OPA4322 TINA model clamping at 100mV

Yao,

Depending on the circuit configuration used, input common mode voltage, Vcm, or output voltage swing, Vo, will be the limiting factor. If you use OPA4322 in a follower configuration, it is not Vcm but the output will be clamped 100 mV from either rail - please see PDS for conditions used for open-loop gain, AOL: 100mV<Vo<(V+)-100mV.

Hi Marek,

if simulated in integrator configuration, Vout is clamped/clipped in 100mV. is it normal? 

(measurement would not be clamped by 100mV)

cheers,

Yao,

OPA4322 output linear range, as shown in the condition column of AOL test (AOL=ΔVout/ΔVos), is defined as 100mV<Vout<(V+)-100mV. Thus the simulation correctly shows the output being clamped 100mV from either rail - see below.  

Hi Marek,

as I simulated with TINA simulator and measured from the scope. I got different results.

which simulation keeps clipping at 100 mV, while couldn't be observed in measurement.

how do you think about this?

Yao,

Your simulation is in agreement with OPA4322 datasheet - the Vout clamps around 100mV from either rail. You may not see this in your measement because OPA4322 100mV clamping from the rail is a DC measurement while you trying to do it dynamically - slow it down and use the same scale for Vin and Vout. Also, you must realize that the actual clamping might be much softer than the simulation and for that reason you must monitor Vos to detect it.

Merak,

thanks for your explanation.

1. So you meant that I could try with the input Vin, which would result in clamping,  longer to see if such clamping did occur.

2. what do mean by monitoring the Vos in measurement?

is my understanding right?

yao

Yao,

See my comments below:

1. So you meant that I could try with the input Vin, which would result in clamping, longer to see if such clamping did occur.

Yes, apply 3V DC to the input and measure the output voltage with respect to GND.

2. what do mean by monitoring the Vos in measurement?

Since the output will get much closer to GND (~10mV) than the linear AOL condiition calls for (100mV), in order to see the effects of gradual trioding of the output stage you must measure the change of the input voltage offset, Vos, as the output voltage is driven closer and closer to negative rail.

Thank you very much Marek, this help me clarify a lot.

but for another thing. is it possible to provide a model to simulate the dynamical behavior to fit with the measurement(vout could reach 10 mV)?

i know sometimes the datasheet would set in the stringer requirements in case of the worse performance.

thank you,

yao

Yao,

In order to make the model match the non-linear performance of the output stage, this would require using actual transistors for macro-modeling that makes the circuit much more complex and leads to convergence problems in simulations.  Since our modern macro-models are purely behavioural (no single transistor inside) they cannot match the non-linear op amp performance.

Also, since the process variation results in changes in the behaviour of the circuit especially in non-linear regions of operation, using such macro-models would lead to unpleasent surprises in actual performance of the circuit.  For that reason the models match only the linear performance of the IC (e.g. Vout condition in AOL).