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.

OPA211: Input protection when power supply is turned off

Part Number: OPA211

Hello!

A lot of op-amps (e.g. the OPA211) require the input voltage at their input pins (inverting and non-inverting) to stay between the power supply rails.

Let's say I want to use the OPA211 as an input buffer, which buffers an external signal supplied to a circuit. When I ever switch off the power supplies to that circuit, without disconnecting the external signal first, this could probably break the op-amp.

What is the best way of protecting the op-amp against this situation? 

I am building a very low noise circuit, so is there a good solution which doesn't add additional noise?

Thank you so much!

Best,

Niki

  • Hi Niklas,

    the usual method is to use a current limiting resistor at the input. Section 8.1.2 and section 8.1.7 of datasheet discuss this issue.

    A more sophisticated method is to use an additional diode clamp in combination with additional current limiting resistors as shown here (the picture is from another thread):

    By this method the total series resistance can be decreased which is helpful in low noise applications. What maximum series resistance can you afford?

    A third way is to short circuit the OPAmp input to signal ground by the help of a relay or P-channel JFET whenever the supply voltage is down, best in combination with a voltage divider to limit the input current.

    Or you use a relay to only switch the signal to the OPAmp input when the OPAmp is powered-on.

    Kai

  • Hi Kai!

    Thank you so much for your reply!

    I would like to limit the input resistance to 50 ohms. And the power supply rails are +/- 15V. The relay/JFET solution sounds awesome!

    Could you recommend a certain relay / JFET ? My expected bandwidth is up to about 1 MHz. And can relays have any problems with noise pickup?

    Thank you so much! You are helping me a lot!

    Best,
    Niki
  • Hi Niklas,

    Kai has provided a good protection circuit utilizing the Schottky diodes from the inputs to the supply lines and TVS diodes off the supply lines to ground. You certainly don't want to add much series input resistance because it does add thermal noise and defeats the point of using a very low voltage noise density op amp such as the OPA211.

    Schottky diodes will turn with a much lower forward voltage than the OPA211 internal ESD diodes. In fact, with the external Schottky diodes the internal ESD diodes should remain off, even as the supplies are turned off. If the supplies are turned off with the input voltage present the OPA211 will simply draw its supply current from the input source until that voltage collapses.

    Since Schottky diodes can have higher leakage current compared to conventional PN junction diodes they should be selected for minimum reverse leakage current. Otherwise, the leakage current can rival, or exceed, the input bias current of the OPA211. Be aware that the leakage current about doubles for every ten degrees of increased temperature. This usually isn't an issue below 25 C, but it can become quite large at high operating temperatures.

    The SDM02M30LP3 surface mount Schottky diode is an example of a low-leakage diode that can handle up to 100 mA forward current, and reverse voltage of 30 V. There are certainly others available to consider.

    Regards, Thomas
    Precision Amplifiers Applications Engineering
  • Hi Thomas and Kai!

    Thank you for the good explanation! That protection circuit you suggest makes a lot of sense! But I'm still wondering, if I want at most 50 Ohms of series resistance, an input voltage of 10V would still cause a 200mA current, when the OPA211 is shut off. And I really want to avoid a larger input resistance.

    I'm still excited about Kai's relay-idea! Just disconnect the signal to the op amp input, when the op-amp looses power. I just don't have any low-noise experience with relays yet. May I ask you a few more questions?

    Can relays add noise the signal, which is more than 1nV/sqrt(Hz)? Do they have a bandwidth of up to 1MHz? Would you recommend a mechanical or a solid-state relay?

    Thank you so much!

    Best, Niki

  • Hi Niklas,

    the lowest contact resistance can be achieved with a mechanical relay. Search for a relay with massive gold contacts positioned in an inert atmosphere which allows "dry switching".

    The only contribution to noise comes from the contact resistance (which is negligible here) and from a bit stray capacitance which can inject noise from the ambient or the relay coil voltage. But when you take a shielded relay (or a relay with added shielding) and when you filter the relay coill voltage sufficiently there's no considerable noise contribution from the relay. There are also HF relays arround which show only sub-pF stray capacitance.

    Can you give us a schematic which shows the input section of your circuit and helps to understand whether a relay suited for "dry switching" is needed?

    Kai

  • Hi Niki,

    You state, "But I'm still wondering, if I want at most 50 Ohms of series resistance, an input voltage of 10V would still cause a 200mA current, when the OPA211 is shut off. And I really want to avoid a larger input resistance."

    When your power supplies are turned "off "they likely become a high impedance circuit instead of a zero ohms circuit. Ideally, just the OPA211 current would be drawn through the 50 Ohm resistor when they are off, but in actuality anything else connected to the the same supply lines as the OPA211 will attempt to draw current from the input source as well. The OPA211 and those other circuits connected to that supply line will load the input source, but they have limited ability to draw current. The actual current under that condition is possibly much less than the anticipated 200 mA.

    I suggest taking a look at the total circuit consisting of the OPA211 and any other subcircuits connected to the common supply lines, and evaluating what the current can be when the supplies are "off." I suspect the current will be much less than 200 mA. If so, the simple diode protection circuits can provide an easily implemented, low cost solution.

    Regards, Thomas
    Precision Amplifiers Applications Engineering
  • Niklas

    We haven't heard back from you so we assume this answered your questions. If youi need additional help just post another reply below.

    Thanks
    Dennis