Input protection ADA4817

Hello on and all!

Currently i am working on a CCD video amplifier circuit. The ADA4817 build up an important part as a programmable an as a instrumentational amplifier.

Now i have measured a behavior i can not explain myself...

For a simplified representation we look at the ADA4817 as a voltage follower. R_L is 500R0 and on the input we have a sample and hold stage. Supply voltage  +VS= 6.5V; -VS= -3.5V

The input voltage is from 0 to about 4V.

In OrCAD PSPICE 16.5 i have simulated a DC sweep from 0 to 5V. (blue:input; green: output; red: input curren ->1Ohm Shunt)

For me the result looks like an input protection diode that becomes conductive, due to the input current...more or less ?!

This result is the absolute horror, cause i have to reach the 4.096V... that is not possible with the current result!

Now with a 1kOhm resistor series to the input voltage. -> Current limited to Vin_max/1kOhm (4mA)

This time i can`t see an effect like in the first picture,here it looks like saturation or something else. Output current is limited to 8mA maximum and the output swing minimum is 6.5V-1.3V -> 5.2V, so what do i have here?

In conclusion:

The datasheet of the ada4817 doesn`t say anything about a limitation of the input current. In the maximum ratings, i can see that the input voltage can be +-Vs + -0.5V. So it is necessary to limit the input current to reach the maximum input voltage without an effect like in the first picture?

Do i have overlooked something in the datasheet?

Thanks a lot!!

  • I think i have found the answer...

    The Key-Word is "Common mode input voltage range" (VCM )...

    -Vs to +Vs -2.9V

    in my case, it is limited to 6.5V-2.9V=3.6V (typical)...

  • 0
    •  Analog Employees 
    on Nov 19, 2013 6:41 PM

    Hi, TinoK.

    You are correct that the cause of your problem is driving ADA4817 with an input voltage above the specified value. However, the IVR that you should follow is -Vs to +Vs -2.8V because you are using ±5V supply. The -Vs to +Vs -2.9V you referred is for a 5V operation. IVR is usually overlooked by many but despite this little slip, I'm glad that you had solved your problem immediately.

    If you need a buffer for a CCD or for other video applications, you may look at ADI's Video Op-amp portfolio which can be found here. It has varieties of high speed amps (up to 1GHz), low cost amps and rail-to-rail output amps which would hopefully be helpful to you.

    Good luck!


  • 0
    •  Analog Employees 
    on Jan 22, 2014 7:42 PM

    Hi, TinoK.

    Are you only simulating ADA4817? If so, try using Multisim as your simulator for this is used for the evaluation of ADI spice models. You can download the Multisim here -

    If you are evaluating the part in actual or in lab, then set your IVR between -3.5V to 3.7V for a +Vs=6.5V and    -Vs=-3.5V. Using a 5.3V (assumed to be maximum peak voltage) input is not recommended because it is above the required IVR. Also, operating the part at or near the absolute maximum rating would cause degradation to long term reliability thus this is not guaranteed. Better stay within the values in the specification table to make the part work properly.



  • Dear Anna,

    i have to pick up this topic again.

    Could you tell me what`s the reason for this nonlinearity?

    Internal protection diodes or a nonlinear area of the FET?

    The other important question is, is this condition critical for the OpAmp and could this lead to damages?

    The input voltage can be described as a rectangle with 5.3V (worst case) amplitude. Hold time is around 200ns.

    A series resistor of 33.7Ohms is placed on the positive input.

    -> ADA4817 configured as voltage follower (Gain=1)

    By simulating this case, the input current is about 24mA. The input voltage is less then the absolute maximum ratings. (0.5V-Vs to Vs+0.5V)

    Thanks a lot!