Simple op amp integrator in Orcad Pspice

[Qacer]

Hi all,

I have a simple op amp integrator schematic in Orcad Pspice Capture. I'm using a ua741 op amp. My voltage source is VSIN with 15v amplitude and 60Hz. I connected a 10k resistor to the volt source and the inverting input terminal. I grounded the noninverting input terminal. Also, I connected a 10uF capacitor between the inverting input terminal and the output. A 100k resistor is connected to the output so I can measure the voltage. I'm biasing my op amp with 15 VDC.

The problem is that I can't get the output plot working right. I'm trying to get a cosine graph since I'm inputting a sine signal. The voltage probe on the output shows 15 VDC throughout the whole simulation. I'm using Transient Analysis (voltage vs. time). What's wrong?

Thanks..

 

[rimshaker]

When you put the voltage marker on the input side, do you see a sinusoidal signal?

 

[Qacer]

Yes. I see a sinusodial wave with my input probe. I'm still trying to find out why my output is straight up 15 VDC. The output is suppose to alternate since it is equal to - (1/CR) integral(input voltage) dt..

 

[Qacer]

By the way, here are some screen captures of my circuit and output:

Linky link

 

[rimshaker]

It works, just keep experimenting with different values for input resistor and the feedback cap.

 

[Superdoopercooper]

I think your problem may lie in the fact that you have no DC-path for bias current in you feedback path. Granted, the 150 uF capacitor is fairly leaky, but I don't think that it will provide enough bias-current for a 741. IF you had an LF356 you may be ok.

My suggestion is to try placing a 100kOhm or even 1MOhm resistor in parallel with the capacitor. Or if you want a gain of 1 still, then use 10kOhm, but then your -3dB breakpoint may be too hight for this thing to act as integrator at 60Hz. It's 7 am... and after 6 hrs of sleep my brain doens't have enough power to figure that out. Anyways... your op-amp is likely losing bias and going to the positive rail.

Anyways... I did figure out the gain at 60Hz.... ==> Gain = 1/(2*pi*60*R*C) == 1.768e-3. Therefore, you need to choose a parallel feedback resistor that will give you some gain, but also keep this thing integrating. Know what I'm saying? And/or try a smaller input resistor and/or smaller feedback cap.

 

[Qacer]

I've tried changing the values for my resistor and capacitor. It still did not work. I also tried adding a big resistance parallel to the capacitor. My output seems to be clipped in the negative portions; it never goes down. I tried using a 10uF capacitor and a 100k resistor to at least achieve a Vout that is the "actual, negative" integral of Vin, since Vout = -(1/RC) * integral(Vin). It did not work.

I can't use any other Op amp model since it only has the ua741 available. I am also limited to my capacitors and resistors.

What did you guys do to make it work? I've tried searching the web for similar integrator circuits, but when I tried their values it still did not work. My outputs have been a square wave, humpty dumpty, and half-sine waves, etc.. Thanks!

 

[davesaudio]

4am.. wow

did you try a simple inverting/non inverting cct w gain to ensure the bias is right?
I tend to agree that its a DC bias problem
don lancasters cook book here we come....

 

[pX]

...

 

[highwire]

A split power supply is needed for a 741. I just glanced at it and - yes, you either need a split supply or you must bias the inputs some where between the neg and pos of the power supply.

 

[highwire]

Also, with the values given, the output will only be about 25 millivolts at 60 Hz. Reducing the size of the feedback cap will up the output.

Edit:
Coop - This circuit is a "perfect integrator". It has a pole of 0 Hz. The response plotted log/log is a straight line sloping down to the right at 6db/octave. The only "corner" is one produced by the amp's finite gain somewhere on the left side of the plot.

No "leakage" or extra stuff is needed to make it work. In the real world, the amp would have to be heavily compensated to prevent oscillation at high freqs because of the near 100% feedback. Otherwise, it is a go. Just make sure the inputs are at lease 3 volt or so above the negative rail.

 

[Insidious]

edit: Do you have pins 4 and 11 reversed?

(also might not want to go for the infinite SS gain...)

 

[highwire]

Does anyone have any opinion on the best freebie spice circuit analyzers around? I found three downloads of various implementations on my HD. I even had a pretty good DOS version I used awhile ago.

Insidious: You made me look in the book! Good thing. It reminded me that the 741 0p amp IS compensated.
The pin-outs shown in qacer's diag are consistent for an eight pin metal package and seem to be right - except for needing a neg voltage source to pin 4.

This infinite gain circuit has been questioned by many in the thread.
To stabilize the circuit and center the output, a 1 meg - 100k resistor across the capacitor will make it a practical stand-alone circuit. This might not be necessary or desired if it is part of a larger control circuit with an overall feedback function. It was a good thing for you and other to point it out in various ways.

Qacer: Your circuit diag page loaded very fast - like a good commercial website. Better than AT, even. And also, anyone with "Thevenin" in their sig is on the right track in this biz. Think Thevenin.

 

[Herr Yunta]

Thevenin

 

[Qacer]

Thanks for the reply everyone! I appreciate it a lot. My friend told me to wire the circuit up and test it in real time. He said something about Pspice not being able to simulate certain things. I think he made a reference to his senior design project for which some parts could not be simulated in Pspice.

He showed me that it was actually integrating using an ideal model from the parts list. For some reason, he was having the same problems using the model for my only opamp (hc7411 > ua7411/eval).


highwire: As for the best spice analyzer software, I suggest getting an Orcad Pspice 9.2 Lite version for free. Go to www.orcad.com and sign up for a free CD. This thing is really good. If only it had a VHDL simulator, then I would love it even more.

 

[Menelaos]

Welcome to the world of saturation!

Menel.

 

[Qacer]

Yep.. Another friend of mine suggested that I match by resistor and capacitor with my input frequency. This seemed to do the trick. I'm still trying to find an exact explanation as to why. My other friend said that the transfer function of my integrator is independent of the input frequency. Hopefully, I'll get a clearer picture.

Now my new circuit has a 256k resistor and a .01uf capacitor in parallel with a 1meg resistor. Then, I feed its output to an inverting configuration opamp to get the actual integral. It works on Pspice, but now I need to test it out in real time.

I can't wait to show it off to the elementary kiddies come Expo day.. Btw, if you are in the Tampa Bay area, stop by at the University of South Florida on Friday (Feb 22) and Saturday (Feb 23) 8am - 4pm to see our Engineering Expo. Our group will be designing some air cannons tomorrow. Hopefully, will be able to shoot some stuff by the end of the day.