I'm trying to invert a square wave signal.
Signal:
Vp-p = 10V
freq = 2.44KHz - 2.6KHz

I want a gain of -1

In simulations both the 318 and 741 are able to do this with little distortions When I calculated the Slew-rate, the 318 was greatly higher overspec, and the 741 just had enough.

However, when I build the circuit. Even though I have the circuit hooked up exactly like the multsim like this:

there is a ton of noise on the oscilloscope.

If I replace the 318 with a 741 (without changing the circuit) all the noise goes away. If there any reason why it does this?

(I should have took a picture).
What's confusing me the most, is when the 318 is hooked up, even the input signal gets distorted. I tried setting the oscilloscope to look for averages, but even that was unreadable.

The 741 works, but the fact that I couldn't get the 318 to work this has always bothered me.

p.s I've already tried switching op-amps and oscilloscopes. DMM seems to be getting a different vale from the scope too.

 

LM318 has ten times more bandwidth than LM741.

The slew rate of LM318 is 70V/µs compared with 0.5V/µs for LM741.

Put a 10pF capacitor across R10.

 

LM318 has ten times more bandwidth than LM741.

The slew rate of LM318 is 70V/µs compared with 0.5V/µs for LM741.

Put a 10pF capacitor across R10.

Thanks for the quick reply. I don't have a capacitor that small (0.01uF is the smallest I have) to test with, but multsim saw no changes in the output, which almost guarantees this won't hurt.

So the high slew rate problem. I didn't realize too of a slew rate could cause a problem. Is there an equation you used for find 10pF?

 

If you look at the LM318 datasheet at ti.com, you will find a schematic on how they measured the slew-rate on page 4.

Also, the general specification of the slew rate was 50V/uS, with this caveat in the text:

For inverting applications, feed-forward compensation boosts the slew rate to over 150 V/μs and almost double the bandwidth.

 

The problem is your power to the op-amp is only ±8 volts. How can you expect to get ±10 V out from the LM741 op-amp that needs at least ±12 V on the power rail to get ±10 V on the output? The LM318 also needs 2 V extra on the power rails.

 

The problem is your power to the op-amp is only ±8 volts. How can you expect to get ±10 V out from the LM741 op-amp that needs at least ±12 V on the power rail to get ±10 V on the output? The LM318 also needs 2 V extra on the power rails.

Did you notice the input was 10V peak to peak? The gain of -1 is within the +/- 8 volt rails.

 

Hi,

Yes, you have to watch your input range as it is limited on these kinds of op amps.

Also, you should have 0.1uf caps on each power rail close to the chip package.

 

What you may be seeing with the LM318 is oscillation, not noise. Not only do you have to have ceramic bypass capacitors but you need to keep the input and output signals away from each other on your bread board. Long component leads make keeping the signals apart -- cutting the leads on the components so the bodies of the parts are in contact with the breadboard is best.

 

If you look at the LM318 datasheet at ti.com, you will find a schematic on how they measured the slew-rate on page 4.

Also, the general specification of the slew rate was 50V/uS, with this caveat in the text:

I'm not too sure how to measure the SR for a square wave, but I know for a sine wave, it goes something like

Vp*wo, the result is very small, doesn't not even compare to the 318's rate.

The problem is your power to the op-amp is only ±8 volts. How can you expect to get ±10 V out from the LM741 op-amp that needs at least ±12 V on the power rail to get ±10 V on the output? The LM318 also needs 2 V extra on the power rails.

I've tried pushing it up to 15V, I don't think that's the cause. If it was too low wouldn't it saturate? This OPAMP isn't saturating, it's corrupting everything from the output to the input.

Hi,

Yes, you have to watch your input range as it is limited on these kinds of op amps.

Also, you should have 0.1uf caps on each power rail close to the chip package.

Yup, I placed a 0.1uF capacitor across the +V to ground, and -V to ground.

What you may be seeing with the LM318 is oscillation, not noise. Not only do you have to have ceramic bypass capacitors but you need to keep the input and output signals away from each other on your bread board. Long component leads make keeping the signals apart -- cutting the leads on the components so the bodies of the parts are in contact with the breadboard is best.

How would I do that? Capacitor from input to ground? Or do you mean physically?

EDIT:
This might be it. I re-did the circuit but this time, placed the input single and OP-AMP on other ends, and got a nice square wave, but still this is disappointing. When I build more complex circuits I'm going to have to make it unnecessarily big if I want to use the 318.

Thanks to everyone who responded so far

 

Did you notice the input was 10V peak to peak? The gain of -1 is within the +/- 8 volt rails.

Joe, it says 20 volts p-p and that the current voltage is 10 V.

 

OP stated:
I'm not too sure how to measure the SR for a square wave, but I know for a sine wave, it goes something like ...

The LM318 datasheet shows how to measure the slew rate. If you don't know how to measure it, how did you get the figures?

Square waves are a lot easier than working with sine waves.

LDC3 said ...
Joe, it says 20 volts p-p and that the current voltage is 10 V.

I see where we both got confused ... in the opening text it stated:

I'm trying to invert a square wave signal.
Signal:
Vp-p = 10V
freq = 2.44KHz - 2.6KHz
I want a gain of -1

Yet the OP's diagram stated "20V pk to pk". I followed the text ... and didn't pay particular attention to the input signal on the diagram. We went with opposite directions. Sorry for not investigating it further. My apologies.

 

Rpy360,

Do you know how to measure the rise time in a signal?

Here are the simulations of the slew rate for both op amps you are working with ....

 

Hi,

One more quick note...

You may wish to look into a comparator like the LM319 or something like that. If you have square wave in and square wave out you dont really need an op amp all the time In fact, sometimes you can just use a transistor. It depends what you have to use the output for.