I am scaling a 10.64 Vp-p (+5.32 to -5.32) signal to ~4.6 Vp-p (0 to 4.6) using an OPA4192 as a differential opamp. The feedback configuration results in attenuation (less than unity gain.) The OPA4192 is a rail to rail opamp. The common mode input voltage and output is within the specified range in the datasheet.

The issue is that when the output goes below 0.5 V it is clipped and the clipped part appears at the inverting input of the opamp (both inputs should be at same potential {virtual ground.}) When I use the same opamp with feedback such that gain=>1 the output goes near 0 V also when it is used as a buffer or comparator.

What am I doing wrong?

The op amp is used with a single supply and the dc offset at non inverting input is 1.6 V to keep the output above or near (0.03 V) ground. Please look at the pictures attached:

  1. Schematic.
  2. Real time output of op amp on oscilloscope.
  3. Real time inverting (yellow) and non inverting (green) input of op amp on oscilloscope.

edit:I am sorry I uploaded the wrong schematic with very low resolution. Also the input would be DC or less than 15Hz signal.




  • \$\begingroup\$ Even at ideal conditions as far as i can calculate you will get clipping when input sine (at V_ctrl?) reaches more than 4.8 V with respect to ground (assuming pot1 is short and you have 1.6 V at noninverting input of u1a). \$\endgroup\$ Commented May 15, 2023 at 9:02
  • \$\begingroup\$ Are you sure R2 is 10k? Where did you get the op amps from? \$\endgroup\$ Commented May 15, 2023 at 9:58
  • \$\begingroup\$ I'd be interested to see the 12V power supply voltage trace, with Vctrl, and to see if anything changes if D1 and C9 are removed. \$\endgroup\$ Commented May 15, 2023 at 11:36
  • \$\begingroup\$ @MakePlatanGreatAgain the datasheet specifies under no load condition to output reaches near zero i would have no worries even if it clips at 100mV but 500mV is too much and doesnt make any sense to me because the output reaches zero when gain >= 1 \$\endgroup\$ Commented May 16, 2023 at 10:11
  • \$\begingroup\$ @AliSiddique Hi, i made that assumption on your previous scheme, which was uploaded wrong as you mentioned, also it is shown in LTspice simulation from Rohat Kılıç \$\endgroup\$ Commented May 16, 2023 at 14:32

1 Answer 1


and the dc offset at non inverting input is 1.6 V to keep the output above or near (0.03 V) ground.

From the values shown in the schematic it should be 0.8V (assuming the pot is at mid):

$$ \mathrm{V_{off}=\frac{1.6}{8.3+1.6} \ 5=0.81V} $$

And this is not the value you'll get as the output offset. With your configuration it might be difficult to adjust the output offset. I simulated your circuit with AD822 which is a R2R CMOS op amp, played with values and here's what I ended up with (Designators don't match):

enter image description here

The output has 1.64V offset and the peaks at 3.4V when the input is 5Vp.

I'd go for AC coupled non-inverting amplifier instead:


simulate this circuit – Schematic created using CircuitLab

R2-R3 pair brings a constant offset of 1.7V. The non-inv amplifier with equal gain setting resistor will bring a voltage amplification of 2. CC2 blocks the DC so the DC offset won't be amplified. This means that the output will have 1.7V offset.

You want the ±5V input signal to be converted to 1.7 ± 1.7V signal. Since the AC amplification factor is 2 (1+R4/R5) the input signal should be attenuated to 0.85V i.e. a factor of 5.88. At AC R2 and R3 will be paralleled so their equivalent will be 6.67k. R1 and this equivalent will form an attenuator. So R1 should be (5.88-1) time R2-R3 equivalent: \$R1 = 6.67k (5.88-1)\approx 33k\$.

NOTE: CC1 and CC2 should be selected for the lowest frequency. For the circuit above the minimum frequency should be around 20 Hz.

  • \$\begingroup\$ Kilic thankyou for the effort 1) Sorry i uploaded the wrong schematic i have updated it if you could give it a look one more time please. 2 )I simulated OP234 in proteus which is also R2R with the same configuration and it was working perfectly fine. I could not find OPA4192 for proteus simulation. 3) I forgot to mention the input would be DC or very slowly changing frequency not more than 15Hz i am certain it would be a DC signal so i dont want to design it for an AC. 4) I have been asked to attenuate the signal with a feedback config and not use a voltage divider at the input. \$\endgroup\$ Commented May 17, 2023 at 5:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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