OP amp stability gain < 1

[crutschow]

..................................
It has its high frequency response reduced with an internal capacitor so that at a frequency where the phase shift would cause oscillation, the gain is less than 1.
...................................

Perhaps this is a nit but, to perhaps clarify, the "gain is less than 1" refers to the open loop gain of the op amp, not the closed loop gain.

 

[LvW]

About the plane gain or frequency response in AC (less than 3dB) I think it is very important because if you need an unit gain amplifier and it does not have linear gain along the work frequency spectrum, output level could fluctuate between a half or twice input level and it must to be once.

Juan - you should not mix "linearity" with "constant".
Each unity-gain amplifier will show a gain drop above a certain frequency - that is no "fluctuation". It is just a decrease for higher frequencies caused by capacitive influences.

 

[chemelec]

Stability of Op Amps can also be affected by Lack of Bypass Caps on these Power Supply Pins.

In a Single supply Op-Amp, Put a .1uf Cap Directly between the Positive and Ground Pins.
In Dual Supply Op-Amps it should have a .1uF Cap be between the + & - Supply Pins.
Or Individual .1uF caps to Ground on each of these.

These Caps will help prevent Oscillations.

 

[Jester]

R10 in the original circuit might be intended as compensation for input current generated offset, but it's completely useless for a CMOS amplifier like OPAx348. The method is only appropriate for bipolar OP without internal input current compensation, in other words for |Ib| >> |Ios|.


Did you learn anything from the previous discussion? The circuit is stable with any passive source impedance connected to J2.

I think so....

Assuming the op amp is unity gain stable:
1) A non-inverting topology will be inherently stable because the gain is always > 1, Av=1+Rf/Rin
2) The main cause of instability is that the op amp introduces a phase shift that is frequency dependent, so at high frequencies the negative feedback is no longer negative feedback creating an oscillator (Thank you Audiguru)

The inverting amplifier shown in post 17, has a gain of 40k/120k or 0.33, well below 1, so I would think this circuit may be unstable, however your comment "the circuit is stable with any passive source impedance" implies otherwise.

So I'm getting there, but not there yet.

Thanks for all the replies.

 

[Audioguru]

An oscillator needs the gain to be a little higher than 1. So with a gain of only 0.33 it will not oscillate.

 

[crutschow]

Double check all your circuit connections, especially the MOSFET package pinout.
Sounds like something is misconnected.

 

[FvM]

A non-inverting topology will be inherently stable because the gain is always > 1, Av=1+Rf/Rin

The "because" in this statement is simply wrong. How did you arrive at this conclusion?

Stability isn't related to gain Av (= closed loop gain). Instead it's constituted by the phase and magnitude of the loop gain, the product of amplifier open loop gain Avol and feedback factor β. An unity gain compensated amplifier is stable for real feedback factors β <= 1, which includes any inverting amplifier resistor ratio. Stability is possibly not achieved for feedback factors with magnitude <= 1 and additional lagging phase shift, but that's beyond the scope of this question.

For a general analysis you'll refer to the Nyquist stability criterion mentioned by LvW. For most practical OP circuits, it's sufficient to check the loop gain phase at the transition frequency where |Avol*β| = 1, the so-called phase margin. If it's > 0, the amplifier is stable.

 

[tggzzz]

Re: OP amp stability gain &lt; 1

1) A non-inverting topology will be inherently stable because the gain is always > 1, Av=1+Rf/Rin

False, pure and simple. For examples, have a look at the data sheets and application notes of opamps that are not fully compensated.

But practical examples only indicate you are wrong; they don't why you are wrong nor what you have to do the avoid the problem. For that you have to understand the theory.

I suggest you look as some first year undergraduate textbooks on control theory. There you will find that the key concepts are loop gain (i.e. not amplifier gain nor gain), phase margin, and the nyquist diagrams are a sufficient methodology and a useful analysis technique.

I'm sure a little googling for those terms will be highly beneficial for you.

- - - Updated - - -

An oscillator needs the gain to be a little higher than 1. So with a gain of only 0.33 it will not oscillate.

False. An oscillator needs the loop gain to be slightly higher than 1.

It is trivial to design a circuit with a gain of 0.33 that oscillates. Consider, for example, an amplifier with a gain of 2 preceded by a volume control with a gain of 1/6. Now make the amplifier with an opamp which isn't fully compensated, and which will oscillate.

 

[crutschow]

Ignore my post #26. It obviously is in the wrong thread.
Because of the unique and strange rule in this forum that you can't modify a post after 30 minutes, I couldn't delete it.

Tggzzz, this post has been largely talking about op amps that are unity-gain stable so your lambasting of other posters by using non-fully-compensated op amps as an example is a red herring.
If you want to be critical of other's abilities I suggest you first pay more attention to all the posts in the thread.