Input Offset Voltage of Op Amp and Integrator

Thread Starter

forbi

Joined Sep 11, 2012
37
Hi,

Firstly, i would like to know how could i set up a circuit to test for the input offset voltage ?
Because for TL071 Op amp, i saw a parameter measure information which shows a input offset-voltage null circuit. Can i use a function generator to generate a square wave of +- 10v at the V- and ground V+ and using oscilloscope to Vout to see if there is any deviation from the reference point of 0v? The supply voltage into the Op amp is +- 10V .

Secondly, i understand there is a integration constant (c) in an integrator. How can we get the value of the integration constant ? Do we do a setup as similar on on how i test for input offset voltage and check the t=0 value ?

Third, is there a way using theortical method to check on the frequency that will cause the op amp to be saturated ? Or i've to manually adjust the input signal from the function generator to visually look if the output is being saturated.

Lastly, is there really a different if we have the integrator circuit to have its capacitor being charged up before giving an input signal ? I can switch because i've attached an multiplexer to allow me to switch between charging up the capacitor and doing integration.

Really appreciate for your help.
 

crutschow

Joined Mar 14, 2008
34,285
Firstly - You can measure the offset by connecting the op amp as a high gain non-inverting amp (say a gain of 1000) and measure the output voltage with the input grounded. The input offset voltage is the output voltage divided by the gain.

Secondly - To check an integrator time (RC) constant input a voltage step. The time it takes the output to reach the input step voltage value (polarity inverted) is the time constant.

Third - What do you mean by "saturate"? Frequency does not saturate the op amp output. At high frequencies the output rolls off as determined by the gain-bandwidth product and/or distorts due to the output slew rate but neither of those is saturation.

Lastly - An integrator doesn't change its response if the capacitor has a initial charge. It just starts integrating from that value instead of zero.
 

Thread Starter

forbi

Joined Sep 11, 2012
37
Firstly - You can measure the offset by connecting the op amp as a high gain non-inverting amp (say a gain of 1000) and measure the output voltage with the input grounded. The input offset voltage is the output voltage divided by the gain.

Secondly - To check an integrator time (RC) constant input a voltage step. The time it takes the output to reach the input step voltage value (polarity inverted) is the time constant.

Third - What do you mean by "saturate"? Frequency does not saturate the op amp output. At high frequencies the output rolls off as determined by the gain-bandwidth product and/or distorts due to the output slew rate but neither of those is saturation.

Lastly - An integrator doesn't change its response if the capacitor has a initial charge. It just starts integrating from that value instead of zero.
Can i confirm my understanding on your answers?

First point: If i plainly just supply supply voltage to the op amp without any input signal, will i be able to find the offset voltage ? do i need to connect the inverting or non-inverting input to ground ?

Second point: what you mean by input a voltage step ? Is it a square wave or impulse ? So i just have to read off from the oscilloscope the time for it to reach the "input step voltage" ?

Third: what i mean is like if you input a square wave into this integrator it should give you a nice triangle wave without being clipped off or showing that the triangle is misaligned (not sure the technical to describe this ) . How could i know at which frequency will the integrator provide a nice triangle ? Is there like a theoretical calculation method or is shown in the datasheet of TL071 ? If so which portion should i look for ?

Lastly, Somehow or rather when i connect the integrator and input a square wave from +10v to -10v for instance it will never integrate from 0 even if i charge the capacitor for awhile before switching into integration mode using the multiplexer.
 

crutschow

Joined Mar 14, 2008
34,285
Can i confirm my understanding on your answers?
Yes.

First point: If i plainly just supply supply voltage to the op amp without any input signal, will i be able to find the offset voltage ? do i need to connect the inverting or non-inverting input to ground ?
As I stated, if you configure it as a non-inverting amp with high gain, you ground the input (resistor).

Second point: what you mean by input a voltage step ? Is it a square wave or impulse ? So i just have to read off from the oscilloscope the time for it to reach the "input step voltage" ?
It can be a square wave of a sufficiently low frequency so that you can read the time it takes for the output to reach the step voltage value.

Third: what i mean is like if you input a square wave into this integrator it should give you a nice triangle wave without being clipped off or showing that the triangle is misaligned (not sure the technical to describe this ) . How could i know at which frequency will the integrator provide a nice triangle ? Is there like a theoretical calculation method or is shown in the datasheet of TL071 ? If so which portion should i look for ?
You want to keep the triangle frequency well below the gain-bandwidth frequency (no greater that 1/10th) or slew rate limits of the op amp.

Lastly, Somehow or rather when i connect the integrator and input a square wave from +10v to -10v for instance it will never integrate from 0 even if i charge the capacitor for awhile before switching into integration mode using the multiplexer.
Why do you sat that? Except for offset drift, the integrator will start integrating from whatever the initial capacitor voltage is.
Answers are in blue.
 

MrChips

Joined Oct 2, 2009
30,720
Correction. In an RC integrator, the output will approach the final voltage asymptotically, i.e. in theory it will never reach the final voltage.
The RC time constant is the time it takes to reach 63% of the amplitude.
 

Thread Starter

forbi

Joined Sep 11, 2012
37
Answers are in blue.
Thanks for the promptly reply.

I've kinda get the idea of it, will be testing it again tomorrow.
I've also calculated the highest frequency that will provide a distortion free operation which is approximately 200Khz.

Looking forward, if the input frequency is very low (15Khz square wave for example), the output signal on my previous test show that it will reach its max amplitude and stay constant for a few ms when looking at (t/2) on the "high" of the square wave.

Question will be, how will i be able to calculate the lowest frequency input signal to prevent the signal to stay constant for a few ms after reaching the max amplitude ?
And also the optimal frequency that the output signal will always be a nice triangle wave?

Thanks !!
 

crutschow

Joined Mar 14, 2008
34,285
................................
Looking forward, if the input frequency is very low (15Khz square wave for example), the output signal on my previous test show that it will reach its max amplitude and stay constant for a few ms when looking at (t/2) on the "high" of the square wave.

Question will be, how will i be able to calculate the lowest frequency input signal to prevent the signal to stay constant for a few ms after reaching the max amplitude ?
And also the optimal frequency that the output signal will always be a nice triangle wave?
By "stay constant" I assume you mean the output saturates at output maximum or minimum. If so then you either need to increase the RC time constant or increase the input frequency to reduce the amplitude of the triangle wave.

For good triangle fidelity you should probably go no higher in frequency than about 1-2% of the gain-bandwidth-product of the op amp.
 
Top