0
\$\begingroup\$

I am using a sensor having high impedance and have been suggested by the manufacturer to use a buffer (unity gain buffer) so that when we measure the values on ADC, they don't vary.

Now i am using LM324 Quadruple Operational Amplifier as a unity gain amplifier using this circuitenter image description here and this link page 9 (but not using RL and CL).

Input on the V + of the circuit is 1.37 but the output V out is coming to be 1.41 or 1.42 instead of 1.37. I am not able to find out why this gain is coming and how to resolve this.

\$\endgroup\$
  • 1
    \$\begingroup\$ Have you looked with an oscilloscope? Is there any noise/oscillation/interference on the op-amp output pin? Did you bypass your power supplies? \$\endgroup\$ – The Photon Apr 21 '17 at 17:02
  • \$\begingroup\$ No, we have used an Oscilloscope. I tried using a 0.1uf capacitor between O/P and GND and 4.7 uF capacitor between +5VCC and GND, but that leads to much unstable readings which were stable earlier. \$\endgroup\$ – srj0408 Apr 21 '17 at 17:14
  • 2
    \$\begingroup\$ Don't add capacitance on the output of the op-amp. That can lead to oscillation. But capacitance on the supply pins can reduce interference from ripple or noise on the power supply net. \$\endgroup\$ – The Photon Apr 21 '17 at 17:18
  • \$\begingroup\$ Not addressing your offset problem (which is likely a voltmeter loading problem), but be aware that with a +5v supply, this LM324 doesn't pull its output up very easily above +3V. So if your ADC is using +5V as its reference, the top 200 of 1024 (approximately) of your ADC range may be non-linear. \$\endgroup\$ – glen_geek Apr 21 '17 at 17:20
  • \$\begingroup\$ @glen_geek : I am using +2.56 as ADC reference, as the voltage levels of the sensor never goes above 1.8. \$\endgroup\$ – srj0408 Apr 21 '17 at 17:22
1
\$\begingroup\$

This is not necessarily a gain error but an offset error. It could also be an error in measurement.

The output voltage will not be exactly the same as the input but who'll be within a few millivolts - your error seems to be larger than that.

To determine the gain you need to measure the output at at least two input voltages so you can determine the actual gains and offset. Since you are using a single supply amplifier don't use zero volts as the opamp isn't specified to work with the output lower than 20mV (page 5 of data sheet).

The relevant parameter in the data sheet is the "Input offset voltage" on page 5. For this device it should only be 9mV max over the full temperature range.

How are you measuring the voltage? What meter are you using? What are you supplying the input voltage from?

Ideally you should measure both the input and output at the same time. If for example you are feeding the input from a source that has a high output resistance the act of putting the meter on that point could cause it to drop. When you remove the meter it then increases by a few millivolts - the LM324 has a bias current of only ~20nA, less than the 100's of nA or microamps of the meter.

\$\endgroup\$
  • \$\begingroup\$ Voltage is provided by a sensor as Analog O/P and then it is fed to LM324. From LM324 is it fed to 10 bit ADC of a micro-controller. I am supplying voltage to the Sensor as well as LM324 via a external 5 V circuit with 1 Amp current capacity. \$\endgroup\$ – srj0408 Apr 21 '17 at 17:08
  • \$\begingroup\$ What sensor? What meter? Do you have two meters so you can measure both input and output simultaneously? \$\endgroup\$ – Kevin White Apr 21 '17 at 17:12
  • \$\begingroup\$ I am trying to work with SPEC sensor and they suggested to use two Amplifier on each of the output. Their recommendation was to use this Op-Amp but we used LM324 because of its availability. Also, i can definitely measure both I/P and O/P, its just adding one more ADC channel. \$\endgroup\$ – srj0408 Apr 21 '17 at 17:20
  • \$\begingroup\$ I would recommend using a DVM to measure the voltages - is that what you have been doing? \$\endgroup\$ – Kevin White Apr 21 '17 at 17:31
  • \$\begingroup\$ Yes, i used a fluke 101 DVM to measure the voltages. \$\endgroup\$ – srj0408 Apr 21 '17 at 18:01
1
\$\begingroup\$

If the manufacturer says you need a high impedance buffer, I would assume your act of measuring the voltage is in fact changing the value.

You can test this by one of two methods.

  1. Get two meters, check they give equivalent readings on the output, then connect one to the input while maintaining the other on the output. If the meter is affecting the voltage then both meters should now read close to the same lower voltage.

  2. Disconnect the sensor and apply a low impedance test source. Verify the output now matches the input voltage.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.