Summing amp output with an LPF turns the output into a triangle wave.

[Yamin]

Hi guys, I wanted to play around with some passive filter design. I wanted to add two or more sine waves using a summing amp and play around with different passive filter topologies.
I am using a 741 opamp and I have attached a screen shot of my circuit. I'm inputting around 120Hz and 1kHz sine wave and I started with a simple RC LPF, the value I choose for Rs is 100ohms and 1uF capacitor which should result in a corner frequency of around 1.5kHz. With the simulation I am able to play around with the values and am able to filter out the 1k signal.
 But in real life the output looks like a triangle wave when measured on the scope, it is filtering out the 1k signal I know this because when I remove the 120hz signal there is no output.
Any help on this would be greatly appreciated.

[VinzC]

The 741 has a terrible slew rate aka 0.5V / µs so you're probably experiencing these effects. Try faster op amps such as TL07x, for instance. You should see significant improvements. Check the datasheet https://www.ti.com/lit/ds/symlink/lm741.pdf for details.

Also note that the minimal slew rate, given an amplitude A and a frequency f is 2πfA, so in your case, with 15V output swing, that gives around 0.15V/µs at 1.5kHz. If your real-life test is worse, then it most probably means your op amp doesn't meet the required specs.

[Wolfram]

1kHz at 10 V peak is only 63 V/ms, so slew rate limiting should not be a factor here. Limited output current capability of the 741 could be however. 100 ohm is pretty heavy for a 741 to drive. Try with 1 k + 100 nF or 10 k + 10 nF instead

[VinzC]

1kHz at 10 V peak is only 63 V/ms, so slew rate limiting should not be a factor here. Limited output current capability of the 741 could be however. 100 ohm is pretty heavy for a 741 to drive. Try with 1 k + 100 nF or 10 k + 10 nF instead

Your explanation makes more sense indeed. At 1kHz, the output capacitor together with the resistor present a 180 ohms load (rough calculation) and the output of the 741 switches to constant current above ~4V. With that said, it would be interesting to see the actual waveform.

[Yamin]

Hey guys, I read the dials on the audio generators wrong . I was actually inputting 1.2kHz and 50kHz sine waves. I came to know of this after double checking them on the scope today.
I have attached a photo of the triangle waveform I mentioned, I am still using the 100 ohm resistor and the 1uF capacitor.
I have 3 more follow ups:
1. Why is showing the triangle waveform in this case?
2. How can you calculate the current draw from the opamp output when you've got a capacitor on the output
3. When a load say speaker is connected to listen to the output, would it alter the result if so how to mitigate it?

I didn't get to play around with it today, but setting the audio generators to 120Hz and 1kHz, I was able to filter out the high frequency nicely.
Thanks for the help

[Sensorcat]

I have 3 more follow ups:
1. Why is showing the triangle waveform in this case?
2. How can you calculate the current draw from the opamp output when you've got a capacitor on the output
3. When a load say speaker is connected to listen to the output, would it alter the result if so how to mitigate it?

• All the standard circuits for OpAmps work with the assumption that the limits of the amplifier used are not exceeded. VinzC and Wolfram have mentioned two of these limits: The slew rate is the maximum speed at which the output of the amplifier can change. There's also a current limit. If you load the amplifier such that more current would be required to provide the desired signal, the amplifier will output the maximum current. And charging a capacitor with constant current creates a triangle waveform.
• VinzC calculated the impedance of the capacitor @1kHz.
• Most speakers will have an even lower impedance than the load you use now. There will be a voltage drop in the amplifier that changes the result. Use a different amplifier to drive a speaker.

[Terry Bites]

741s only exist because people keep ordering it by accident.
Kinda like doner kebabs.

[ArdWar]

741s are kept afloat by sale orders from students trying to replicate circuits they see in their never updated textbooks.

Oh, and spare part for legacy equipment.

[VinzC]

741s only exist because people keep ordering it by accident.

3. When a load say speaker is connected to listen to the output, would it alter the result if so how to mitigate it?

I'd add to what Sensorcat said that (most) operational amplifiers are not designed for driving moderately high loads such as speakers (unless explicitly specified). Regardless always refer to the datasheet and check operating parameters.

Short: don't connect a speaker directly to a 741.

Additionally: if you want better quality, there's the NE5532, much better suited to audio. The 741 is crap. Really. Nice one back then but absolutely worthless for audio nowadays.

[Yamin]

Thanks everyone for your input.
I wanted to demonstrate the effects of filter so instead of just viewing from the scope I wanted to pass it through some speakers to hear the effect.
The 741 are the opamp in the training boards at college hence why I am using them hehhe.
I'm thinking of having a buffer stage after the filtering stage and pass the audio through to a powered speaker. That shouldn't effect the results right?