Op-Amps Keep Railing

[Sirius24]

I am working on a circuit with someone and our op-amps keep railing. Even when just working with the op-amp, on its own, with an inverting configuration in negative feedback, everything rails. From my understanding, there should be a gain equal to the feedbak resistance divided by the input resistance, but I am not seeing this. It always rails. Does anyone know why it might do this?

[berkeman]

I am working on a circuit with someone and our op-amps keep railing. Even when just working with the op-amp, on its own, with an inverting configuration in negative feedback, everything rails. From my understanding, there should be a gain equal to the feedbak resistance divided by the input resistance, but I am not seeing this. It always rails. Does anyone know why it might do this?

Can you post some of your circuit configurations? Which opamp, and what supply rails? If you are using split power supply rails (like +/-12V), what does the output do when you hook up the opamp in a buffer configuration (- input = output) with the + input grounded?

[Sirius24]

Can you post some of your circuit configurations? Which opamp, and what supply rails? If you are using split power supply rails (like +/-12V), what does the output do when you hook up the opamp in a buffer configuration (- input = output) with the + input grounded?

The op-amp is an LM358 with rails at 0 V and 12 V.
With an any sort of input, I am receiving a 10.9 V output.

[berkeman]

The op-amp is an LM358 with rails at 0 V and 12 V.
With an any sort of input, I am receiving a 10.9 V output.

If you want to use it in single supply operation like that, you will need to use a non-inverting configuration. You are trying to run an inverting amp above ground with positive input voltages? What happens if you set Vin to -0.1V?

[Sirius24]

With all the lines we have going everywhere, it would be difficult to move things to test at -0.1V, but switching from inverting to non-inverting fixed the problem. Thank you.

~Sirius

[Sirius24]

With all the lines we have going everywhere, it would be difficult to move things to test at -0.1V, but switching from inverting to non-inverting fixed the problem. Thank you.

~Sirius

Out of curiosity, though. Why can't an inverting case work for 0 to positive rails?

[vk6kro]

The non-inverting input should go to a "virtual ground" made from two identical series resistors connected across the 12 V power source.

The bottom resistor should be bypassed and the + input connected to the 6 V centre point of the two resistors.

4.7 K resistors would be OK as a start.

http://dl.dropbox.com/u/4222062/opamp%20bias.PNG

If you want the input to include zero volts, there will need to be a negative as well as a positive supply.

This is just because of the internal construction of the opamp. The inputs and output connect to transistors which have voltage drops across them.

[Sirius24]

Would it be possible to make a differential amplifier go from ground to positive rail by putting V-at zero and making absolute ground into virtual ground?

[berkeman]

Would it be possible to make a differential amplifier go from ground to positive rail by putting V-at zero and making absolute ground into virtual ground?

What are the range of voltages for V1 and V2 where that might work? Refer to vk6kro's post...

[vk6kro]

Would it be possible to make a differential amplifier go from ground to positive rail by putting V-at zero and making absolute ground into virtual ground?

You would have to look at the internal circuitry of the amplifier.

If an input results in a transistor's base-emitter junction being reverse biased, or not biased at all, then the output will certainly be unacceptable.

Usually this happens if the input is within 0.5 volts of the negative supply, although certain opamps ( like the LM324) have special circuitry that does allow the inputs to become zero or even slightly negative even with single voltage power supplies.

[sophiecentaur]

When you realise what a differential op amp actually does, then then vk6ro's cirircuit modification says it all. It amplifies the difference between the input voltage on the - terminal and the 'reference voltage', on the + terminal (which has been given a value of half rail volts). This is just what you want: an inverting amplifier which will give you the gain equal to the ratio of feedback resistor to the input circuit resistance.
There's no need to discuss the minutiae of the internal circuitry of an individual chip in order to get a working circuit. An op amp does exactly what it says on the tin.