Oscillator using Sallen-Key Low pass filter

Thread Starter

Wai Khang Lim

Joined Oct 2, 2018
7
Hi, I am currently trying to construct an oscillator using a sallen-key low pass filter and is using LTSpice to simulated the circuit. Based on the transfer function of the circuit, I calculate the value of K(gain) which is equals to 3 exactly. The output of the circuit is zero when the input is grounded, and the values of C=C1=C2=10u and R=R1=R2=10k. But, the output of the circuit is zero when i try to simulate.

Any suggestions or advises?
 

Attachments

ericgibbs

Joined Jan 29, 2010
10,516
hi Wai.
Look at this, note the Initial Condition IC=....
E
Basically the circuit requires a little 'disturbance' as would be the case in the 'real' world.


Capture.PNG

EDIT:
This is another option.

Capture2.PNG
 
Last edited:

LvW

Joined Jun 13, 2013
1,078
Hi Wai - for a safe start of oscillations you need a gain whichis slightly larger than "3" (for example 3.1) and - as mentioned by Eric - a kind of asymmetry (an initial condition/preloading for one of the capacitors) or a kind of kick-off pulse at t=0.
 

ericgibbs

Joined Jan 29, 2010
10,516
hi Wai,
In LTSpice the set Initial Condition is a useful tool, check Help/Search for .IC.
The conditions at a simulation start up in LTS are ideal and the circuit is stable and quiet often not self start oscillating.
By setting an Initial Condition voltage at a Node in the circuit, the circuit is not 'balanced/stable' at power up and this is usually sufficient to start the circuit oscillating.

E
Check this link.
https://www.google.com/search?client=firefox-b-d&channel=crow&q=examples+of+.IC+in+LTSpice
 

ericgibbs

Joined Jan 29, 2010
10,516
hi,
Why would you want to modify the basic circuit in order to make it oscillate, when there is provision for setting the Initial Condition of a Node voltage.???
Adding components to the circuit will most likely change its performance characteristic as a Sallen.?

Try this test, remove or Comment out the .IC statement, run the Sim and then use the cursor probe to check all the circuit Nodes, they are all 0V.!, that's why it will not self start oscillating,
E
 

LvW

Joined Jun 13, 2013
1,078
So, if i wanted to start the oscillation by using circuitry instead of setting the initial conditions what should i do?
The most realistic method is to use not constant supply voltages but to switch on (at least) one of the two supply voltages at t=0 during TRAN analysis. This is the most secure method to allow the start of oscillations - provided the circuit fulfills the oscillation condition.
 
Top