2
\$\begingroup\$

The schematics: enter image description here

It should convert 0-3.3V input from the DAC to -4.1 + 4.1V (roughly)

The positive part works great, the negative output voltage saturates at -2.95V.

Is it ok? The simulation looks ok (and it is the expected behavior): enter image description here

The part: https://www.st.com/content/st_com/en/products/amplifiers-and-comparators/operational-amplifiers-op-amps/standard-op-amps/tl084c.html#design-scroll

\$\endgroup\$
5
\$\begingroup\$

the negative output voltage saturates at -2.95V.

Yup. If you look at Figure 2 of the data sheet, it shows a nominal maximum for +/- 5 volt supplies to be 6 volts pk-pk. In other words, about +/- 3 volts, and you're getting -3. Just about what you'd expect.

The positive side is doing better than expected, but that's just an unexpected bonus.

TL;DR - You should never count on getting +/- 4.1 with your power supply and this op amp.

\$\endgroup\$
  • \$\begingroup\$ Yes . Thanks. The positive part is actually described in the DS as almost RR. So the output will be -2.5 to 2.5V \$\endgroup\$ – P__J__ Jun 23 '18 at 23:26
4
\$\begingroup\$

You should not only rely on simulations only1. Check the datasheet.

That shows:
On a +/-15V supply the output swing is between +/-10 and +/-12.

For running on +/5V select a low-voltage rail-to-rail opamp.

1In LT spice I recently found an opamp swinging between +/-220V on a +/-20 V supply

\$\endgroup\$
  • \$\begingroup\$ I took the model from the STM website :) \$\endgroup\$ – P__J__ Jun 23 '18 at 21:03
  • 1
    \$\begingroup\$ @PeterJ_01 They usually still don’t behave correctly at clipping or even close to the rails. \$\endgroup\$ – winny Jun 23 '18 at 21:51
  • \$\begingroup\$ There are lies, damn lies, and SPICE models... -- Disgruntled Engineer \$\endgroup\$ – Richard the Spacecat Jun 23 '18 at 22:27
  • 1
    \$\begingroup\$ Spice is reasonably good at telling you something will NOT work, but take anything it says works with a pinch of salt. In particular spice opamps are all too often really three terminal models (and real ones have at least 5 pins, which should hint at the problem), any model that references spice node 0 should get the stink eye. \$\endgroup\$ – Dan Mills Jun 23 '18 at 22:51

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.