Differential amplifier output frequency differing from input

Hi,

I have a differential amplifier with an adjustable gain between 30-200 with input signals of 1V DC + 3mV sine and 1V DC. The idea is of course to amplify the AC part, not the DC part. Transient simulations however show that the shape of the output sine signal is a little distorted, with the top end of the sine wave looking thicker than its bottom end, if the output voltage nears the supply voltage level (3.3 V) and vice versa if it nears GND level. FFT of the output shows that there are smaller fractions of the signal around 1kHz ( for instance at 1.4kHz, 2kHz, 3kHz ) (input signal is set to be a 1kHz sine wave with a 1V DC, the signal band I am interested spans from 100Hz to 10kHz). Can anyone tell me the reason for this and how I can avoid it?

Thank you very much

Ugur

It sounds like non-linearity in the amplifier near the rails. You either need to not drive it so close to the rails or look for an amplifier which maintains its linearity closer to the rails.

A circuit and some results may help produce a more specific answer.

Keith

Re: Differential amplifier output frequency differing from i

Ugur Yegin said:

Hi,
.............
I have a differential amplifier with an adjustable gain between 30-200 with input signals of 1V DC + 3mV sine and 1V DC. The idea is of course to amplify the AC part, not the DC part. Transient simulations however show that the shape of the output sine signal is a little distorted, .................

Click to expand...

To be exact, each amplifier adds some distortion to the (amplified) input signal. Thus, each output signal looks "a little distorted". And in each case, one has to decide if the distortion level is acceptable or not. Perhaps in your case not!
As KEITH has mentioned, most probably your output signal comes to close to the supply rail. Therefore, first check if the the available range around the operating point is large enough and more or less symmetric around this point.
What about the dc part of the input signal? Why do you mention it? Don't you cut it using a coupling capacitor?