This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Maintain Input impedance when Amplified is not powered

Other Parts Discussed in Thread: LMC8101, OPA2341, LMC6001, LPV521, LPV531, LMC6442, LMV341-N, LMV601


I am using LMC6572 op amps, and need to maintain low input current ( <1pA) even when the circuit is not powered.  As it is now, current flows through the IN+ pin to the supply and output pins when power is off.


  • Hello Ord,

    This will probably not be possible.

    Mostly, what you are seeing is the leakage of the ESD diodes to the power supplies (which are probably at ground). Normally, when the power is on, the upper and lower ESD diodes "leak" identically, and cancel each other out. With no current flowing, you are only going through one diode and see all the leakage of that diode. The leakage path is through the ESD diode to the supply pins.

    The only thing that would not leak with no power applied is a relay. Even CMOS "switches" or muxes would leak with no power (same problem as the amp). You may be able to use a very low leakage discrete MOSFET to disconnect the input - but that may complicate or interfere with your circuit operation.

    What you may want is a device with a shutdown pin, which brings the supply current into the nano-amps. Shutdown turns off the internal stage biasing to drop the supply current. The input leakage may go up a little, but will not be as bad as an un-powered device. But you would need to provide "standby" power - but at a very low current (a few uA's or nA's -use a super cap?). Unfortunately, there are no dual LMC's with shutdown. The only LMC with a shutdown is the single LMC8101 (which is a relative of the LMC6572/6482 family). The OPA2341, 2363 and 2373 are CMOS R-R Input devices with shutdown.

    If we knew more about your application, we may be able to offer more solutions.


  • Thanks Paul,

    The op amp is being used as a voltage follower for a pH sensor in a very low power application. Power is supplied intermittently, but the electrode is permanently connected to the amplifier input. The goal is to keep the input leakage current down below 50fA at all times.

    Something like the LMC8101 could work, although the input bias current is too high on that particular device.

    So far we have not bee able to find a reed relay with a low enough coil current. The whole system has only ~5mW available.

    Is there any type of op amp or buffer available without input protection?



  • Hi Ord,

    A device with no ESD protection will not last very long in the real me...

    We did experiments with the LMC6001 where we cut out the ESD diodes, just to see what the gate current really was. They did not last very long. An abs-max of 16V is 16V...period...and with no input protection, just looking at it funny kills the input. If you forgot to hard-ground yourself, or your tools, or discharge anything that may carry a charge - you will zap the gates. That is why you do not see any commercial devices without input protection of some sort.

    You could use a discrete unprotected MOSFET stage - but you may run into the same "dead-junction" and "zapping" issues. And you would have to guarantee that the MOSFET gate never sees more than a few volts. This may be possible if the amp and the sensor are packaged and sealed together. I would not do this if there is a cable or connectors involved.

    Or...If you are clever, you can use the little glass enclosed SMT reed switches (think reed relay w/o a coil) and a magnet connected to the power switch, cover or similar mechanical "tun-on" device to actuate the reed. The reeds are available in both N.O and N.C configurations. Even SPDT...which I would recommend so that you ground the input to the amplifier while the probe input floats.

    I have used the Coto line of guarded reed-relays in many femto-amp applications. You only need to turn the relay on for the time you are measuring. You can save a little power by under-voltaging the coil (~60% voltage is usually enough for reliable pull-in). The "load" is practically non-existent, so you do not need to slam the reeds shut. If you have a few ohms of contact resistance,  that is nothing compared to Gigohms...

    Just some other thoughts...

    What about a micro or nano-power op-amps that can be on continuously. I'm assuming your Ph probe does not need a lot of bandwidth.

    Have a look at the:

      LPV521 (400nA supply current, 10fA typ ib)

      LPV531 (5uA to 475uA adjustable, 50fA typ ib)

      LMC6442 (1.3uA supply current, 50fA typ)

    With Shutdowns:

     LMV341-N (20fA typ, 45pA typ shutdown current)

     LMV601  (50fA typ, 33nA typ shutdown current)

    Just remember that if you use a nano-power device, you will need to scale up your feedback resistors, as they may be drawing more current than the amplifier itself. But larger resistors will increase the noise of the circuit....but you can add more filtering since it will be on continuously and will not need to settle...but you will then slow down your sensor response time....Ah, but "engineering" is all about tradeoffs...

    Hope these suggestions can help...or give you some new ideas.



  • Thanks Paul,

    those are some very helpful ideas, and some quick tests have shown that there are a few options that may work. So far the simplest is pulsing a relay and sampling only when the contacts are closed - a low enough duty cycle keeps us within the available power.

    The nano powered amps are also an interesting solution, since the size & cost of a relay is a bit of an issue in this application.

    best regards,