# Differential Amplifier with voltage divider for reference voltage I'm trying to convert a 0-5 V voltage generated by a DAC to a -10 to 10 V range by using an LM324N OpAmp in differential amplifier configuration.

By choosing R1/R2 = R4/R3 = 4, I expect the output to be SIG1 = (DAC1 - VREF_2) * 4.

Therefore, I have to set VREF_2 to 2.5 V to achieve the desired output range of -10 to 10 V. I'm planning to use a simple 10k/10k voltage divider from +5V generated from an Arduino for this purpose. Does the voltage divider create the desired voltage reference or could there be an undesired output because of the subsequent differential amplifier?

• You should buffer your reference voltage with voltage follower for a low output impedance or you will have to take into account the equivalent resistance of your voltage divider. – Captainj2001 Nov 14 '19 at 21:13

Either buffer the reference voltage, OR:

Consider the Thevenin equivalent of your reference circuit. It's 2.5V with an impedance of 5K by inspection. If you added that to your diff amp you would have 15K in the non-inverting input branch and 10K in the inverting input branch.

That would destroy your common-mode rejection, so you would want to replace R2 with a 5K resistance (Maybe 2 10K resistors in parallel) to keep the CMR from degrading.

Note that the CMR degrades rapidly if the resistor matching isn't good. Off the top of my head I think the best you can do over tolerance with 1% resistors is about 48dB. So if you need better consider using 0.1% tolerance resistor.

Also, you have a 0.1uF cap across R4, which will ruin your AC common mode rejection. If you need AC CMR consider putting the same capacitance across R1, or removing the cap and filtering in a later stage.

• At the moment you have positive feedback. Swap the input connections over on the op amp. – James Nov 16 '19 at 10:44
• @James good catch, I missed that. – John D Nov 16 '19 at 19:32
• Awesome, thanks! So I could adjust the resistors or use a voltage follower. Would a voltage reference like the LM336-2.5v also work? I inserted the 0.1 uF Cap initially because of the MCP4922 datasheet but will remove it. I switched to 14k/56k resistors and 1k/1k as voltage divider for the moment and observed a distorted output compared to the expected (DAC1 - VREF_2) * 4, could this be the impact of a slight resistor mismatch? – tribol Nov 18 '19 at 20:13

Using E24 series resistors.... Note that the op amp's positive supply needs to be at least +12V in order to achieve up to 10V output.