what has even or odd number of inverter to do in ring oscillator

[jp619]

In a ring oscillator, I believe the oscillation takes place due to the delay provided by not gate. pls can anyone tell me the relationship between inverter being in even n odd no. have to do with oscillation... :-?

 

[trav1s]

Does ring oscillator with an even number of inverters really oscillate? Think about it.

 

[jp619]

to be specific, there were 5 inverters and asked what will happen if 6 are connected???? though after 6 output should be as input all i want to know is the role played by delay given by each inverter and how it can be anticipated????

 

[trav1s]

If you chain an even number of inverters together, there will be no oscillation. The inverters will stabilize and hold their value. Try drawing the circuit diagram and trace the signal as it passes down through the inverters and wires and you will understand why. If not, you are lacking basic understanding and should go back and study the basics.

 

[trav1s]

Does this answer your question?

 

[jp619]

plzzz would you explain a bit elaborately from the posted image point of view??

 

[trav1s]

Yeah no problem. Lets name the inverters from left to right: inv1, inv2, inv3 and so on.

We will look at the top circuit first.
If this was a physical circuit you would not need an initial input, as the circuit would self-initialize. However for the sake of completeness lets say the input at t=0 is 1. After t=0 the input is not driven. Follow the signal down the chain.

At t=050ns the output of inv1 is 0
At t=100ns the output of inv2 is 1
At t=150ns the output of inv3 is 0
At t=200ns the output of inv4 is 1
At t=250ns the output of inv5 is 0

Now the output of the ring oscillator has stabilized at 0. This is the same signal as the input to inv1. The input to inv1 was initially 1, but now it is 0. Therefore,

At t=300ns the output of inv1 is 1
At t=350ns the output of inv2 is 0
At t=400ns the output of inv3 is 1
At t=450ns the output of inv4 is 0
At t=500ns the output of inv5 is 1

An observation as to the period of the ring oscillator can now be made. At t=250ns the output of inv5 was 0 and it changed to 1 at t=500ns. Therefore since 250ns is half of the period, 500ns is the full period. In fact, the formula for the period for an arbitrary ring oscillator will be 2x(number of inverters)x(inverter delay).

Now, lets look at the case of 6 inverters. Again say the input at t=0 is 1. After t=0 the input is not driven.

At t=050ns the output of inv1 is 0
At t=100ns the output of inv2 is 1
At t=150ns the output of inv3 is 0
At t=200ns the output of inv4 is 1
At t=250ns the output of inv5 is 0
At t=300ns the output of inv6 is 1
At t=350ns the output of inv1 is still 0

A ring oscillator must have an odd number of inverters. If you have a chain of an even number of inverters it will behave rather like a D Flip-Flop. Can I get some points now?

 

[jp619]

thank trav1s

---------- Post added at 09:03 ---------- Previous post was at 08:57 ----------

hey trav1s, can you hekp me with this thread???
Windows Live Gallery

 

[trav1s]

I think you posted the wrong link, jp.

 

[jp619]

https://www.edaboard.com/threads/203913/#post859436

 

[LvW]

If you chain an even number of inverters together, there will be no oscillation. The inverters will stabilize and hold their value. Try drawing the circuit diagram and trace the signal as it passes down through the inverters and wires and you will understand why. If not, you are lacking basic understanding and should go back and study the basics.

In the analog world there is a similar effect:
A linear oscillator can oscillate only if there is NEGATIVE dc feedback. Otherwise there is no stable operational point.
This can be established only if the number of inverting functions within the loop is odd.