Op-Amp integrator for high frequency application

Thread Starter


Joined Apr 28, 2014
hi everyone, I am currently doing a project which used PLL as a carrier for amplitude modulation. Since I can obtain triangular waveform from the PLL(LM567), I would like to integrate the triangular waveform into a normal sine wave to be feed into my circuit input. However, the integrator circuit doesn't seem to be working at very high frequencies.

I would really appreciate it if anyone can shed some light regarding this. Thank you

P.S: I hereby attached my integrator circuit below. This circuit works when I change the frequency of function generator to 100Hz, but failed when I tweak it up to 100kHz.



Joined Jul 17, 2007
kevinraphael, don't let #12's sarcasm get to you; please realize that we see people trying to use the ancient relic 741 opamps every day, and it really does get old.

When they were first introduced, they were the cutting edge of a revolution in electronics; before the 741, vacuum tubes were being used - bulky, hot, unreliable vacuum tubes. And then the 741 burst upon the scene, and the entire game changed.

However, that was a half-century ago, and opamps have evolved; there are a great many that are far better for your purpose.

Try something like an MC33078 or MC34063 or OPA134.


Joined Nov 30, 2010
Did that come across badly?
I am quite serious. The 741 chip is a miserable antique that can't produce a sine wave of more than 2.8 volts RMS at 20 KHz, no matter how much voltage you give it for power supplies.

Look at slew rate. SR = 2 Pi F Vpeak
Plug your 100 KHz into that and see how much voltage it can produce.

math math math...less than .8 volts peak.


Joined Mar 6, 2009
Plus the 22nF integrator capacitor will drop the overall gain to a very low value at 100KHz. The capacitive reactance at 100KHz is about 72 ohms. The effective stage gain then becomes very small, even with a more suitable op-amp. You might do better to use a 2nd order filter type such as a Sallen-Key topology or if a narrow bandwidth is acceptable then a tuned amplifier stage with suitable Q.


Joined Aug 1, 2013
Without getting into the opamp for a moment, your overall circuit has a low frequency gain of 10. It functions as a lowpass filter with a corner frequency of 72 Hz, so at 100 KHz the output would be almost 43 dB below the input. Your corner frequency is over 10 octaves below the fundamental, so this certainly will turn a triangle wave into a sine wave. In fact, it will turn just about anything into a sine. But a very small sine.

All of that is without figuring in the internal 10 Hz lowpass filter built into the 741 open loop gain function. As noted above, faster, contemporary parts will work better. In round numbers, a feedback circuit needs 20 dB of headroom to function. So for an opamp to pass a 100 KHz signal as a normal opampy circuit with a gain of 10, it should have at least 40 dB of open loop gain at 100 KHz. Very few devices do. The MC33078 has around 40, MC34063 is a switching power supply regulator, and the OPA134 has around 40.

What are the input voltage amplitude and the output voltage acceptable amplitude range?



Joined Mar 14, 2008
An integral of a triangle-wave does not give a pure sinewave. It is a series of parabolas. But it can be a good enough approximation for non-critical uses, having odd-harmonic distortion, the 3rd harmonic being about 29dB below the fundamental.

Thread Starter


Joined Apr 28, 2014
Thank you guys for all your opinions, seems like I really have a shallow understanding on the integrator circuit. Unfortunately I am unable to get opa134 currently. By reading all your comments, am I to assume that it is not possible to proceed with LM741?


Joined Aug 1, 2013
Withoug knowing where you are or what access you have to comtemporary components, here is a generic qualification method.

Figure out how much gain you need at your highest frequency of interest.

Add 20 dB.

For any opamp under consideration, get the full data sheet. There should be a plot of open loop gain versys frequency. Find your highest frequency on the x axis, move up to where it intersects the plot line, then move to the left to see what the gain is at that frequency. If it is less than your requirement, move on to the next possible part. If the datasheet doesn't have the plot, move on. It can be frustrating work, but it is part of the design function.

There is no magic search engine for this. If you have a favorite (or company-specified) IC vendor, they might have a parametric search page that lets you search for parts with a minimum GBW. Also, big distributor sites like Digi-Key and Mouser might have one that covers the vendors they handle.