Low pass Active RC filter + PSPICE

Hey.. below is my lowpass schematic in PSPICE and the simulation result.
The circuit is taken from the opamp datasheet so there should be nothing wrong with the circuit.. This opamp support single supply at 5V..

However, the simulated AC sweep result shows a negative gain..? 600uV..? How can that be possible when my input signal is 1mV..?

Anythoughts..? By the way, I have spent no lees than 4 hours trying to get just the lowpass filter to simulate correctly.. And i am out of ideas.. Any ideas..?
Thanks..

Opamps behave strangely (phase reversal and saturation) when you give them inputs outside their common mode range. In my experience this happens at about -0.3VDC, so I'm at a loss to explain your result. The 1 mV AC input does take the input below ground. If you want to work with a single supply then you have to bias the inputs to V/2 and AC couple them.

Before you tackle that I would suggest that you add another 9V battery to make a negative supply and rerun the simulation. Once you have some confidence you can try for SS operation.

MrNobody said:

Hey.. below is my lowpass schematic in PSPICE and the simulation result.
The circuit is taken from the opamp datasheet so there should be nothing wrong with the circuit.. This opamp support single supply at 5V..

However, the simulated AC sweep result shows a negative gain..? 600uV..? How can that be possible when my input signal is 1mV..?

Anythoughts..? By the way, I have spent no lees than 4 hours trying to get just the lowpass filter to simulate correctly.. And i am out of ideas.. Any ideas..?
Thanks..

Click to expand...

Spice does an AC analysis by first computing the DC operating point, then linearizing all components at that operating point. It does not matter what AC amplitude you use. The gain will be the same, because the network has been linearized.
Your problem is, the op amp output is at (or very near) zero volts. At that DC level, the op amp has very little gain, because the pulldown transistor is saturated.
You need to raise the DC bias on the input, or add a negative supply. Try making the DC level of your input 1V. Also try outrageous AC voltages, like 1000V. If Pspice is the same as all other spices I have used, you will get 6dB gain (gain=2) in the passband, regardless of the AC voltage.

Did you calculate the transfer function? Do you know what the answer is supposed to be, obviously a filter, but what exactly is the frequency response supposed to look like?

Maybe its working as designed, run through the theory and double check that.

3iMaJ said:

Did you calculate the transfer function? Do you know what the answer is supposed to be, obviously a filter, but what exactly is the frequency response supposed to look like?

Maybe its working as designed, run through the theory and double check that.

Click to expand...

If he is measuring the gain within the passband, the gain should be 2, by inspection.

Yeah I could see the gain, but it was the two caps that concerned me and the associated roll off. Aside from the gain the general trend looked correct which I guess is a sign to me to go back and double check the math that I used to design it.

A bipolar supply would help w/o a dc offset on the input, probing the actual time domain output might help show why the gain is incorrect if the op amp is cutting off the lower swing of the input that might be why the gain is incorrect.

But I guess its been so long since I've seen an actual schematic for an op amp circuit the only way I could figure whats going on anymore is to look at the transfer function, so I suppose it was mainly for my own sanity.

The frequency response of the Sallen and Key lowpass filter will have a peak near the cutoff frequency when the resistors and capacitors have equal values and the gain of the opamp is more than about 1.6, when the opamp is biased correctly.

audioguru said:

The frequency response of the Sallen and Key lowpass filter will have a peak near the cutoff frequency when the resistors and capacitors have equal values and the gain of the opamp is more than about 1.6, when the opamp is biased correctly.

Click to expand...

I test the circuit on hardware using a single 5V up to 9V supply but the opamp only output the positive half of the sinewave.. the negative half is chopped off.. Finally, I use MAX232's charge pump to supply both positive and negative voltage to the opamp and it works.. The gain is close to what audioguru said..

Hmm.. using MAX232's chargepump.. Somebody once told me that it is not recommended to use MAX232 to provide negative supply to opamp.. However, when I observe the output from 100Hz up to 10kHz, i see quite a good sinewave (with an exception of phase shift as the frequency gets higher).. What do you guys thing..?

