opamp roll-off

[Kevin Weddle]

Is utilizing opamp roll-off a good idea? A 741 opamp has a gain of 100 at 100KHz, and the output is stable. External filtering components are most often used, so is opamp roll-off avoided? If I set the gain at 1000 and use a higher frequency to bring the gain to 500, would that be acceptable design?

 

[Hero999]

I'll answer in case a newb reads this and gets confused. . .

Where did you get your information from? I couldn't find a frequency response graph on the datasheet.

You couldn't be right or wrong, either way it's not an acceptable design because the slew rate limits the output voltage to an unacceptably low level at 100kHz, increasing the frequency only makes things worse.

The amplifier circuit should be correctly designed so the gain is reduced using negative feedback rather than the frequency roll-off which is unpredictable.

 

[audioguru2]

Is utilizing opamp roll-off a good idea? A 741 opamp has a gain of 100 at 100KHz, and the output is stable. External filtering components are most often used, so is opamp roll-off avoided? If I set the gain at 1000 and use a higher frequency to bring the gain to 500, would that be acceptable design?

You make no sense.
The datasheet shows that the voltage gain of a lousy old 741 opamp is only about 8 at 100kHz, not 100.
The filter capacitor is built inside the opamp (not external) to prevent it from oscillating when its phase-shift causes 180 degrees delay. The capacitor causes a roll-off so that the voltage gain is less than 1 at a frequency that will cause oscillation.

If you set the gain of a 741 opamp to 1000 then its output has a cutoff frequency at about 800Hz. The gain is 0.7 times (-3db) at the cutoff frequency. View attachment 40875

 

[Hero999]

Where did you get that graph from?

The datasheet I found on Google did not include that graph, although I know I've seen it somewhere before.

 

[audioguru2]

Where did you get that graph from?

The datasheet I found on Google did not include that graph, although I know I've seen it somewhere before.

Most datasheets do not have enough frequencies listed. This one is from Philips and it has correct logarithmic frequencies.

 

[Hero999]

Thanks, I've found the datasheet and saved it, here's a link.

http://www.nxp.com/documents/data_sheet/UA741X.pdf

 

[Alex Tsekenis]

Guys dont bother with Kevin, he is a post bot to generate search engine hits.

 

[Hero999]

But some of his posts are wrong and might confuse a newb into believing incorrect information.

 

[Kevin Weddle]

You couldn't be right or wrong, either way it's not an acceptable design because the slew rate limits the output voltage to an unacceptably low level at 100kHz, increasing the frequency only makes things worse.

The amplifier circuit should be correctly designed so the gain is reduced using negative feedback rather than the frequency roll-off which is unpredictable.

Filter circuits attenuate the same as an opamp would above it's bandwidth. It's not recommended to allow the opamp to reach it's 40db/decade attenuation because it might oscillate. I'm wanting to know if the distortion is more appreciable above the opamps bandwidth with the attenuation at 20db/decade. In which case two opamps cascaded might be better.

 

[audioguru2]

An opamp uses a single capacitor frequency compensation filter that attenuates at only 20dB/decade so that the phase shift is only 90 degrees. Since the RC reduces the open-loop gain then the negative feedback is less at high frequencies so the distortion is increased.

The slew-rate limitation at 100kHz for good opamps converts signals to triangle-waves which have extreme distortion.

Cascading opamps produces less attenuation due to frequency compensation so the bandwidth is more and the distortion is less than if a single opamp is used.

 

[Kevin Weddle]

And that is why operating the opamp at 100Khz doesn't matter. The external capacitors dictate a 20db/decade roll off. But what if you don't have external capacitors. Operating the opamp beyond the bandwidth doesn't make more distortion?

 

[audioguru2]

And that is why operating the opamp at 100Khz doesn't matter.

The open loop gain of an opamp decreases above 10Hz. Then there is less negative feedback at high frequencies so the distortion is much higher than at lower frequencies.

The external capacitors dictate a 20db/decade roll off. But what if you don't have external capacitors.

Opamps have internal frequency compensation capacitors for the last 40 years. The first opamps (uA709 and others) used one or two external capacitors.

Operating the opamp beyond the bandwidth doesn't make more distortion?

If you exceed the slew-rate limit then the distortion is severe.

Look at the graph of distortion vs frequency for a TL07x opamp on its datasheet. The closed-loop gain is set to 1 and the output level is 6V which is a little below clipping.
The distortion is only 0.003% up to 12kHz.
It is 0.01% at 40kHz.
It is 0.2% at 100kHz when slew-rate limiting will increase the distortion much more.

EDIT:
Added the graph. View attachment 40905

 

[Kevin Weddle]

I don't understand slew rate. It is a fuction of voltage and time. Better opamps have great slew rates.

 

[audioguru2]

Slew Rate is the speed that the output of an opamp can change its voltage. At the frequency and amplitude where it is limited by capacitance in the circuit, changes become ramps.
Sine-waves and square-waves become triangle-waves.
above the slew-rate frequency limit the ramps reduce the max output level.

 

[Kevin Weddle]

Sine-waves and square-waves become triangle-waves.
above the slew-rate frequency limit the ramps reduce the max output level.

A triangle wave? A sinewave becomes a triangle wave exceeding the slew rate? Any high frequency approaching the  slew rate is too high. Slew rate is a characteristic of the performance of the device. It's the same as device frequency limitation.

 

[Hero999]

I'll have a go at explaining.

The slew rate is the maximum rate of change in output voltage.

Suppose slew rate is 1V/

 

[Kevin Weddle]

A capcitor fe

And that is why operating the opamp at 100Khz doesn't matter. The external capacitors dictate a 20db/decade roll off. But what if you don't have external capacitors. Operating the opamp beyond the bandwidth doesn't make more distortion?

Hero99 said operating the 741 opamp on a 20db/decade curve is a good idea. I agree. It is only a general purpose device.

 

[Hero999]

Where did I say that?

Post a quote?

I wouldn't say that, I'd say that the 741 should be operated so it still has plenty of gain at the frequency you want to amplify  and that the output should be kept within the slew rate.

Anyway I tried, maybe I shouldn't have bothered.

 

[Kevin Weddle]

Hero99, the post is about roll-off. A 20db/decade roll-off of an opamp approaches the limitation of the device. You extend the bandwidth of an opamp by reducing gain.