0
\$\begingroup\$

Consider a simple two stage CMOS operational amplifier. When it's designed for given specification and is being used as an Op-amp block, doesn't its bias gets disturbed when any arbitrary voltage is applied to inputs by the user?

Say if one of the terminals is grounded, wouldn't the differential amplifier inside be off?

enter image description here

\$\endgroup\$
1
\$\begingroup\$

The configuration of op amp you have drawn would certainly be unable to properly operate at the most negative end of the power rails. This is why a lot of op amp applications use a proper negative rail hence 0 volts (ground) is somewhere close to the midpoint between supplies.

The circuit you have drawn would work with inputs close to (or slightly above) the positive rail of course but. if you want an op-amp that can operate with inputs across the whole supply range you need a far more complex input circuit.

If you look at this BJT op amp you can see the complexity: -

enter image description here

There are two diff input pairs - one formed by Q1 and Q2 and another formed by Q3 and Q4 - at somewhere mid supply one pair takes over from the other pair thus allowing the full input range to be resolved. Note that the circuit above is also rail-to-rail output too. The diagram is of an LT6220

\$\endgroup\$
  • \$\begingroup\$ Thanks. But isn't it necessary for the differential pair to have a DC bias? Users could apply any voltage to the inputs +IN and -IN (within the ICMR) without knowing the internal circuitry. If there is no DC bias given to differential pair, they might not operate in saturation region(MOSFETS) or active region(BJTs). \$\endgroup\$ – Aditya Patil Sep 26 '15 at 10:23
  • \$\begingroup\$ The DC bias comes from the input signals for absolutely every op-amp ever created. \$\endgroup\$ – Andy aka Sep 26 '15 at 10:27
  • \$\begingroup\$ What if the input is a very small signal(Like a sensor output in mV)? How do you expect the mosfets to be biased properly? \$\endgroup\$ – Aditya Patil Sep 26 '15 at 11:46
  • 1
    \$\begingroup\$ @AdityaPatil The supplies of an op-amp for that sort of application will normally be +/- X volts relative to 0V. The sensor output might be at 0V for no output signal therefore the op-amp is appropriately biased exactly half way between + and - outputs. It sounds like you need to be doing rudimentary research about op-amps before jumping into the make-up of an op-amp. \$\endgroup\$ – Andy aka Sep 26 '15 at 14:28
  • 1
    \$\begingroup\$ The DC level on both signals has to be somewhere around mid-rail between gnd and Vdd. \$\endgroup\$ – Andy aka Sep 26 '15 at 15:37

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.