are as follows. BJT minimum beta|hfe
25 BJT collector current(mA) 40 load and/or source resistance(Ohm) 50 DC supply voltage(V) 15 oscillation frequency(MHz) 150 loaded quality factor Ql 40 collector resistance Rc 187.500000 Ohm emitter resistance 56.250000 Ohm base bias resistance RB1 50.986842 K Ohm base bias resistance RB2 97.039474 K Ohm emitter signal resistance - re - 0.625000 Ohm emitter bypass capacitor 1.887238e-10 F input coupling capacitor 7.318724e-11 F output coupling capacitor 2.689314e-11 F Colpitts series inductor 5.307856e-08 H Colpitts shunt capacitor 1 4.246285e-11 F Colpitts shunt capacitor 2 4.246285e-11 FThe SPICE netlist is as follows:
.PARAMS C0=4.246e-11 L0=5.308e-08 AMPL1=15 .PARAMS CBYP=1.887e-10 CIN=7.319e-11 RB0=50.99K
.SUBCKT ampcomemitter 1 2 3 7
** 1 VCC ** 2 IN ** 3 OUT ** 7 GND|REF FOR CURRENT CONSERVATION C0 2 4 {CIN} C2 6 7 {CBYP} Q0 3 4 6 2N5179 RB0 1 4 {RB0} RB1 4 7 {RB1} RC 1 3 {RC} RE 6 7 {RE} .ENDS.SUBCKT PIN 1 2 3
** 1 IN ** 2 OUT ** 3 GND|REF FOR CURRENT CONSERVATION C0 1 3 {C0} C1 2 3 {C0} L0 1 2 {L0} .ENDSVCC 1 0 DC {AMPL1} AC 0.0 C0 3 4 {COUT} R1 4 0 {RL} XCEA 1 2 3 0 ampcomemitter XPI 4 2 0 PIN
.OPTIONS METHOD=GEAR NOPAGE RELTOL=1m MINBREAK=5ps
** COLPITTS OSCILLATOR .IC V(2)=10.0 V(3)=8.0 V(4)=9 .TRAN 0.25ns 2.0ms 1.0ms UIC .PRINT TRAN V(4) .ENDWith the 50.0 Ohm external load, the output signal amplitude is in the range of a 10^-12 V. Howevwer, when the external output amplitude is in the range of +/- %V, which is meaningful amd measureable.
The 50 ohms refers to XL. You cannot put a 50 ohm load on Vout.
Why not install LTspice. That way you can see the actual schematic and waveforms, plus get an advanced SPICE engine. For example see Mike Engelhardt's Spice Differentiation at
You already have two SPICE versions installed. Adding one more will let you cover all the bases.
Heh. Playing Nethack was not waste of time.
Cutoff is where I_C goes to zero. Saturation is where V_CE goes to zero (or nearly). In both cases, the gain of the stage is zero, so either or both can be the amplitude limiting mechanism in an oscillator.
Saturation is slow and jittery, whereas cutoff is fast and clean.
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510 http://electrooptical.net http://hobbs-eo.com
The transistor is cut off for most of the cycle. It only conducts at the positive peak of the base waveform, so you cannot use emitter-base cutoff to limit the amplitude. The only way to adjust the amplitude is to change the emitter resistor. I specifically mention avoiding saturation or excessive base-emitter reverse bias in the notes:
01.ASC Classic Colpitts ~~~~~~~~~~~~~~~~~~~~~~~ This is the simplest and easiest to get working. Here are the steps:Run the LTspice model.
NOTE: In all these oscillators, you should monitor the signal on the base to ensure it doesn't approach VCC, and check the base-emitter voltage at the negative peak to make sure it doesn't exceed the reverse breakdown voltage of the transistor.
You know, Steve, you write a lot better than you read.
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510 http://electrooptical.net http://hobbs-eo.com
There is nothing wrong with my reading. Maybe you would like to download Oscillator.zip, run 01.ASC, and look at the actual waveforms.
Then tell me how you can adjust the amplitude by changing the base-emitter cutoff.
You don't _change_ the base-emitter cutoff, but cutoff is the physical mechanism that limits oscillation amplitude.
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510 http://electrooptical.net http://hobbs-eo.com
V_be does not "cutoff" according to Phil. It does, according to you. Who is right? Can we define terms agreeably, please?
