1
\$\begingroup\$

I am trying to design a voltage follower to drive a resistive load using MOSFET as a peripheral circuit. I am using this simple circuit using two NMOS transistors. The output is following the input. But the problem is output is becoming load-dependent. If I change the resistance of the load, I am getting an increase/decrease in output voltage. How can I solve this problem? Is there any CMOS based stable voltage follower/Unity gain buffer circuit which can be used to drive a resistive load?

schematic

simulate this circuit – Schematic created using CircuitLab

\$\endgroup\$
4
  • 1
    \$\begingroup\$ If you incorporate an op-amp and take feedback from Vout you can make it basically load independent. \$\endgroup\$
    – user57037
    Nov 20, 2020 at 22:08
  • \$\begingroup\$ @mkeith Thank you for your comment. I have area and power constraints so I cannot use op-amp. Are there any other ways or any other CMOS based circuits? \$\endgroup\$
    – Apprentice
    Nov 20, 2020 at 22:30
  • 1
    \$\begingroup\$ If you can't use an opamp, then a single NPN will give better results than a single MOSFET. \$\endgroup\$
    – Mattman944
    Nov 20, 2020 at 23:58
  • \$\begingroup\$ @Mattman944 Thank you. I will check if this works. \$\endgroup\$
    – Apprentice
    Nov 22, 2020 at 20:54

1 Answer 1

4
\$\begingroup\$

You can use an op-amp to provide feedback control to mostly eliminate the load dependence:

schematic

simulate this circuit – Schematic created using CircuitLab

In comments you said the reason you have not considered using an op-amp is "I have area and power constraints". But your proposed solution is to use a second MOSFET.

Very likely you can find an op-amp that uses just as little area and less power than any solution that involves adding a second MOSFET.

For example, Mouser lists 215 op-amps in SC70-5 package (2.0 x 1.25 mm), with typical supply current as low as 350 nA (and almost all below 1 mA). Other options are available in packages (mostly "wafer-level" and "chip-scale" types) smaller than 1 x 1 mm.

The best choice of op-amp will depend on the required bandwidth and Vdd voltage, which you haven't shared.

\$\endgroup\$
8
  • 1
    \$\begingroup\$ Upvoted because I agree 100%. Packaged op-amps can be just as small as packaged transistors, because it is the package, not the die that is limiting size. I do suggest a series resistor to the gate, and a feedback capacitor from the op-amp output (which may not be needed, but...). \$\endgroup\$
    – user57037
    Nov 21, 2020 at 0:59
  • \$\begingroup\$ @mkeith, like so? \$\endgroup\$
    – The Photon
    Nov 21, 2020 at 1:02
  • \$\begingroup\$ Yeah but now that I see it I realize that the cap can't do much because vout will likely be low impedance. The cap will be "swamped". Maybe add a resistor between Vout and the bottom of the cap to allow it to have some effect. The concern is that the op-amp may not be stable if it takes its feedback from the transistor output. Taking some feedback directly from the op-amp output was intended to stabilize it (I mean that was my motivation). \$\endgroup\$
    – user57037
    Nov 21, 2020 at 1:05
  • \$\begingroup\$ @mkeith, careful, though. If we add too many extra parts OP's concerns about board area might start to gain some validity. \$\endgroup\$
    – The Photon
    Nov 21, 2020 at 1:07
  • \$\begingroup\$ Yes. Maybe better to just go to BJT instead of mosfet and get rid of the capacitor. I think the BJT will make a nicer load for the op-amp rather than the possibly high capacitance gate of the MOSFET. \$\endgroup\$
    – user57037
    Nov 21, 2020 at 1:41

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.