Op amps(compensated) behave like low pass circuits, with 'Wb' as the 3dB frequency.
If let me assume 10 Khz is the 3 dB frequency the is the bandwidth of the op amp from 0kHz to 10 Khz?
And what happens to the bandwidth when opamps re cascaded, does it increase or decrease?
Any kind of suggestion is welcome.

 

Each op amp type will vary in this respect. The figure for bandwidth is what determines the limit of response. The frequency stated is divided by the gain of the op amp, so the frequency rolloff becomes the bandwidth divided by the gain.

Say an op amp gas a bandwidth of 1 MHz. If it is used with a gain of 100, the frequency limit is 1,000,000/100, or 10,000. Why this happens is due to the limited ability of the op amp to slew the oputut in response to a change at the input. When the output has to change faster than the op amp can respond, the output level drops from the set gain.

This is referred to as the gain-power bandwidth. Here is a link to an article that explains it more fully - http://en.wikipedia.org/wiki/Gain-bandwidth_product.

 

It should be noted that GBP is a phenomena of voltage feedback amplifiers. Current feedback amplifiers have GBP independence up until its break point which is high in frequency.

 

sndpgr,
did you mean 10khz is the 3db rolloff frequency of the openloop gain?
and did you mean bandwidth as to the closed loop gain?
if so, it is this way. the bandwidth of a closed loop op amp is at a much higher freq than that, and it goes like what beenthere said.
that's because the openloop gain is usually very high (say 100k to a few million times), while we usually set the closed loop gain at only fractions (way below thousands times) of that.
when you cascade two opamps, they do affect the total openloop gain and thus the/each closed loop gains. but the closed loop gain (which we are using) is very near to flat along our intended bandwidth. Phase shift is somewhat affected along the frequency though.

 

did you mean 10khz is the 3db rolloff frequency of the openloop gain?
and did you mean bandwidth as to the closed loop gain?

Simply look at the graph of the opamp's open-loop frequency response on its datasheet.
An old 741 opamp has an open-loop voltage gain of about 200,000 from DC to about 3Hz. At 1kHz its voltage gain is reduced to about 900 and at 1MHz its voltage gain is reduced to less than 1.
In addition, its output is slew-rate limited so at its full output of 28V p-p any frequency above only 9khz results in a lower level triangle wave.

If the closed loop gain of a 741 opamp is 100 then its frequency response rolls off above 8khz. If the voltage gain is 1000 then its frequency response rolls off above 800Hz.

when you cascade two opamps, they do affect the total openloop gain and thus the/each closed loop gains. but the closed loop gain (which we are using) is very near to flat along our intended bandwidth. Phase shift is somewhat affected along the frequency though.

Opamps that are the same cannot be cascaded with a single overall negative feedback. They each must have their own negative feedback. Then if each one has a rolloff frequency of 8kHz where the response is down 3dB (0.707 times) then two cascaded will be down 6dB (0.5 times) at 8kHz.

 

It should be made abundently clear, Gain-Bandwidth product is a small-signal parameter. For a desired gain and signal, one must check both the amplifier's gain-bandwidth product along with the slew rate. Also, it would be nice to see if you are still within supply rails.

To put an answer simply, gains are cumulative. Two amplifiers with gains of 5 can be cascaded with a total gain of 25. When you get familiar with THD and input noise, then you will see that using several amplifiers cascaded should be avoided. Or at the very least, use most of the gain at the first stage.

If you want a very precise gain at a specific frequency, do not count on the G-B product! Remember, this is specified at the -3dB point. Be conservative

Steve

 

Thanks for the replies
So the bandwidth is determined by the unity gain frequency(bandwidth product) and not by roll off frequency.
And
rwmoekoe
I mean open loop op amp.
And 10khz was just an example now since bandwidth is dc gain*roll off frequency.So I got the answer.
But I still have doubts with the resultant bandwidth of cascaded open loop amplifiers, I mean if two similar op amps are cascaded then the dc gain is (dc gain)^2, what about the roll off frequency.

 

Just treat them as first order low-pass filters that are frequency dependent. Your -3dB point is your G-B product.

When in doubt, prototype it if you can. You'll be able to check waveforms and feel comfortable about it.

Steve

 

But I still have doubts with the resultant bandwidth of cascaded open loop amplifiers, I mean if two similar op amps are cascaded then the dc gain is (dc gain)^2, what about the roll off frequency.

One opamp has its output down -3dB (0.707 times) at its cutoff frequency and two cascaded identical opamp amplifiers will have the output down -6dB (0.5 times} at the cutoff frequency of each one. The response is not Butterworth but is a rather droopy cutoff.