I have a circuit that is supposed to measure a battery. It takes the form of a simple text book diff amp made from op amps. The picture is attached. The 10v in the schematic represents the battery, the 20 volt represents a voltage bias of the battery.

My question is does the bias (20v shown) affect the output? My impression is that the answer is yes. My impression is that the input terminals on the pins can only handle a voltage that is between the positive and negative rails, of the opamp supply. I've simulated this on two different simulators now.

Circuit maker seems to be ok with changing the 20 volt to even 1200volts. That seems ludicrous. Websim, digikeys on-line simulator, gives me the results I expect.

I don't want to be a cherry picker here. Does the bias(shown as 20v) matter? About when can I expect things to become undefined? My impression is about 36 volts, since thats the supply multiplied by 3 and the circuit attenuates by 3.

 

First the LMX324 is a 5 volt part. You can't run it at 12V. Even if you had a 12V rail to rail op-amp it would be operating at its maximum input voltage. I suggest you lower the gain and choose a different op-amp.

The voltage at the positive input should be 11.95Volts. Something is wrong with your simulation.

 

First the LMX324 is a 5 volt part. You can't run it at 12V. Even if you had a 12V rail to rail op-amp it would be operating at its maximum input voltage. I suggest you lower the gain and choose a different op-amp.

thanks for the reply. LMX324 was all I could find that I felt familiar with. It turns out its much different than the LM324. In reality the opamp is a TL074. Which is not available in all simulation programs. Can the bias(20v shown) be whatever with the TL074?

 

Regardless of the op-amp, per the resistances and voltage values on the schematic, the voltage at the positive input should be 11.95 volts. This is the same as your positive supply voltage, 12 volts. The voltage at the positive input MUST be below the supply voltage. Lowering the gain will fix this.

Also a TLo70 is an old op-amp. The inputs are limited to 2 volts less than the power supply voltage and the output only go to power supply voltage minus 2 volts.

 

This is what I don't get. The simulation is shown, with a different part and different bias. It's giving the proper output voltage and the bias is 120v. It's showing 36.8volts sitting comfortably on the input terminals. It that possible? On the data sheet the input voltage range is given as 16v. Does the circuit work like this in reality or is it a simulation problem?

 

The LMX324 using a single +2.3V to +7V Supply Voltage Range.
If you still want to using this op amp, then you should using the voltage divider to reducing the input voltage first, and lower the 12V to 6V for power supply or using a resistor to limiting the current to <540uA.

 

here is another one. The simulator is reporting everything is going fine and there is 304 volts on the +, - terminals. I want to believe myself and lestraveled, so are the simulators just not reliable in this scenario?

 

The LF356 can not function when the voltage at the inputs are greater than the supply voltage.

 

Why don't you try a 10000 volt bias.

 

I think you should try another simulator. LTspice is a favorite.

 

My boss actually tried to connect hundreds of volts to the circuit shown. Two of the circuits existed on the board, and he connected two batteries 100 volts apart. It exploded in his face. I tried to tell him before the board was designed that opamps don't work like he thinks. He thinks they are like magic dc transformers, and only the difference matters and nothing else is relevant. He even thinks that opamps provide isolation. The burned up board is attached.

Another 3 month design cycle down the drain ... or rather ... up in smoke

 

here is another one. The simulator is reporting everything is going fine and there is 304 volts on the +, - terminals. I want to believe myself and lestraveled, so are the simulators just not reliable in this scenario?

Simulators are only as reliable as the models you feed them. The common opamp, transistor, and diode models do not model junction voltage breakdown, or other voltage limits.

What I do is to model a circuit I plan to build using an "ideal" op-amp model before picking a specific opamp. With the supplies applied to the circuit, a .DC or .TRAN solution will quickly show the peak voltages/currents/power dissipation that the real op-amp or other part has to contend with.

Using a simulator doesn't obviate the need to have the data sheet for the parts in your circuit in front of you (and know how to read it).

 

To check my post on #6 again, if you know the rules, voltage is not the problem.