An n-channel MOSFET is an inexpensive way to connect and disconnect circuits to 0V. But if I have a fixed voltage (in this case, 2.5V) and want to connect fixed resistances of higher or lower voltage, the diode in the MOSFET will interfere?

There are analog switches, but they are expensive by comparison. I am looking for the equivalent of an inexpensive MOSFET that is 1Mohm when the gate is 0V, 0.5ohm or less when the gate is 5V (or 10V) above the reference voltage, but I want to handle voltages both higher and lower than the reference voltage.

The power in the circuit is tiny, in the milliwatt range, so dissipation is not an issue.

Is there any such device?

Here is an op amp circuit. It will allow variable gain depending on the whether the gates of M1 and M2 are high or low. This works when the resistors are connected to ground, and it should work when the resistors are connected to 2.5V as shown, as long as Vin is higher than 2.5V. I would like to know what to do if Vin is less than 2.5V.

enter image description here

  • 6
    \$\begingroup\$ A common solution to this dilemma is to use 2 MOSFETS back-to-back in series, so that their diodes point in opposite directions. \$\endgroup\$ – brhans Oct 13 '16 at 13:36
  • \$\begingroup\$ Have you considered JFETs? \$\endgroup\$ – Ignacio Vazquez-Abrams Oct 13 '16 at 13:41
  • 2
    \$\begingroup\$ Why don't you describe your application more. There is probably a better solution than the one you have constrained yourself to. \$\endgroup\$ – Daniel Oct 13 '16 at 14:45
  • 1
    \$\begingroup\$ It looks like OP is trying to built a variable-gain amplifier with binary-code control, and looking for analog switches. Next questions will be about parasitic capacitance... I just don't get the cost considerations... \$\endgroup\$ – Ale..chenski Oct 13 '16 at 15:51
  • \$\begingroup\$ Hmm. A very 'beefy' 4066 would be in order. Too bad 4-term discrete mosfets are unobtainium. \$\endgroup\$ – jonk Oct 13 '16 at 18:32

HEF4066B <200 Ω @12V only cost $0.15usd/1k


simulate this circuit – Schematic created using CircuitLab

enter image description here


You can use a solid state relay.

TLP222AF (2 ohms, $1.08)

VO14642AT (1/4 ohm, $2.66)

Or use a regular mechanical relay.
Coto 9007-05-00 (0.2 ohms, $1.27)

  • \$\begingroup\$ \$2\:\Omega\$ from datasheet. OP says \$\le \tfrac{1}{2}\:\Omega\$, as I read it. \$\endgroup\$ – jonk Oct 13 '16 at 19:24
  • \$\begingroup\$ @jonk you are correct. I added some other parts that meet the 1/2 ohm spec. \$\endgroup\$ – user4574 Oct 13 '16 at 20:53
  • \$\begingroup\$ Sometimes what the OP wants is not what they need \$\endgroup\$ – Sunnyskyguy EE75 Oct 13 '16 at 21:53
  • \$\begingroup\$ A relay? I want to change amplification more than 10 times/second. I liked the 2n7000 @ .05, so I'm not in love with these solutions, but thanks. \$\endgroup\$ – Dov Oct 14 '16 at 1:33

If you need a really good DC isolation from any offsets, and speed of switching is not important, you can use "reed relays", like this one, and this one. These devices will give you parasitic capacitances under 1-2pF and will provide good bandwidth for variable-gain designs.


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.