Oh yeah, another thing is during the testing, my input voltage to opamp is 100mV. Just wondering, if I use the OMNI microphone similar to **broken link removed** what would the output voltage from the mic be..? Is it around 1mV, 10mV or 100mV..?
I tried to test mine by connecting the 2 terminals directly to the oscilloscope and then say something while looking at the oscilloscope but i got nothing..

An opamp doesn't need two supply voltages. Its input can be biased at half the supply voltage and have a coupling capacitor from the signal source. Then the output will be able to swing almost the entire supply voltage.

You show an electret microphone. It has a polarity and needs to be biased correctly. Without a bias voltage its output is nothing.
Its output level is about 5mV to 10mV when you speak with a normal voice level about 10cm away.

audioguru said:

An opamp doesn't need two supply voltages. Its input can be biased at half the supply voltage and have a coupling capacitor from the signal source. Then the output will be able to swing almost the entire supply voltage.

You show an electret microphone. It has a polarity and needs to be biased correctly. Without a bias voltage its output is nothing.
Its output level is about 5mV to 10mV when you speak with a normal voice level about 10cm away.

Click to expand...

Thanks.. Based on what you have said, I have made some changes to the preamp circuit.

The circuit is attached below, based on the assumption that the opamp will work at 5V.. Does it seem right to you..?

Let's say I want to test the circuit without using mic, can I replace the mic with 1.69 K resistor (impedance of the microphone) and input a 10mV sine wave to the coupling capacitor to observe the output..?

You changed the entire circuit:
1) The source impedance for a Sallen and Key lowpass filter must be very low. So it should be fed from an opamp. The new opamp should have its input biased at half the supply voltage and it should be the opamp with the input coupling capacitor.
2) If the gain of your lowpass filter is more than only 1.6 then the frequency response will have a peak.
If the gain of the opamp is more than 3.0 then it will oscillate.
Use gain in the first opamp.
3) Both opamps need a capacitor in series with their gain resistors to ground so that they don't amplify their DC bias voltage.

An LM324 is a quad opamp in a 14 pins case. An LM358 has the same opamps but it is a dual in an 8 pins case.
They are both horrible old opamps for audio because they are very noisy and have crossover distortion.
The MC34071 single, MC34072 dual and MC34074 quad opamps have low distortion and work with your low supply voltage. But they are not low noise.

Thanks for your update..
hmm.. i am confuse between noisy opamp and an opamp that is not low noise..
i mean, if an opamp is noisy, it is not low noise type right..?
aren't they the same parameter..?

The LM324 and LM358 are so noisy that the datasheet doesn't say anything.
The noise from an MC34071 is 32nV/root Hz. Its minimum sopply voltage is a total of 3V.
The noise from an LM833 audio dual opamp is only 4.5nV/root Hz. But its minimum supply voltage is a total of 10V.

Thanks..

Hey.. i built your preamp circuit and it works fine..
The only thing is that it doesn't swing a full 5V.. When i observe it on the oscilloscope, the positive part get clipped off at around 4V when I power the opamp at 5V. Is it because the opamp is saturated..? Well, I can operate the opamp with an output V(pk-pk) of 3V but I am just curious of why it doesn't swing a full 5V..

MrNobody said:

Thanks..

Hey.. i built your preamp circuit and it works fine..
The only thing is that it doesn't swing a full 5V.. When i observe it on the oscilloscope, the positive part get clipped off at around 4V when I power the opamp at 5V. Is it because the opamp is saturated..? Well, I can operate the opamp with an output V(pk-pk) of 3V but I am just curious of why it doesn't swing a full 5V..

Click to expand...

The opamp has two emitter-followers cascaded making a darlington transistor at its output to pullup the output. The output of the darlington goes as high as about 1.2V less than the supply voltage.


The output is pulled down close to ground if its load resistor goes to its negative supply voltage that you have as ground. Its output voltage has a voltage loss if the output transistor must sink current.

audioguru said:

The opamp has two emitter-followers cascaded making a darlington transistor at its output to pullup the output. The output of the darlington goes as high as about 1.2V less than the supply voltage.


The output is pulled down close to ground if its load resistor goes to its negative supply voltage that you have as ground. Its output voltage has a voltage loss if the output transistor must sink current.

Click to expand...

I see.. so, that is the nature of the opamp (internal costruction) and not the external preamp circuit which causes that..
Thanks..