How to analyse input and output offset voltage of opamp?

[kankanas]

Hello all,

I need help understanding the calculation of offset voltage for OP-AMP. Currently, I am trying to find the offset voltage of TL074 in LT-Spice. As per the datasheet given, Input offset voltage (Vos) is typically 3mV and Max 10mV. (supply is +-15V)

My understanding is if we short the two inputs of the OPAMP to ground and measure the
output voltage, whatever value will get at output node is the OUTPUT offset voltage not input offset voltage. To measure input offset voltage a test voltage has to be added in one of the input terminals and have to vary. The value at which output becomes 0V must be INPUT offset voltage. But I am not sure as the input offset is not matching with the datasheet and any information about OUTPUT offset I could not see. So, My doubts are in general theoretically----

1. How to find the input offset as well as the output offset of an OP-AMP? (Please provide circuit configuration for both)

2. When the OPAMP is included in big circuit configuration how we can approach for offset voltage at each node?

3. Can the offset change with different configurations and different feedback strategies (negative feedback and open loop mainly)?

Thanks in Advance

KS

[TimFox]

The input offset voltage is the DC voltage that exists between the two input pins when the output is zero.
Since the DC open-loop gain is so high, if you short the two pins with no feedback, the output will almost certainly hit one or the other maximum output (near power rails).
The normal way to measure the offset quantitatively is to use two resistors in a feedback network to set the closed-loop gain to something reasonable like 100 or 1000;  the input offset is then the output voltage divided by the "noise gain" (q.v.) of the feedback amplifier.  The change in input voltage due to the few volts at the output in this test is negligible.
"Output offset voltage" is not a useful parameter for normal op amps.  Obviously, the actual output voltage depends on the input offset voltage and the noise gain.

[kankanas]

Hi, Thanks for your reply.

One very basic question is coming in my mind... What input voltage should I take for the test, Should I apply some DC input?

the second thing is that, are u mentioning the gain of 100 or 1000 in the feedback configuration as the noise gain?

Regards,
KS

[TimFox]

Look up "noise gain" in the literature (or internet) since it's a very important topic.
For gain 100 or 1000, there is very little difference between the closed-loop gain and the noise gain.
When measuring input offset voltage with such a gain configuration, using split +/-15 V supplies, you would normally connect the bottom end of the feedback network to the non-inverting input, and then connect that node to ground, unless you really need to measure the offset voltage at a certain common-mode voltage.
If you are using a single supply, then connect that node to an intermediate voltage (such as half the supply voltage) and measure the output voltage with respect to that node at the intermediate voltage.

[kankanas]

Thanks for the information. Certainly, I will look for the "noise gain" in detail.

As per the circuit configuration and measurement, is it the same as measuring the DC operating point of the ckt attached as an image?

Please correct me if I am wrong.

[TimFox]

Yes, that is a good circuit.  The physical op amp will have an input DC offset voltage, and you would short out the generator at the left.  It would be easier to use 10 ohms and 10 kohms  for the two resistors when measuring input offset.  When the parallel combination of those two resistors is large, you may see an extra voltage from the input bias current flowing through the combination (especially with a bipolar input op amp).  Note the terminology for currents:  “bias” means the current through one input, “offset” is the difference between the two currents (usually smaller).

[David Hess]

The simple way to measure input offset voltage is to configure the operational amplifier with a high fixed gain, typically 100 or 1000, apply zero volts, and then measure the output.  Now the input offset voltage is the output voltage divided by the gain.

Amplifiers do have "output offset" as a separate characteristic however usually input offset voltage dominates so it can be ignored.

[kankanas]

I was just wondering if this method works or not(Below attached images). I found somewhere about this test concept. V5 is the common-mode voltage to be considered if required. The result of Vos is |15.98uV|.

But still either of the cases(attached earlier and this) Vos is not following as per the datasheet.

Datasheet image is attached.

[David Hess]

I was just wondering if this method works or not(Below attached images). I found somewhere about this test concept. V5 is the common-mode voltage to be considered if required. The result of Vos is |15.98uV|.

It can work for zeroing the input offset voltage but not for measuring it because the open loop gain is both very high and poorly defined.