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.

LMP2232 - Input structure when power to device is off

Other Parts Discussed in Thread: LMP2232

One of my customers is using them LMP2232 in a bridge sensor design.  The bridge impedance is approximately 10K ohms with a 3.3V excitation voltage.  LMP2232 has a nominal 3.3V V+ rail.

Concern - The design will have a FET switch in the power circuit to to turn off the power rail when not in use to reduce even further the circuit quiescent current.  The bridge drive voltage will still be present.

Question(s) -  

1) Will the presence of voltage on the input, with no V+ rail be detrimental to the device.

2) Will the input impedance change significantly with no V+ rail ? (will new leakage paths occur)

Regards,

Dave Johnson

  • Hi Dave,

    #1. Maybe

    #2. Depends.

    Hope that helps... :*)

    The inputs are protected by ESD structures to each supply. Basically there is a diode from V+ to the input, and V- to the input.

    These structures are there to clamp any voltage on the input that exceeds the supplies. They basically act like diodes, so they will start conducting around 300mV and will be fully conducting at around 600-800mV.

    When the power is off, the diode will clamp the input to the V+ supply by a diode drop. So if the supply is at 0V, then the diode will start to conduct starting about 300mV and will be fully clamped by 700-800mV.

    How much you can abuse the inputs is usually described in the Absolute Maximum Ratings section as either a maximum current or a maximum voltage. The LMP2232 does not mention a maximum current, but a maximum voltage. The LMP2232 Abs Max spec is ±300mV beyond the rails. Theory is, that because it is a diode, if you are exceeding 300mV, the diode is tuned on and you are already drawing a large current and violating the spec (I am not a fan of this way of thinking...and prefer a max current rating).

    In general, the LMP series can withstand at least 5mA through the ESD diodes without "damage". "Damage" is defined as when any spec becomes out of tolerance (usually Ibias). However, the diodes are rated for a one-time event. The ESD diodes are NOT designed to provide long-term or continuous input clamping. It is highly recommended that you use an external clamping diode (preferably Schottkey with low Vr) if you expect these conditions to occur regularly, to take the load off the internal diode.

    Another thing to keep in mind is that the ESD or clamp diode will be conducting the voltage into the supply line. If the input source impedance is low enough, you could start "lifting" the floating power supply line. Bypass capacitors will start to charge, and provide a low-impedance charge reservoir to things that may try to start-up as the supply voltage slowly increases (can cause "chirping" of switchers or other devices on the same supply line - or a dimly glowing power LED).

    So to really answer your questions:

    #1 -  If you limit the current into the pin to less than 1mA, then there should not be any problems. Limiting the current can be as simple as adding series resistance - which could be part of the existing feedback or voltage divider network. If the maximum available voltage is 3.3V, then 3.3V/1mA = 3.3K minimum resistance needed to limit the current.

    With a 10K sensor impedance, this should suffice as a current limit - but make sure the sensor will not have problems with the large, unbalanced current flow because it's output is tied to ground by a diode drop. If this is a battery powered device, the current through the diode may shorten the battery life due to the leakage current (3.3V-0.7V=2.6V, 2.6V/10K = 260uA...times two).

    #2 - You are looking into a conducting diode - so the "input impedance" will be changing depending on the voltage applied to the input.

    The only way to keep the input impedance high is to use a device with shutdown. This way, the supply pin is at 3.3V and the diode is not conducting, but the internal circuitry is disabled so the input impedance does not change. The input will "look" roughly the same, but the amplifier is disabled. But keep in mind in some circuit configurations, the feedback resistors can create a current path since they are not disconnected during shutdown - unless you go directly into the inputs (like a 3-amp instrumentation amp).

    If they want to save power - they may have to turn off the bridge power, too.

    Regards,