Amplitude stabilization in Oscillator using Diode ? What is mechanism ?

Thread Starter

Ngô Đình Nhân

Joined Jan 8, 2017
45
Hi everyone
Now I am making a oscillator circuit using Quartz Crystal (Zq). But actually I do not clearly understand about mechanism of amplitude limitation using Schottky diodes. Does any one have experience about this ? Please help me !

You see that the Gain (A) depends on R4, so in my opinion, I think that when the diode on, the Gain (A) will be Zero because that that time, R4 is bypassed. Is this right ?
Image 4.png
 

Bernard

Joined Aug 7, 2008
5,784
I believe it is the non linearity of the diodes. In the far distant past a incandescent bulb was used for same purpose.
 

AnalogKid

Joined Aug 1, 2013
10,986
Now I am making a oscillator circuit using Quartz Crystal (Zq). But actually I do not clearly understand about mechanism of amplitude limitation using Schottky diodes.
You see that the Gain (A) depends on R4, so in my opinion, I think that when the diode on, the Gain (A) will be Zero because that that time, R4 is bypassed. Is this right ?
Nope. Two things.

1. When the diode is "on", it does not completely bypass anything. Whenever a diode is in forward conduction, it has an equivalent series resistance that varies inversely with the forward current. For example, if a small signal diode has 1 mA DC going through it and a Vf of 0.6 V, then the equivalent series resistance is 600 ohms, not 0 ohms. If you increase the current to 2 mA and the (perfect) diode's Vf stays at 0.6 V, the ESR now is 300 ohms. Audio and RF circuits use diodes as on/off shunt switches and current-variable resistors by manipulating this mathematical consequence.

2. The diodes in your schematic are not ideal, so there is no magical 0.6 V Vf brick wall. Forward conduction starts around o.4 V for low forward currents, and increases as the forward current increases. The increase with increasing current tapers off above 0.62 V or so, but for higher currents the Vf easily can hit 0.8 or 0.9 V. So the diodes act as a soft-knee clipper whose clamping voltage varies throughout each sinewave cycle.

So Bernard is correct, it is the diode non-linearities that do the work, but more as a limiter than a gain-stabilizer. The light bulb he mentions is a classic part of the classic Wein Bridge oscillator. In that circuit the light bulb is part of the gain-setting shunt resistance in a amplifier negative feedback loop. Its thermal mass lowpass filters the signal, and its resistance varies directly with the total energy in the waveform (usually a sinewave). As the waveform amplitude increases, the bulb resistance increases, turning down the loop gain. Because the bulb filament is a true resistor, the circuit is capable of very low distortion. However, the amplitude changes when the furnace or air conditioner comes on.

ak
 
Last edited:

ErnieM

Joined Apr 24, 2011
8,377
An oscillator has 3 basic parts: a high gain amplifier, a frequency selector network or device, and some sort of gain limiter. The limit may be part of the amp.

The over all loop gain can be less than, greater than, or equal to one. When less, the oscillations damp out and stop. When greater than one they grow without bound... Or till they saturate the amp in the real world. That signal may be distorted.

Only when the gain is exactly one will the wave be stable and have the lowest distortion. The diodes serve as an amplitude limiter to adjust the gain to exactly one.

Any further discussion requires math I forget decades ago. :)
 

MrAl

Joined Jun 17, 2014
11,389
Hi everyone
Now I am making a oscillator circuit using Quartz Crystal (Zq). But actually I do not clearly understand about mechanism of amplitude limitation using Schottky diodes. Does any one have experience about this ? Please help me !

You see that the Gain (A) depends on R4, so in my opinion, I think that when the diode on, the Gain (A) will be Zero because that that time, R4 is bypassed. Is this right ?
View attachment 119230
Hi,

As others have pointed out, the diodes in this circuit probably provide more of a limiting (by clipping) action than by actual gain adjustment and they have no memory so the clipping would occur on each and every half cycle no matter how long the thing runs.

The circuit is questionable however in that although the caption under the circuit reads "practical realization" there is no apparent initial energy supply for the collector circuit in that drawing. That makes the circuit look like it wont really work exactly the way it is drawn because there is no energy to start the oscillations. It could be that R4 really goes to +Vcc or something like that.

The analysis of these circuits isnt as difficult as it might seem at first glance. It's just a single feedback system and methods for much more complicated systems have been around for decades. We would need to know more about the elements of the circuit though and how it gets it's initial energy for start up. Maybe it depends on a forward biased base collector to start the oscillations :)
 

Sensacell

Joined Jun 19, 2012
3,432
Note that very few practical crystal oscillator circuits require any form of amplitude stabilization, unless the intention is ultra-low distortion.

Usually the amplitude is limited by the voltage compliance of the amplifier stage.
 

ramancini8

Joined Jul 18, 2012
473
In many crystal oscillators the output signal is a pulse or similar non-linear wave shape, and the feedback circuit take the form of a crystal pi filter thus yielding a sinewave.
 

ErnieM

Joined Apr 24, 2011
8,377
Hi,

As others have pointed out, the diodes in this circuit probably provide more of a limiting (by clipping) action than by actual gain adjustment and they have no memory so the clipping would occur on each and every half cycle no matter how long the thing runs.

The circuit is questionable however in that although the caption under the circuit reads "practical realization" there is no apparent initial energy supply for the collector circuit in that drawing. That makes the circuit look like it wont really work exactly the way it is drawn because there is no energy to start the oscillations. It could be that R4 really goes to +Vcc or something like that.

The analysis of these circuits isnt as difficult as it might seem at first glance. It's just a single feedback system and methods for much more complicated systems have been around for decades. We would need to know more about the elements of the circuit though and how it gets it's initial energy for start up. Maybe it depends on a forward biased base collector to start the oscillations :)
I quite agree this is not the print one builds to, it is an AC model of some other circuit.
 
Top