Lm258 non-linear behaviour in differential amplifier configuration

Samuel,

When both inputs exceed upper common mode range the input PNP transistor cut off. Once this happen the device is no longer controlled bu the inputs and the device it will its own thing. Usually the output goes high, but it is also possible that it would go low. 

Samuel,

Spice model show normal behavior well but out of range results don't always simulate correctly.

Hi Ron,

Thanks for your reply. I understand that the non-linear behaviour of an op-amp cannot be simulated correctly/ explained easily. Like you said, the opamp output did go high in this case.

I imagine the correct way to solve this issue is to ensure that the common mode voltage goes down at the same time as the supply voltage to the opamp and ensure that the common mode range is never violated.

Thanks

Samuel

Samuel,

That could work or you could could detect that 24V supply is incorrect and mask the output as invalid or flag a error that overrides the current measurement error.

I suspect that you are getting an error signal during power down that is causing a problem.

Hi Ron,

Yes, you are right. The problem appears during power down. As the supply of the opamp drops, the common terminal is still at the same level (~15V). The concern that I have is that during the shutdown, the output shoots to 14V and that output is connected to an ADC pin. There is a protection diode to the 3V3 rail and there is a 3.6V zener that clamps the voltage to approx. 3.8V (given the range of the zener).

The concern is that voltage is still being seen by the ADC that may cause an issue over time. I will have to think of other way to disable the 15V regulator at the same time or before the supply drops to a certain voltage.

Would it helps if I used different opamp for this circuit?

Thanks for your help Ron

Samuel

Samuel,

Another op amp could work, but I haven't seen any data sheets that guarantee an low output low when common mode range is exceeded.
Could you add an open collector/drain comparator output to the LM258 non inverting input that could pull it low when 24V is falling?

Hi Ron,

Thanks for your reply. I managed to control the terminal voltage by shutting down the 15V regulator around 22V input (i.e. when supply to LM258 is 22V, the 15V regulator has been shut off). However, I still have the same problem where the output still goes high, mimicking the 15V rail even though the supply is higher.

Attached scope plot shows the behaviour:

CH3 (purple) is scoped at the output of LM258. If you look at the time when the output jumps high, the supply voltage is approx. 20V (CH4 - dark blue) and the 15V rail is approx. 14.8V (CH1 - yellow).

Here is another scope plot of showing the same behaviour of the opamp:

CH1-yellow is 15V the input terminal voltage; CH2 - light blue is Vin to the board; CH3 - purple is opamp's output; CH4-dark blue is opamp supply
All 4 plots on the same offset. As you can see the supply is higher than the input terminal voltage but the opamp output still goes high, following the input terminal voltage.

I also tried rail-to-rail opamp that can handle the common mode voltage, LM7322 and supply it with the same terminal voltage (i.e. 15V rail) but the issue still persists:

CH1-Yellow (15V rail: opamp supply and terminal voltage), CH2 - Green (opamp's output).

This makes me wonder that the issue is something else than the common mode input range, which was thought to be the issue initially. Would you have any idea on where to debug?

Regards

Samuel

Samuel,

The output seems to go high a little sooner than I would expect. However, I don't think any new effect is happening.
Because the high pulse is shorter, I think you are getting closer to a solution.