# Is 3k ohm too high impedance to an op amp differential amplifier?

I have a pressure sensor, MPX2100 and I need to read its output with an Arduino. To do so I want to use a differential amplifier with a gain of 40 times (3k and 120k resistors). The impedance of the sensor's output is 3k. So is this input impedance too high that I will get non-linearities in the output?

The amp op that I am going to use is LM258.

My schematic:

Even though a In-amp is better I assembled the circuit above just to test my sensor while I wait my AD623 to arrive, however the gain is about 3 times! When my sensor's output is 50mV the output of the amplifier is about 170mV!

Another odd thing that is happening is that when I connect the sensor to the op-amp the voltage at the sensor's output changes to -4mV. I am not sure whether my sensor is not working properly or the amplifier circuit is messing everything up.

• A schematic of the circuit you are proposing to use would help us understand exactly what you are doing. – The Photon Sep 24 '13 at 2:12
• 3K is not a high input impedance, and usually you want high input impedance. Unnecessarily high can be bad, but erring on the side of too high is better than too low. – Kaz Sep 25 '13 at 0:48

I'll assume you're using the common op-amp differential amplifier like this:

with R1 = 3 kOhm and R2 = 120 kOhm.

The input impedance of this circuit is approximately 2xR1, or 6 kOhm in your case. This won't necessarily cause nonlinearities but it will cause you to lose about 33% of your signal to the the voltage divider formed by the sensor's output impedance and the amplifier's input impedance.

If you are not tightly cost-constrained, consider using an instrumentation amplifier instead of an op-amp. In-amps are generally optimized for differential amplification with high accuracy at dc and high input impedance. TI has in-amps in their catalog with budgetary pricing in the $0.50 to$0.60 range, which probably means $1 -$2 in onesy-twosy quantities.

• Ditto on the instrumentation amp. You want the input impedance high. One prob w/ the differential amp is that the impedance of the two inputs is different. – Scott Seidman Sep 24 '13 at 2:30
• Is the AD623 a good choice? I don't have any symmetrical power supply. It's a "normal" 12V regulator, I also have got a 5V and 3.3V regulator. The output voltage cannot be above 3.3V as my arduino device runs at 3.3V. – Tiago Queiroz Sep 25 '13 at 0:38

The standard instrumentation amplifier topology was created for just this purpose.

It uses two op amps as infinite-impedance buffers and a third op amp as a fully-differential gain stage.

You can buy this configuration, prepackaged, off the shelf, for not much more than you're paying for your LM258.

• To add, you lose many of the benefits of this configuration if you build it yourself out of op amps. – Scott Seidman Sep 24 '13 at 13:24

Whatever amplifier configuration you choose, the specification for the pressure sensor recommends a typical operating voltage of 10V and this usually means that the standing offsets on each output are at the 5V level. This should tell you that your "front-end-amplifier" needs to run from the same supplies as your sensor.

This makes life easier because then you are almost guaranteed not to approach input common mode range end-stops no matter what the differential output voltage is of the device. It also tends to influence your decision as to which route to take.

Because of the standing 5V (half supply voltage) on each output a designer would tend to choose an instrumentation-amp (IA) because it fully "cancels" this voltage and it does not appear at the output pin PLUS, you get an offset pin on an IA which allows you to offset the output to (say) 2.5v making it convenient for whatever ADC you might be using.