I'm not taking sides, I'm just pointing out that either you did not read well, or did not bother to respond to what you read.
I'm interested in the outcome of the discussion, so it's not great that one or both of you seems to repeatedly miss the other's point.
Clifford Heath.
Cutoff has little or nothing to do with the oscillator amplitude. The transistor is off most of the cycle.
The amplitude is determined by the energy fed into the tank, which is dissipated by the losses, specifically the inductor ESR. Keeping everything else constant, the emitter resistor is the only way to change the energy.
Notice I do not use the phrase gm. I think transistor transconductance has little to do with the energy fed to the tank. Most of the time, gm is zero.
I think the issue has more to do with the charge drained from the tank capacitors through R1 during most of the cycle. This is restored at the brief positive peak when the transistor turns on hard and acts as an emitter follower with positive feedback. This dumps energy back into the tank and continues the oscillation.
The reason the transistor turns on hard is because the base-emitter forward bias voltage is greater when more charge is drained by R1. Changing the emitter resistor R1 changes the amount of energy drained from the tank. With a higher value, less energy is drained and the oscillator amplitude decreases, and vice versa.
All good stuff, but entirely beside the point.
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510 http://electrooptical.net http://hobbs-eo.com
Yup. This topology (Colpitts, constant current LC biased) works very nicely. It doesn't squegg at these or nearby values (whereas RC bias tends to). Proof that it's not saturation limited: the amplitude is freely adjustable by base voltage.
Tim
-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Design Website: https://www.seventransistorlabs.com/
I don't know how you change the amplitude, but that's how I do it.
Yes. Also beside the point, which is _the physical mechanism of amplitude limiting_.
If we make the approximation that the voltage gain of the stage is constant when the transistor is conducting, the conduction angle will be
theta_C = 2 pi/A_V.
This is because in order to have a stable amplitude, the average gain has to be 1.000.
But this can happen either by collector saturation or by emitter-base cutoff, or both. Cutoff is better.
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510 http://electrooptical.net http://hobbs-eo.com
As you state above, the transistor is cut off for most of the cycle. When it is cut off, it has no effect on the amplitude. The cutoff occurs when the base-emitter voltage drops to zero and goes negative. This is controlled by tha amplitude of oscillation. So you change the amplitude to ahcnge the cutoff, not the other way around.
The physical mechanism of amplitude limiting is draining charge from the feedback capacitors, which occurs throughout the cycle by means of the emitter resistor.
This changes the amount of energy fed back into the tank when the transistor turns on.
The amount of energy delivered exactly balances the amount lost, so the amplitude remains constant.
To change the amplitude, you have to change the emitter resistor or the base bias voltage.
The Griffith amplitude stabilization method described in my article changes the base bias voltage. He uses the forward voltage across a LED as a reference.
Ulrich Rohde describes another method that measures the voltage on the emitter and changes the base bias voltage to stabilize the amplitude.
In both of these cases, the amplitude has to be within the control range of the feedback.
This is accomplished by adjusting the emitter resistor to change the amplitude.
This is my last post on this topic.
Oh....?
That's quite a bold claim. By "better" what measurable performance characteristics are you claiming are better?
For example, the key most important characteristics in a precession oscillator are typically phase noise, supply current, supply and temperature stability, with the same design that in production might, for example, be used over a 10MHz to 60MHz frequency range and run on, say a
2.5V supply, or less.-- Kevin Aylward
This is usually the case, but not always.
In general, this is not true for all feedback loops though. Its quite possible to have a multiple of 360 Degs, with gain greater than one, and still have the system fully stable. i.e. does not oscillate. Technically, such a system is called "conditionally stable". This somewhat non-intuitive condition is what led to the Nyquist stability criterion, to wit, a system is unstable if there is a pole in the right hand plane. This can be determined without finding the roots of the loop gain transfer function by determining if a polar plot of gain and phase encloses the -1 point.
In practice, a rule that works pretty much all the time, is simple ensuring at at the last time the phase goes below N.360 that the unity gain point is crossed at a 20dB/dec slope, i.e. a single order rate.
Sure, if one wants trivial performance. If low noise, low current and good temperature, power supply and ageing stability are required, it is another matter entirely. Indeed, the hardest part noise wise, is squaring up the oscillator in a limiter.
-- Kevin Aylward
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.