# Difference amplifier error

Hi all,
I am currently investigating a circuit I built which uses a difference amplifier in order to amplify the difference between two signals. One is a reference signal of 4.5V and the other is a varying signal which oscillates around 4.5V. Ideally when my varying signal is 4.5V, the output of the amplifer should be 4.5V.
I have simulated two op amps, one is the LT1013 and the other is the TL082. I have compared these to the ideal op-amp model. The gain of the difference amplifier is 22/4.7 = 4.68.

However in my simulations I see an error term appear. This error term is larger for the TL082 circuit than when using the LT1013 circuit. It is zero for the ideal op-amp model.

LT1013 = 0.069V
TL082 = 1.553V
ideal = 0V

I have read that error is introduced by low CMRR (in dB). My questions are:

• How can I calculate what the error will be without simulation?
• Are there any other factors (other than CMRR) that can influence this error?
• What are important parameters to look for in the datasheet regarding this error term?

Is there any experienced designer out there who can help me understand this better?

One other factor - which might be assumed to be 0 in the model is input offset voltage and current.

The TL082 is not rail-to-rail at all, you'd need a -5V and a +9V supply to get it to work at all. It also has a very large offset voltage of 15mV, not suitable for precision.

The LT1013 is not rail-to-rail, it does go down to the negative supply, but from a 5V setup you are limited to the range 0.0 to 3.5V approximately. It is genuinely precision with 250uV max offset.

CMMR is not going to give you the kind of errors you see, that's entirely due to trying to output outside the working range.

BTW the differential amplifier circuits you give will try to output centred on 0V, as the reference input (ie the 22k resistor currently drawn to ground) is at 0V.

One is a reference signal of 4.5V and the other is a varying signal which oscillates around 4.5V. Ideally when my varying signal is 4.5V, the output of the amplifer should be 4.5V.

That doesn't make sense. Since the difference is zero the output should be (ideally) zero no matter what the gain.

I have simulated two op amps, one is the LT1013 and the other is the TL082.

As mark says, your op-amps-aren't rail-to-rail so you can't get to zero without a negative power supply... Power the op-amps from +9V & -9V and the simulation should work. Most op-amp circuits use dual power supplies (although there are exceptions).

(I'm not sure that a "rail-to-rail" op-amp will go exactly to zero or if it will just get very close.)

I am currently investigating a circuit I built which uses a difference amplifier in order to amplify the difference between two signals. One is a reference signal of 4.5V and the other is a varying signal which oscillates around 4.5V. Ideally when my varying signal is 4.5V, the output of the amplifer should be 4.5V.

V1 : *One is a reference signal of 4.5V *

V2: *the other is a varying signal which oscillates around 4.5V. *

*Ideally when my varying signal is 4.5Vthe output of the amplifer should be 4.5V *

ideal = 0V (this is my experience using a real world op amp (LT1215). Admittedly , the difference wasn't EXACTLY 0, but it was a very small voltage.

A difference amplifier is a subtractor. (V2-V1)

If both are 4.5V then V2-V1 =4.5-4.5 = 0.

If you are not getting close to zero then your offset or something else is messed up.

I have used the LT1215 (the Instrumentation Amp circuit on the first page and it behaves as described.

DVDdoug: That doesn't make sense. Since the difference is zero the output should be (ideally) zero no matter what the gain.

As mark says, your op-amps-aren't rail-to-rail so you can't get to zero without a negative power supply... Power the op-amps from +9V & -9V and the simulation should work. Most op-amp circuits use dual power supplies (although there are exceptions).

(I'm not sure that a "rail-to-rail" op-amp will go exactly to zero or if it will just get very close.)

These days most opamp circuits do not use dual supplies. This is partly why most low voltage opamps handle inputs down to below the negative supply and have outputs capable of driving to the negative supply.

Hi all, thanks for your replies! Yes there was a type, I meant when both inputs are 4.5V we ideally want an output of zero. I think that the non- rail-to-rail outputs of my amplifiers are causing these problems! Thanks for your help. I will make the changes and post the results...

My experience with the LT1215 is the output is pretty close to zero if both inputs of the subtractor are the same.