How to address gain error issues with AD817?

I am using AD817 with a Burr Brown BUF634 in the feedback loop (boosts current to 250mA) as an instrumentation buffer. Specifically I am using AD817 for it's ability to drive capacitive loads because my buffer goes to a 50 ohm coaxial cable that eventually goes to a digitizer in a PC. I noticed an offset that is reasonable as per the AD817 datasheet, but I also see a gain error that is up to 15% for voltage readings captured by the digitizer. I quantified this by forcing a series of DC voltages and digitizing them. At lower voltages (0 to .5V) there is little error, but by the time it is 1.5 to 2V I am seeing 80 - 100mV error. It seems to change slope around 1V, making the issue worse.

Here is my circuit:

I have voltages higher than 16V present at the input, so I have a HV mosfet with 5V on the gate in series with the BufIn and the signal I want to capture as a voltage blocker. But the measurements were all taken with scope probes on BufIn and BufOut and the oscilloscope agrees with the digitizer. I also have clamp diodes to ground and 12V as additional protection for the AD817. The 1K resistor to ground is do not load because it will allow current to flow to ground through my voltage blocking mosfet and it can become damaged.

Any ideas how to improve the gain error in my circuit? I could measure voltage very accurately with the digitizer but for this problem. Thanks a lot!

  • 0
    •  Analog Employees 
    on Feb 6, 2015 12:14 AM

    Hi, Jon.

    Are you trying this in the lab? If so, please check your ground connection especially the probe ground pins. You can use a ground coil which is included in the probe accessories bag for your ground in order to eliminate this error. Be sure that your ground is near to the probe tip. Also, please clean your board with an eraser (pencil eraser would do) or a flux remover for there might be dirt causing this error.

    Let me know if this works.



  • Hi again, I found that my clamp diode to ground was damaged and leaky to the tune of 30uA. After I replaced the diode, the DC gain error was gone:

    However, I still see around 80-100mV difference between the buffer in and buffer out voltages in my Rdson test:

    I tried the following things to improve my results:

    1. Implemented a 10k potentiometer as offset null adjust, very sight improvement.
    2. Switched the 3K resistors at the input an int he feedback loop to 1.5K 0.1% resistors.
    3. Changed my 100 ohm resistor at the Buf634 input to a 0.1% 100 ohm.
    4. After these changes I still see the buffer output ~50-80mV lower than the input.

    Any suggestions would be appreciated. Thanks.

  • 0
    •  Analog Employees 
    on Aug 2, 2018 2:53 PM
    This question has been assumed as answered either offline via email or with a multi-part answer. This question has now been closed out. If you have an inquiry related to this topic please post a new question in the applicable product forum.

    Thank you,
    EZ Admin